You are on page 1of 426

._... ~ .._,__ ... .,..,.. - -....• ,._...,. •• tr., .,.,.--...~----~...,....._...... _ _ _ ~-· -·-~· ~ - ~ .... .....


- . ~ - --- .




Published by

Sixth Printing-March, 1973
Library of Congress Catalog Card Number: 62-21097

Copyright 1962



All rights reserved. This book, or parts thereof, may not be re-
produced in any form without permission of the publisher, ex-
cept by a reviewer who wishes to quote brief passages in
connection with a review written for inclusion in a magazine or

Printed in U.S. A.

This book was prepared by the engineering and technical

staff of Betz Laboratories, Inc. Many individuals contributed
to this edition and previous editions. We particularly wish to
thank the following for their editorial contributions to this
Sixth Edition.

H. L. Boyer A. J. Piluso
A. B. Cherry 0. H. Preis
R. B. Conlan M. U. Priester
J. M. Donohue L. F. Probst
C. M. Frova R. L. Reed
P. J. Gaughan J. H. Richards
H. B. Haldeman C. J. Schafer
R. H. Hayman J. R. Schieber
J. B. Lord B. F. Shema
J. J. Maguire W. A. Tanzola
S. 0. Meyer J. G. Weidman
W. L. Nieland G. A. Wilhelm
C. A. Noll


Chapter Page
1. Industrial Water Conditioning 8
2. Water and Its Impurities ................................................ . 12
3. Aeration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
4. Subsidence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
5. Coagulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
6. :Filtration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
7. Precoat Filtration ......................................... , . . . . . . . . . . . . . . 42
8. Chlorination . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
9. Hydrogen Sulfide Removal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
10. Iron Removal and Iron Retention . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
11. Lime-Soda Softening . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
12. Hot Process Phosphate Softening . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
13. Silica Removal By Magnesium Salts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
14. Sodium Zeolite Softening . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
15. Cation Exchange by Hydrogen Zeolite . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
16. Demineralization ......................................................... 102
17. Dealkalization by Chloride-Anion Exchange . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
18. Hot Lime-Hot Ion Exchange Softening ..................................... 115
19. Oil Ren1oval . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
20. Deaeration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
21. Chemical Feed Systems ................................................... 135
22. Boiler Scale Control ...................................................... 151
23. Boiler Corrosion Control .................................................. 167
24. Carryover ............................................................... 174
25. Silica Turbine Blade Deposits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182
26. Measurement of Steam Purity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189
27. Condensate Return Line Corrosion ......................................... 194
28. Embrittlement of Boiler Metal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 204
29. Boiler Blowdown Control .................................................. 210
30. Chemical Cleaning of Boilers .............................................. 221
31. Cooling Water Treatment: Once-Through Scale Control ..................... 229
32. Cooling Water Treatment: Once-Through Corrosion Control .................. 236
33. Cooling Water Treatment: Open Recirculating Scale Control .................. 247
34. Cooling Water Treatment: Open Recirculating Corrosion Control .............. 258

35. Cooling Water Treatment: Closed Recirculating Systems ...................... 270

36. Pretreatment and On-Stream Cleaning ..................................... 274
37. Acid Feeding to Cooling Water Systems ..................................... 279
38. Slime and Algae Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 288
39. Cooling Tower Wood Deterioration ........................................ 300
40. Air Conditioning Systems .......................... .' ...................... 311
41. Industrial Waste Treatment ............................................... 319


42. Methods and Equipment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 329
43. Expression of Analytical Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 335
44. Acid (Free Mineral) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 338
45. Alkalinity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 339
46. Alkalinity (Hydrate) ..................................................... 343
4 7. An1monia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 345
48. Calcium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 347
49. Carbon Dioxide ......................................................... 349
50. Chloride . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 353
51. Chlorine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 356
52. Chron1ate ............................................................... 359
53. Color .................................................................. 363
54. Conductance (Specific) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 365
55. Hardness ............................................................... 370
56. Hydrazine .............................................................. 374
57. Iron ................................................................... 377
58. Manganese . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 380
59. Nitrate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 382
60. Nitrite . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 384
61. Oxygen (Dissolved) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 386
62. pH .................................................................... 390
63. Phosphate 394
64. Silica . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 398
65. Solids (Total, Dissolved, Suspended) ....................................... 402
66. Sulfate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 404
67. Sulfite . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 407
68. Turbidity 409
69. Composition of Prepared Reagents ......................................... 411
Conversion Factors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 415
Index .................................................................. 417

Industrial Water Conditioning

What Is Industrial Water Conditioning?. . . . . . . . 9

External Treatment . . . . . . . . . . . . . . . . . . . . . . . 9
Internal Treatment . . . . . . . . . . . . . . . . . . . . . . . . 9
Water Conditioning Control. . . . . . . . . . . . . . . . . 10
Industrial Water Problems. . . . . . . . . . . . . . . . . . 10

used directly or indirectly in an

W indu~trial process, is classed as indus-
trial water. The use of w~ter .in boil:rs for
steam generation is an obvwus mdustnal us~.
Cooling water, either o~ a once-t~roug~ basiS /
or with cooling towers, Is a promment mdus-
trial use. Water is essential to large air con-
ditioning systems. But, water has a multitude
of industrial uses-some as commonplace as
washing of equipment-others specific t~ an
individual industry or plant, such as use m a
high pressure spray for debarking logs. Water
is incorporated in many products and foods.
Water is used as a carrier in industrial proc-
esses and as a convenient means of disposing
of wastes. Figure 1-1 • External Treatment
by Hot Process Softening
INDUSTRIAL WATER CONDITIONING? use of equipment, which may include hot or
cold process lime-soda softeners, zeolite and
To many people, industrial water condition- other ion exchange systems; deaerators, filters,
ing is shrouded in an air of mystery. Much of clarifiers, etc. Such equipment is used for the
this confusion is due to a lack of understand- purpose of reducing hardness and alkalinity,
ing of what industrial water conditioning is eliminating dissolved oxygen and for the
. . . and what it is supposed to accomplish. removal of suspended solids.
Basically, industrial water conditioning em- The use of external treatment should be as
braces the broad field of fitting water to the complete as can be justified by the severity
job. Its purpose is twofold. One-it involves of the water problem and engineering
removing or minimizing undesirable charac- economies.
teristics of water, such as removing hardness Regardless of the purity of the water pro-
by softening to avoid boiler scale. Two-it vided by the use of such equipment, addi-
involves adding desirable properties to water, tional internal chemical treatments normally
such as adding chromate ion to give cor- are required for complete protection against
rosion inhibitory properties. scale, corrosion, and a host of other potential
Industrial water conditioning may be sources of trouble.
grouped into two major parts or spheres of There are often cases where an investment
operation-external treatment and internal in external treatment equipment is not war-
treatment. ranted because of non-critical requirements,
a good water supply, small size of the plant,
EXTERNAL TREATMENT or a combination of these reasons. In such
cases, internal chemical treatment alone, will
This phase of industrial water conditioning produce successful operating results.
has many names: preliminary treatment,
pre-treatment, indirect treatment, primary
purification, etc. Whatever it is called, it
means doing something to a water to make it Like external treatment, this phase of indus-
more suitable for its intended application trial water conditioning is known by several
before it reaches the point of use. names: secondary treatment, after-treatment,
External treatment usually requires the direct treatment, etc. In this case, it means

doing something to a water at the point of use ing systems. These cas~s include the use of
to make it suitable for its intended application. chemicals at various points in process appli-
In boiler water systems, the objective of cations to impart desirable properties to
internal treatment is to make possible the a water.
control of scale formation and corrosive Water conditioning is in many respects
action. Steam purity and caustic embrittle- unlike any other field of engineering. It de-
ment must be controlled. It also may be mands an unusual variety of talents including
necessary to use internal chemical treatment · those of the chemical, mechanical, corrosion
to prevent return line corrosion due to dis- and sanitary engineer, microbiologist, physi-
solved gases liberated in the boiler system. cal chemist and bacteriologist.
Internal chemical treatment is used also
in cooling water systems for the prevention WATER CONDITIONING CONTROL
of scale and corrosion, as well as conditions
brought about by biological growths. Proper Control water analyses are an essential part
application of internal chemical treatment of any industrial water conditioning program.
results in improved heat transfer by eliminat- The processes of corrosion, scaling, slime for-
ing these insulating deposits. mation and the like are dynamic, not static.
The treatment measures employed for their
There are cases, too, where internal chemi-
control must be adjusted frequently on the
cal treatment is used in the solution of specific
basis of control water analyses. Such analyses
water problems other than in boiler and cool-
need not determine every ion or substance
present in the water, but only those required
for control of treatment process. For example,
in sodium zeolite softening the important
tests for control purposes are hardness and
chloride. The hardness test is employed to
determine the end of the softening cycle and
the chloride test to determine the end of the
rinse cycle.


Following is a list of the more common water

problems that plague industry. Each industry
has its individual problems. This list is not
intended to be complete, but only to indicate
the great variety of problems common to
most industries.

Boiler Water Problems

Scale Deposits
Excessive Sludge
General Corrosion
Feedline and Heater Deposits
Figure 1-2 • Internal Treatment High Blowdown Rate
Chemical Tanks and Pumps Carbonized Oil

Steam and Condensate Problems

Carbon Dioxide Grooving

Oxygen Pitting
Turbine Blade Deposits
Superheater Fouling
Oil Contamination
Condenser Leakage
Process Contamination

Cooling and Process Water Problems

Deposition of Scale
General Corrosion
Fungus Attack of Cooling Tower Wood Figure 1-3 • Plant Operator
Product Contamination Conducting Control Ana(yses
Pitting and Tuberculation
Slime and Algae Deposits
Red Water

Preliminary Treatment Problems General Problems

(removal of impurities)
Poor Quality Water Supply
Turbidity, Mud, Silt Waste of Water-Lack of Re-use
Excessive Color Inadequate Pretreatment Equipment
Iron and Manganese Improper Control of Existing Equipment
Soluble Silica Unsatisfactory Chemical Feeding
High Hardness Stream Pollution
Acidity or Alkalinity Waste Disposal
Objectionable Gases-C0 2 , H 2 S, NH 3 High Costs
High Solids Improper Over-all
Microorganisms Water Conditioning Program

Water and its Impurities

Sources of Water . ......................... 13

Rain Water. . . . . . . . . . . . . . . . . . . . . . . . . . . 14
Sw:face Water. . . . . . . . . . . . . . . . . . . . . . . . . . 14
Ground Water. . . . . . . . . . . . . . . . . . . . . . . . . . 14
Impurities. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

F water as it occurs in nature were "pure" of hydrogen and one volume of oxygen, chem-
I water and nothing else there would be no
need for water analyses and water condition-
ically combined.

ing. Whatever the source, water always con- SOURCES

tains impurities in solution or suspension. It
is the determination of these impurities that The various sources of water can be con-
makes water analyses necessary and the con- veniently classified as:
trol of these impurities that makes water con- Rain water.
ditioning essential. Surface water (streams, ponds, lakes and
Pure water is tasteless, colorless and odor- impounded reservoirs) .
less and is one of the most universal solvents. Ground water (from springs, shallow wells
The chemical formula for water is H 2 0 (or and deep wells) .
HOH) and it is composed of two volumes Water is evaporated from the surface of the

Figure 2-1 • Calcium Sulfate Crystalline Scale in 5 inch Line

Figure 2-2 • Pipe Section Pitted from Dissolved Oxygen

earth by heat furnished from the sun. Water organic matter and may pollute the water
precipitated from the atmosphere, high above with animal waste. Swamps may discharge
the earth's surface, is chemically pure. As their waters during floods and carry decay
water descends through the air oxygen, nitro- and vegetable matter, color and microscopic
gen and carbon dioxide (the normal gases organisms into the stream. In addition, sur-
present in the atmosphere) are dissolved in face waters are exposed to pollution by ani-
the falling water in proportion to their partial mals and humans, the sewage of cities, and
pressure in the atmosphere. Rain water also the wastes of industry.
encounters dust, smoke and fumes which are SuRFACE WATERS. Surface waters obtained
dissolved or retained in suspension. Bacteria from ponds, lakes and impounded reservoirs
and the spores of microscopic organisms may are similar to those from flowing streams.
be picked up. However, within these large bodies of water
RAIN WATER. Rain water in descending considerable self-purification may take place
through the air and in percolating through and the water quality may change to a large
the upper layers of soil absorbs carbon di- extent.
oxide with which it forms carbonic acid. GRoUND WATER. Ground water obtained
This action increases the solvent power of from shallow wells may be hard or soft de-
the water so that it dissolves a certain pending upon the mineral characteristics of
amount of the mineral matter of the soil the surrounding area. Natural filtration
or rock with which it comes in contact. through the sand will usually provide a water
Water obtained from surface streams may be relatively free of turbidity and with low
rendered turbid by the presence of clay and organic matter. The water from shallow
silt. Agricultural land may contribute to the wells is usually softer than deep wells.

The water from deep wells usually contains

relatively high concentrations of dissolved
minerals. These waters are usually clear and
colorless due to the filtration taking place
through the layers of sand.
Spring water is usually similar in character-
istics to water obtained from shallow or deep
wells in the surrounding area. However,
springs are more likely to be affected by the
normal surface pollution due to their rela-
tively shallow origin.


The impurities that may be present in a

water supply can be divided into suspended
and dissolved solids. Suspended solids are
those which do not dissolve in water and
which can be removed or separated by filtra-
tion. Examples of suspended solids are mud,
clay or silt. Dissolved solids are those which
naturally dissolve in the water and which
therefore cannot be removed by filtration. The
presence of hardness or chloride in a water
supply is an example of dissolved solids. Gases
may also dissolve in water, but unless they
combine chemically with other impurities also
in solution, these gases will be expelled from
the water on boiling and are not considered
as dissolved solids.
Fig. 2-4 is a listing of the analytical deter-
minations made in the examination of most
natural waters. Illustrated are the chemical
formula, if one can be given, the difficulties
commonly caused by the presence of that con-
stituent and the usual means of treatment to
prevent these difficulties. This listing covers
the usual determinations made in the analysis
of a water prior to treatment. Of course, for
certain special water uses, additional analyses
are necessary. For example, boron is of im-
portance in irrigation waters because of its
effect on the growth of certain crops. Boron
possesses little significance industrially and
usually is not determined in a water analysis.
Similarly, water analyses are made for certain
constituents that are added to water as a
:nethod of treatment. For example, chromate Figure 2-3 • Delignification
IS employed as a corrosion inhibitor in cooling of Cooling Tower Wood

water systems, phosphate as a precipitating the constituents that normally are present
agent in boiler water and sulfite is used to naturally in untreated water supplies and is
control oxygen corrosion. Analyses must be not intended as a complete listing of all of
made of such treating agents for control pur- the analytical determinations that should be
poses. Figure 2-4, however, deals only with made where specific problems are involved.

Fig. 2-4-Common Impurities Found in Water


TURBIDITY . . . . . . None-expressed in Imparts unsightly appearance to Coagulation, settling and filtra-
analysis as units. water, Deposits in water lines, proc- tion.
ess equipment, boilers, etc. Inter-
feres with most process uses.

COLOR . . . . . . . . None-expressed in May cause foaming in boilers, Hin- Coagulation and filtration. Chlor-
analysis as units. ders precipitation methods such as ination. Adsorption by activated
iron removal, hot phosphate soften- carbon.
ing. Can stain product in process
HARDNESS ..... Calcium and mag- Chief source of scale in heal ex- Softening. Distillation. Internal
nesium salts ex- change equipment, boilers, pipe boiler water treatment. Surface
lines, etc. Forms curds with soap, active agents.
pressed as CaC03.
interferes with dyeing, etc.
ALKALINITY ..... Bicarbonate (HC03), Foaming and carryover of solids lime and lime-soda softening.
carbonate (C0 3), and with steam, Embrittlemenl of boiler Acid treatment. Hydrogen zeolite
steel. Bicarbonate and carbonate softening, Demineralization. De-
hydrate (OH), ex-
produce C02 in steam, a source of alkalization by anion exchange.
pressed as CaC03. corrosion, Distillation.
FREE MINERAL H2S04, HCI, etc. ex- Corrosion Neutralization with alkalies.
ACID ........ . pressed as CaC03.

CARSON Corrosion in water lines and par- Aeration. Deaeration. Neutraliza-

DIOXIDE ticularly steam and condensate tion with alkalies. Filming and
lines. neutralizing amines.
pH ........... . Hydrogen ion con- pH varies according to acidic or pH can be increased by alkalies
centration defined as: alkaline solids in water. Most natu- and decreased by acids.
·ral waters have a pH of 6.0-8.0.
pH= log ------c:---
(H +)

SULFATE ....... Adds to solids content of water, Demineralization. Distillation.

but, in itself, is not usually signifi·
cant. Combines with calcium to form
calcium sulfate scale.
CHLORIDE.,.,., Cl- Adds to solids content and in- Demineralization. Distillation.
creases corrosive character of
NITRATE ... , .. , Adds to solids content, but is not Demineralization. Distillation.
usually significant industrially. High
concentrations cause methemoglob-
inemia in infants. Useful for con-
trol of boiler metal embrittlement.
FLUORIDE ... , , . F- Cause of mottled enamel in Ieeth. Adsorption with mangnesium hy-
Also used for control of dental de- droxide, calcium phosphate, or
cay, Not usually significant indus- bone black. Alum coagulation.

Fig. 2-4-Common Impurities Found in Water (cont.>


SILICA . . . . . . . . . Si02 Scale in boilers and cooling water Hot process removal with magnes-
systems. Insoluble turbine blade ium salts. Adsorption by highly
deposits due to silica vaporization. basic anion exchange resins, in
conjunction with demineraliza-
tion. Distillation.

IRON ......... Fe ++(ferrous) Discolors water on precipitation. Aeration. Coagulation and filtra-
Fe+++ (ferric) Source of deposits in water lines, tion. Lime softening. Cation ex-
boilers, etc. Interferes with dye- change. Contact filtration. Surface
ing, tanning, paper mfr., etc. active agents for iron retention.

MANGANESE . . . Mn + + Same as iron. Same as iron.

OIL . . . . . . . . . . . Expressed as oil or Scale, sludge and foaming in boil- Baffle separators. Strainers. Co-
chloroform extract- ers. Impedes heat exchange. Un- agulation and filtration. Diatoma-
desirable in most processes. ceous earth filtration.
able matter.

OXYGEN . . . . . . . 02 Corrosion of water lines, heat ex- Deaeration. Sodium sulfite, Cor-
change equipment, boilers, return rosion inhibitors.
lines, etc.

HYDROGEN Cause of "rotten egg" odor, Cor- Aeration. Chlorination. Highly

SULFIDE ...... . rosion. basic anion exchange.

AMMONIA . . . . . NH3 Corrosion of copper and zinc alloys Cation exchange with hydrogen
by formation of complex soluble zeolite. Chlorination. Deaeration.

CONDUCTIVITY Expressed as mi- Conductivity is the result of ioniz- Any process which decreases dis-
cromhos,specific con- able solids in solution. High con- solved solids content will decrease
ductivity can increase the corrosive conductivity. Examples are de-
characteristics of a water. mineralization, lime softening.
DISSOLVED None "Dissolved solids" is measure of Various softening process, such as
SOLIDS ....... total amount of dissolved matter, lime softening and cation ex-
determined by evaporation. High change by hydrogen zeolite, will
concentrations of dissolved solids reduce dissolved solids. Deminer-
are objectionable because of proc- alization. Distillation.
ess interference and as a cause of
foaming in boilers.
SUSPENDED None "Suspended Solids" is the measure Subsidence. Filtration, usually
SOLIDS •...... of undissolved matter, determined preceded by coagulation and
gravimetrically. Suspended solids settling.
plug lines, cause deposits in heat
exchange equipment, boilers, etc.
TOTAL SOLIDS . . None "Total Solids" is the sum of dis- See "Dissolved Solids" and "Sus-
solved and suspended solids, de- pended Solids."
termined gravimetrically.


Theory Involved. . . . . . . . . . . . . . . . . . . . . . . . . . 19
Basic Aeration Methods. . . . . . . . . . . . . . . . . . . . 19
Water Fall Aerators . .................... 20
Air Diffusion . . . . . . . . . . . . . . . . . . . . . . . . . . 21
Applications. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
Limitations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

AERATION is a process which consists of bon dioxide and no measurable amount of

1"1. intimately admixing water and air in hydrogen sulfide.
some manner. This mixing may be accom- Both carbon dioxide and hydrogen sulfide
plished in the form of a thin film, drops or are relatively soluble in water. At 20 C and
spray. Aeration is based on the establishment sea level, carbon dioxide is soluble to the ex-
of a state of equilibrium between the gases tent of appro~imately 1700 ppm while hy-
present in the water and those present in drogen sulfide solubility is approximately
the atmosphere. 3900 ppm. The partial pressure of each of
Aeration is employed primarily for the these gases in a normal atmosphere, how-
removal of undesirable gases from water. ever, is practically "zero." Consequently, the
Carbon dioxide and hydrogen sulfide are establishment of a state of equilibrium be-
corrosive gases and, when present in appreci- tween water and air by means of aeration
able quantities, must be removed to avoid will result in saturation of the water with
corrosion of lines, pumps and equipment. Al- nitrogen and oxygen and the practically
though aeration results in saturating water complete elimination of carbon dioxide and
with another corrosive gas - oxygen - this hydrogen sulfide from the water.
process is the one most commonly used for The basic requirement in securing removal
the elimination of carbon dioxide and hydro- of gases by aeration is the establishment of
gen sulfide. a state of equilibrium between the gases in
In some cases, the purpose in removing the water and those present in the surround-
carbon dioxide by aeration is simply to re- ing atmosphere. The removal of gases by
duce chemical treatment requirements where aeration is favored by increase in tempera-
the water is subsequently lime softened. Car- ture, increase in aeration time, increase in
bon dioxide can be removed more economi- the volume of air in contact with the water
cally by aeration than by chemical precipita- and increase in the surface area of water
tion with lime. exposed to the air. The efficiency of aeration
In the removal of iron and manganese is greater where the concentration of the
from water, aeration is frequently employed. gas to be removed is high in the water and
The oxygen introduced by aeration encour- low in the atmosphere. For example, aeration
ages the chemical oxidation of iron and efficiency is higher with a water containing
manganese, thus aiding in their precipitation 100 ppm carbon dioxide, than with one con-
as the insoluble hydroxides. taining only 10 ppm.
Aeration may be employed in potable
water treatment for the removal of certain BASIC AERATION METHODS
tastes and odors, such as those resulting
from the decomposition of algae. In sewage - Aeration is primarily a mechanical process
treatment, aeration i~ used to promote the for which two general methods are em-
biochemical oxidation of organic matter. ployed. The first method is that where the
water falls through the air as with the use of
THEORY INVOLVED spray nozzles, cascade devices and forced
draft type aerators. This method of aeration
The removal of gases from water by aera- is the most common in industrial water con-
tion follows Henry's Law which, in brief, ditioning.
states that the solubility of a gas in water is The second general method of aeration is
?irectly proportional to its partial pressure where air is bubbled up through a solid body
m the surrounding atmosphere. At 20 C and of water, as in a tank, and is known as the
at sea level, water in equilibrium with the air diffusion method. In industrial water
atmosphere will contain approximately 15.8 conditioning this method is usually confined
ppm nitrogen, 9.4 ppm oxygen, 0.5 ppm car- to the treatment of relatively low flows.

WATER FALL AERATORS. Many variations of the advantage of smaller space require-
the water fall principle are in use. The sill].- ments. Coke tray aerator~. are widely used in
plest is the use of vertical risers to discharge iron and manganese removal since a catalytic
the water by free fall into a basin. The effect is secured by contact of the iron-
risers can operate on the available head of bearing water with freshly precipitated iron
water and efficiency of aeration is increased sludge. As shown by Figure 3-1, these units
by making the fall as great as practical. consist of a series of coke filled trays through
The addition of steps or shelves to break up which the water percolates with additional
the fall and to spread the water into thin aeration supplied during the free fall from
sheets or films increases contact time and acts one tray to the next. Wood slat tray aerators
as a further refinement. Spray nozzles are are similar to small atmospheric cooling
another variation of this principle and can towers. The tray slats are staggered so as to
provide efficient removal of dissolved gases. break up the free fall of water and to spread
However, space requirements frequently limit out the water surface before finally dropping
the usefulness of this type aeration in indus- into a basin.
trial water conditioning. Forced draft aerators, as illustrated by
Coke tray aerators and wood slat aerators Figure 3-2, are favored for many industrial
are relatively similar in design and possess water conditioning purposes. These units

Courtesy lnfilco, Incorporated

Figure 3-1 • Coke Tray Aerator





Courtesy Pfaudler Permutit Inc.

Figure 3-2 • Forced Draft Aerator

provide horizontal wood slat trays which by pumping air into water through perfo-
divide the falling water into small streams. rated pipes, strainers, porous plates or tubes.
Air is forced up through the unit by a blower Aeration by diffusion is theoretically superior
with uniform air distribution across the en- to water fall aeration inasmuch as a fine
tire cross section. Air flow is counter-current bubble of air rising through water will con-
to the fall of water and positive air circula- tinually be exposed to fresh liquid surfaces,
tion prevents any stagnant areas. Because of providing maximum water surface per unit
these features, forced draft aerators are more of air. Also, the velocity of bubbles ascending
efficient for gas removal and require less through the water is much slower than the
space for a given capacity. velocity of free falling drops of water, pro-
AIR DIFFUSION. Air diffusion is accomplished viding longer contact. However, for gas re-

moval, air requirements by the diffusion Aeration has also been applied as an aid in
method are higher than for odor removal rapid mixing of chemicals, and in backwash-
and the cost of the air supply is a factor ing filters as well as for the esthetic effect in
restricting application of this type of aera- water works plants.
tion in industrial water conditioning.

APPLICATIONS Theoretically, at 20 C it is possible to re-

duce the carbon dioxide content of the water
In industrial water conditioning one of the to 0.5 ppm by means of aeration to equilib.
major applications of aeration is in the re- rium conditions. Practically, it is not eco.
moval of the corrosive characteristics im- nomical to attempt such complete gas re-
parted to water by the presence of carbon moval and reduction of carbon dioxide to 10
dioxide and hydrogen sulfide. On many oc- ppm is normally considered satisfactory.
casions aeration is used to remove the carbon Although removal of free carbon dioxide
dioxide liberated by a treatment process. For increases the pH of a water and renders it
example, in boiler feedwater conditioning, less corrosive from this standpoint, aeration
it is common practice to acid treat the ef- also results in the saturation of the water
fluent from a sodium zeolite softener in order with dissolved oxygen. In some cases, where
to reduce the alkalinity of the boiler feed- the original oxygen content is already quite
water. Carbon dioxide is produced as a re- high, no undesirable effect is thus produced,
sult of the acid treatment and aeration is However, in the case of a well supply, high
employed to rid the water of this corrosive in carbon dioxide but devoid of oxygen,
gas. Similarly, when the effluents of hydrogen aeration may simply exchange one corrosive
and sodium zeolite units are blended, the gas for another.
carbon dioxide formed is eliminated by aera- The efficiency of aeration is greatest where
tion. In anion-cation exchange, aeration is the initial concentration of the gas to be re-
similarly used. Forced draft aerators are moved_ is considerably above its equilibrium
usually employed in these applications. value. Therefore, with waters containing only
Aeration is also applied for the removal of a small amount of carbon dioxide, lime treat-
methane from certain ground waters. Since ment is usually more efficient.
methane constitutes- an explosion hazard on Aeration alone will not secure the com·
release in confined spaces, aeration is em- plete removal of hydrogen sulfide, but must
ployed for its removal. Electric motors in be combined with pH reduction or chlorin·
the vicinity are undesirable and forced draft ation.
aerators are not employed in this appli-
Where odors are not due to the gases of
decomposition or to volatile oils, aeration
Iron and manganese removal can be se- will not remove these odors. Tastes and
cured with a variety of methods. In all but odors due to industrial wastes, phenols and
a few of the processes used, aeration plays a cresols are unaffected by aeration. The me·
part since it is a convenient method of sup- dicinal taste imparted to water by the chlo-
plying the oxygen required for the oxidation rination of water containing phenols and ere-
of iron and manganese. However, aeration sols is not removed by aeration.
alone should not be considered a universally
effective method for the removal of iron and
manganese, but rather only as a part of the
treatment steps required. "Aeration of Water", Journal, Am. Water Works
In municipal water treatment practice, Assoc., Vol. 47, pp. 873-885 (1955)
M. L. Riehl, "Hoover's Water Supply and Treat-
aeration has been employed primarily for ment", 8th Ed., pp. 18-19, National Lime Assoc.,
the removal of odors due to volatile materials. Washington, D.C. (1957)


Theory Involved. . . . . . . . . . . . . . . . . . . . . . . . . . 24
Equipment Employed. . . . . . . . . . . . . . . . . . . . . . . 25
Applications and Limitations . . . . . . . . . . . . . . . . 25

UBSIDENCE is the process which permits The viscosity of the water is also a major
S the settling of suspended substances from
factor influencing settling yelocity and the
water viscosity is, in turn, dependent on the
The main purpose of subsidence is to ef- temperature. For example, the rate of set-
fect clarification of a water by permitting tling at 0 C is only 43% of the settling rate at
the settling of suspended matter and conse- 30 C. It is for this reason that greater effi-
quent reduction in turbidity. Subsidence is ciency in turbidity removal by subsidence
of value chiefly in the removal of coarse may be obtained during the summer months
particles which settle rapidly and which can when water temperatures are higher. During
be more economically removed in this fashion the winter months, with normally lower
than by coagulation. A secondary result of water temperatures, the rate of settling will
su~sidence is removal of bacteria. Percentage be reduced and lower efficiency results or
bacteria removal generally closely parallels else it may be necessary to increase the time
turbidity removal. While dependent on many permitted for subsidence.
factors, turbidity and bacteria removal by Settling velocities may be calculated from
subsidence may vary from 30% to 80%. Stoke's Law.
y = 18.5 D2 (51- S2)
All particles heavier than water possess a
tendency to settle due to the influence of Where:
gravity. The time required for settling is V =Velocity of fall in feet per second.
dependent on a number of factors such as D = Diameter in inches of particle.
the weight, shape and size of the particle to-
S1 = Density of particle in pounds per
gether with the viscosity and frictional re-
cubic foot.
sistance of the water.
S2 = Density of fluid in pounds per cubic
To illustrate the effect of particle size, the
following tabulation is presented:
Z = Viscosity in centipoises.
In employing this equation it is assumed
that the particles are spherical, falling under
lOG (50F) viscous resistance in an infinitely large body
Specific Gravity = 2.65 of fluid.
The use of Stoke's Law and tabulation of
Diameter of
Particle, Order of Time Required settling velocity is based on still water. Theo-
lnmm Magnitude to Settle 1 Foot retically, velocity of flow in a horizontal
plane will not affect settling velocity of the
10.0 Gravel 0.3 seconds
1.0 Coarse Sand 3 seconds
suspended particles because their downward
0.1 Fine Sand 38 seconds movement is caused by gravity. However,
0.01 Silt 33 minutes velocity of flow in a vertical plane and con-
0.001 Bacteria 35 hours
0.0001 Clay Particles 230 days vection currents directly tend to prevent
0.00001 Colloidal Particles 63 years settling. Design of subsidence basins is there-
fore such as to keep such interferences to a
The specific gravity of suspended particles minimum and to assure continued quiescence
is one of the main factors governing the rate of the water.
of settling. Particles of high specific gravity A study of the tabulation of settling veloc-
will naturally settle at a faster rate than ities will indicate that for industrial pur-
those of a specific gravity only slightly poses the practical limit on solids removal by
greater than that of water. subsidence is for particles of a diameter of

0.01 mm or more. Smaller particles have It is necessary in basin design to make

such a low rate of settling that the time re- allowance for the sludge accumulated be-
quired is greater than can be practically al- tween cleanings. Twenty-five per cent of the
lowed. In large storage reservoirs having a tank volume is usually provided for this
retention period of weeks and months, it is purpose. When proper settling is no longer
possible to effect removals of particles of obtained, due, to sludge accumulation, the
less than 0.01 mm diameter. tank is drained and flushed.
In the larger tanks and particularly where
EQUIPMENT EMPLOYED there is a considerable amount of sludge,
mechanical means are provided frequently
Subsidence may be employed by using a for continuous sludge removal. Mechanically
simple tank which is filled with the turbid operated rakes and scrapers travel just above
water and allowed to stand. After the sedi- the floor and collect sludge to the draw off
ment has settled, the clarified water may be point for continuous removal. The design
decanted off and the settled solids removed and speed of operation of the scrapers must
from the bottom. The tank is then ready to be such that a minimum of disturbance is
be refilled for another cycle. This method is presented to the desired settling of the sus-
designated as a "fill and draw" system, and pended solids.
the settling rates will correspond to those
Continuous methods of subsidence are in
greater use in plants of recent design. This Waters containing turbidity due to relatively
is accomplished by passing the water through large size or high density suspended solids
a basin or tank of such size, shape and de- may be clarified to a suostantial degree by
sign that the velocity of the water is reduced subsidence. This applies to natural waters,
to a point sufficiently low to permit settling sewage and industrial trade wastes. Subsi-
of the readily settleable solids. Clarified dence is an economical method of obtaining
water is removed from the tank continuously clarification where it is applicable and in
from an overflow, the settled solids remain- such cases is employed prior to coagulation.
ing in the bottom of the tank. Suspended solids of low specific gravity or
In smaller plants of less than 1 million gal- those of fine particle size cannot be removed
lons per day, usually only one settling basin by subsidence within a reasonable period of
is employed. In larger plants, two or three time. Increase in basin size or retention
basins in series are used. The use of three period may not be feasible. In such cases,
basins is favored to permit the removal of chemical coagulation is necessary to increase
one basin from service for cleaning, while the size of the particles and render them more
two basins remain on line. Where continuous easily settleable.
sludge removal is used, one basin is adequate
even in plants of greater capacity. REFERENCES
Basins are usually rectangular and if only
one basin is employed, the use of a square H. E. Babbitt and J. J. Doland, "Water Supply
Engineering", 3rd Ed., pp. 499-509, McGraw-Hill
layout provides the most economical con- Book Co., Inc., New York, N.Y. (1939)
struction. However, in many cases the basin M. N. Baker, "The Quest for Pure Water", pp.
dimensions must be made to fit the space 286-298, Am. Water Works Assoc., New York,
N.Y. (1948)
available. Sloping bottoms are provided to M. L. Riehl, "Hoover's Water Supply and Treat-
facilitate sludge removal by flushing with ment", 8th Ed., pp. 37-39, National Lime Assoc.,
high pressure jets. Depth of settling basins Washington, D.C. (1957)
will vary usually over the range of 12 to 20 "Water Quality and Treatment", 2nd Ed., pp. 131-
162, Am. Water Works Assoc., New York, N.Y.
feet with an average of 16 feet. ( 1950)


Theory of Coagulation . .................... 27

Types of Equipment . . . . . . . . . . . . . . . . . . . . . . . 28
Coagulants. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
Alum ........ ·........................ 30
Sodium Aluminate . . . . . . . . . . . . . . . . . . . . . . . 30
Copperas. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
Ferric Coagulants. . . . . . . . . . . . . . . . . . . . . . . 31
Magnesium Oxide . ...................... 31
Coagulant Aids . . . . . . . . . . . . . . . . . . . . . . . . . . . 33

OAGULATION is that process whereby finely particles into larger masses, the suspended
C divided particles of turbidity and color,
capable of remaining in suspension indefi-
turbidity particles, bacteria, microorganisms,
etc., become enmeshed by the floc. Any sus-
nitely are combined by chemical means into pended particles carrying a negative charge
mass:s sufficiently large to effect rapid are more easily enmeshed by the floc and it is
settling. believed that color is removed in this fashion.
Ordinary subsidence usually will not pro- For each coagulant there is an optimum
duce a water sufficiently low in turbidity to pH zone for best coagulation. It is desirable
be suitable for domestic purposes or for most to control the pH to the point of maximum
industrial requirements. Additional settling precipitation of the coagulant, equivalent to
time is not usually of benefit because of the the minimum solubility of the floc. The best
small size of the suspended particles and their pH for this precipitation is dependent not
slow settling rate. Filtration, without coagu- only on the coagulant, however, but also upon
lation, will not remove the fine particles of the mineral characteristics of the water. For
turbidity if a filter medium sufficiently coarse example, while alum will coagulate best at a
for modern filtration rates is employed. Coagu- pH of approximately 5.5 for pure distilled
lation is required to agglomerate the sus- water, for most natural waters the ideal pH
pended particles, thus making the water more is higher and may range from pH 5.5 to 8.0.
readily filtered and also causing much of the The temperature of the water also affects
coagulated matter to settle out prior to filtra- the process of coagulation and at tempera-
tion. tures near the freezing point increased mixing
time as well as increased coagulant doses may
In general, successful coagulation involves
Coagulation reactions can be written very three main factors:
simply showing the use of aluminum or iron
1. The presence of a minimum quantity of
salts precipitating to form a floc of aluminum
aluminum or iron ions to form an in-
hydroxide or iron hydroxide. While these re-
soluble floc.
actions will be shown for the sake of simplic-
2. The presence of a strong anion such as
ity, coagulation is a much more complex
sulfate or chloride.
phenomenon than thus illustrated.
3. The pH of the water must be controlled
It is thought that the precipitation of an within a definite range.
aluminum or iron coagulant produces finely
It is not possible to predict from the analy-
divided precipitates of the hydrous oxide.
sis of a water which will be the best coagulant
These precipitates are positively charged and
to use, the quantities that will be required
thus repel each other. Neutralization of the
or the optimum pH for control of the process.
positive charges by negative ions such as sul-
For most efficient and economical coagula-
fate and -chloride in the water causes coales-
tion it is necessary that laboratory tests be
cence of the fine particles, thus forming a
made with different coagulants at varied con-
gelatinous porous precipitate of large volume.
centrations and varied pH levels. Laboratory
Flocculation, as practiced in water treat- stirring devices should be used in these tests
ment, is the further increase in size and volume to provide uniform test conditions. While
of the precipitate caused by bringing together previous practical experience with a given
by gentle agitation of the water a number of water will aid in selecting coagulant dosage
t~e individual precipitated hydrous oxide par- and optimum pH, in order to consistently
ticles together with enmeshed turbidity to secure the desired results it is necessary to
form larger and more readily settleable change coagulant dosage and pH control to
compensate for varying raw water character-
In the flocculation of the hydrous oxide istics.

Coagulation problems vary greatly not only water of the most desirable characteristics.
because of varying raw water characteristics, The presence of iron and manganese in the
but also because of the ultimate use of the raw supply introduces additional factors af-
water. For potable systems the removar of fecting coagulant choice and dosage. Indus-
bacteria and microorganisms is quite impor- trial wastes present in the raw water also may
tant and these requirements may affect coag- complicate the coagulation problem.
ulation control. For example, a slight excess of
coagulant may be used over and above that
necessary to produce a relatively clear water TYPES OF EQUIPMENT
in order that bacteria removal may be in-
creased in the sedimentation basin, thus re- Three essential steps in the proper control of
ducing the residual chlorine concentration re- coagulation processes prior to filtration are:
quired in the finished water and minimizing 1-Mixing
taste and odor problems. These considera- 2-Flocculation
tions would not be important in the prepara- 3-Sedimentation
tion of boiler feedwater where clarity of the It is important that the coagulant be
filter effluent at minimum coagulant dosage mixed rapidly with the water to be treated in
would be the controlling factor. In boiler feed- order to secure efficient coagulation at mini-
water conditioning, however, softening is usu- mum chemical dosage. By rapid or "flash"
ally important and coagulation may be com- mixing the coagulant is uniformly distributed
bined with softening to achieve a finished throughout the water before precipitation is

Courtesy Graver Water Conditioning Co.

Figure 5-1 • Sludge Contact Clarifier



Courtesy lnfilco, Incorporated

Figure 5-2 • Upjfow Clarifier

completed. Mixing is secured usually by six hours are usually provided. It is important
motor-driven propellers placed in small basins to so design the basins as to avoid short-
where the coagulant is introduced or by add- circuiting and to keep the velocity of flow
ing the coagulant at some point of turbulence. quite low. Modern basins are usually designed
Following mixing, the next step is to secure for continuous mechanical desludging rather
flocculation by bringing the small particles than the periodic draining and manual de-
together under the influence of gentle mix- sludging employed in older designs.
ing, thus permitting them to increase in size Upflow clarifiers which combine in one unit
and therefore to settle more rapidly in the the functions of mixing, flocculation and sedi-
sedimentation basin. Flocculation may be mentation have been installed in large num-
achieved by the use of baffied basins where bers in recent years. Particularly in industrial
change in the direction of flow will produce practice, these units are favored because of
the desired agitation. Lack of flexibility is in- the marked reduction in space requirements
herent in this method since the agitation se- and installation cost. These units are designed
cured will vary with the rate of flow. Mechani- for only one or two hours retention in com-
cal drive paddles or mixers are generally pre- parison with two to six hours for conventional
ferred for this purpose since their speed can sedimentation basins.
be controlled to conform with the rate of flow Figure 5-2 illustrates one design of an up-
and floc characteristics. With proper floccula- flow clarifier. In this unit the coagulant pre-
tion a uniform dense floc can usually be pro- cipitates in the presence of a recirculated
duced. slurry of previously formed precipitates. Floc
After mixing and floc formation, the treated formation is encouraged by this slurry re-
water is passed to a sedimentation basin which circulation and finely divided precipitates are
provides the retention time for the floc par- avoided. In the secondary mixing and reac-
ticles to settle. Retention periods of two to tion zone the raw water is circulated in con-

tact with three to five volumes of recirculated The use of sodium aluminate is advantageous
slurry, aiding in bringing treatment reactions under certain conditions because there is no
to equilibrium. Clarified water separates from need for additional alkali ..
the top of the slurry pool and sludge is auto- Sodium aluminate may be used in conjunc-
matically blown from the concentrator com- tion with alum, particularly for color removal.
partments. Units of this type provide efficient Alum alone, or in conjunction with acid, may
clarification of water in a single compact be used to coagulate color at a relatively low
unit. pH with sodium aluminate added after prim-
ary color removal for pH adjustment and
COAGULANTS secondary coagulation.
Sodium aluminate is employed frequently
ALUM Al2(S0 4 )a.l8H 2 0. The most widely
as a coagulant in conjunction with both hot
used coagulant is aluminum sulfate or filter
and cold lime-soda softening. In addition to
alum. Commercial filter alum contains
providing a lower turbidity of the softener
slightly less water of crystallization than in-
effluent, sodium aluminate assists in more
dicated by the above formula and corres-
complete precipitation of magnesium, thus
ponds to Al 2 (S04 ) 3 .14H20. This material
reducing the magnesium content of the sof-
is specified on the basis of 17% Al 20 3 con-
tened water. This effect on magnesium re-
tent. Alum is applicable as a coagulant over
duction is obtained to a greater extent in cold
a pH range of approximately 5.5 to 8.0
process softening, but is also of value in hot
and reacts with either natural or added al-
process softening.
kalinity to produce a floc of aluminum hy-
droxide as illustrated below. Ammonium alum is not widely used in co-
agulation processes, but is useful in some spe-
Alz(S04h + 3Ca(HC0 3h 2AI(0Hh + cial applications. In lump or crystal form am-
aluminum + calcium = aluminum + monium alum dissolves slowly and is employed
sulfate bicarbonate hydroxide
where alum pots are used for coagulant feed
3CaS0 4 + 6C0 2
in small installations as in swimming pool
calcium + carbon
sulfate dioxide practice. Truly proportional feeding is not
achieved. On the basis of aluminum content,
Alz(S04h + 3NazCOl + 3Hz0
ammonium alum costs approximately three
aluminum + sodium + water
times as much as filter alum. The ammonia
sulfate carbonate
content may be of value where chlorine-am-
2AI(0Hh + 3NazS04 + 3COz
monia treatment is used by eliminating the
aluminum + sodium + carbon
hydroxide sulfate dioxide need of a separate source of ammonia.
A(z(S04h + 3Ca(0Hh 2AI(0Hh + Activated alum is a commercial brand of
aluminum + calcium = aluminum + alum containing insoluble silicates to aid in
sulfate hydroxide hydroxide the coagulation and to produce a tougher floc.
3CaS04 This material contains less water of crystalli-
calcium zation than the usual filter alum. Black alum
sulfate is a commercial filter alum containing pow-
SomuM ALUMINATE Na2Al 2 0 4 • Sodium alu- dered activated carbon. Black alum is em-
minate is actually aluminum oxide stabilized ployed where it is desired to obtain both co-
with caustic soda. A good commercial grade agulation and a degree of taste and odor
of sodium aluminate contains approximately removal with a single material.
55% aluminum oxide, 35% combined so- CoPPERAS (Ferrous Sulfate-FeS04 .7H2 0).
dium hydroxide and 6% excess sodium hy- The natural alkalinity of most waters is
droxide. This material may be used in place not sufficient to react with copperas to form
of an alkali in water which must be coagu- the desired ferric hydroxide floc. Lime or an-
lated with aluminum in a higher pH zone. other alkal.i must be added to produce this

result. The first reaction with lime produces as ferric chloride. Reactions with ferric sul-
ferrous hydroxide which in turn is oxidized fate can be shown as follows:
by the dissolved oxygen in the water to form
ferric hydroxide. This oxidation occurs only Fez(S04h + 3Ca(HC0llz = 2Fe(0Hh +
at higher pH values and copperas is a suitable ferric + calcium = ferric +
sulfate bicarbonate hydroxide
coagulant only in the alkaline pH range of
8.5 to 11.0.
3Ca504 + 6COz
calcium + carbon
FeS04 + Ca(OH)z = Fe(OH)z + sulfate dioxide
ferrous + calcium ferrous + Fez(S04h + 3Ca(OH)z = 2Fe(OHh +
sulfate hydroxide hydroxide
ferric + calcium = ferric +
sulfate hydroxide hydroxide
sulfate 3Ca504
4Fe(0Hlz + Oz + 2Hz0 = 4Fe(0Hh sulfate
ferrous + oxygen + water = ferric
hydroxide hydroxide In general, the ferric coagulants will co-
An advantage of both ferrous and ferric agulate at much lower pH values than the
coagulants in comparison with aluminum co- aluminum salts and can be employed over
agulants is that the floc produced will not the wide pH range of 4.0-11.0. Ferric coagu-
redissolve at higher alkalinities. However, the lants have been employed successfully in color
need for coagulating at a higher pH and the removal at low pH values. At high pH values,
consequent use of larger amounts of alkalies ferric coagulants are also useful in combining
restricts the general use of this material for color removal with oxidation and precipita-
coagulation purposes except in conjunction tion of iron and manganese. The ferric hy-
with lime softening. droxide floc is heavier than alum and settles
FERRIC CoAGULANTS. The ferric coagulants, more rapidly. Chlorinated copperas is also
ferric sulfate-Fe2 (S0 4 ) 3 and ferric chloride classed as a ferric coagulant. Chlorine will
-FeCl3 , act to precipitate ferric hydroxide oxidize copperas and form a mixture of ferric
as is the case with copperas. However, since chloride and ferric sulfate.
the iron is already in the ferric state, dis- 6Fe(S04) + 3C(z = 2Fez(S04h +
solved oxygen is not required for oxidation. ferrous + chlorine ferric +
Ferric chloride is available either in the sulfate sulfate
anhydrous form, or as FeCl3 .6H2 0 or as a 2FeCh
water solution containing 35-40% FeCl3 • Fer- ferric
ric chloride is quite corrosive in the presence chloride
of moisture. Reactions with natural or added In practice, the copperas solution is mixed
alkalinity can be illustrated as follows: with the chlorinator discharge prior to the
2FeCh + 3Ca(HC03)z = 2Fe(OHh + entry of the mixed solution into the coagula-
ferric + calcium ferric + tion system. Theoretical chlorine require-
chloride bicarbonate hydroxide ments are 1 part of chlorine for each 7.8 parts
3CaCI 2 + 6C0 2 of copperas. Since this reaction produces both
calcium + carbon ferric chloride and ferric sulfate, chlorinated
chloride dioxide copperas has the same field of usefulness as
2FeCI3 + 3Ca(0Hh = 2Fe(0Hh + 3CaC(z these materials. Because of the comparative
ferric + calcium = ferric + calcium ease of feeding copperas as compared with
chloride hydroxide hydroxide chloride the ferric coagulants, chlorinated copperas is
Ferric sulfate is available as the anhydrous sometimes preferred.
material and as Fe 2 (S0 4 )a.9H 2 0. Ferric sul- MAGNESIUM OxiDE (MgO). Activated mag-
fate is corrosive, but not to the same extent nesium oxide was developed as an efficient

Fig. 5-3-Data on Coagulants and Coagulant Aids

Commercial Grades Weight, Handling
Name Formula Strength Available lb/cu ft Materials Miscellaneous
Aluminum Lump Powder, 38-45 Lead pH (1 "/o
Sulfate Powder Other, 57-67 Rubber solution)
Granules Silicon Iron = 3.4

Sodium Crystals 60 Iron Stabilized

Aluminate Steel with approx.
Rubber 6"/o excess
Plastics NaOH
Ammonium AI2(S04h· Lump 60-68 Lead pH (1%
Alum (NH 4 hS04 .24H 20 Powder Rubber solution I
Silicon Iron = 3.5
Copperas 55"/o FeS04 Crystals 63-68 Lead Efflorescent
Granules Tin
Ferric 90% Fe 2(S0 4 h Powder 60-70 Lead Hygroscopic
Sulfate Granules Rubber
Stainless Steel
Ferric FeCI 3 .6H 20 60"/o FeCI 3 Crystals 45-55 Rubber Hygroscopic
Chloride Glass
Magnesium MgO 95"/o MgO Powder 25-35 Iron Essentially
Oxide Steel insoluble,
fed in slurry
Bentonite Powder 60 Iron Essentially
Steel insoluble,
fed in slurry
Sodium Na 20.3.22Si0 2 41° Be Solution 87 Iron Solid grades
Silicate Steel are available
Rubber with varied
Na 20-Si0 2

means for the removal of soluble silica from unique and useful advantage 1s that no
water. Magnesium oxide is also an efficient soluble salts are added to the water with its
coagulant in the alkaline pH range of 9.5 use and no lime or soda ash is required
and above. Magnesium oxide provides ex- for its precipitation. The additional advantage
cellent coagulation properties in conjunction of silica removal is secured with this coagul-
with hot and cold process lime-soda soften- lant and unlike aluminum coagulants, no ions
ing and can be considered as a coagulant are introduced into the water which may re-
specific to the lime-soda softening process. A sult in complex boiler scales.

Figure 5-4 • Jar Test Coagulation Studies

COAGULANT AIDS ionic, cationic or nonionic. Polymers whose

functional groups in water solution give posi-
Polyelectrolyte coagulant aids are being used tively charged particles are cationic. Polymers
in increasing quantities in water clarification that dissociate to form negatively charged
processes. The increase in use far exceeds the ions are called anionic. Polymers in the non-
increasing demands for water as a raw ma- ionic group provide both positive and nega-
terial indicating that the synthetic polymers tive charges in solution.
have definitely proven themselves superior in The mechanisms of the aids are not com-
many applications. pletely understood but a great deal of re-
Polyelectrolytes are high molecular weight search is being done in order to develop fur-
water soluble polymers that contain groups ther knowledge of this function. Such factors
capable of undergoing electrolytic dissocia- relating to colloidal behavior as Zeta poten-
tion to give a highly charged, large molecular tial, Brownian movement, van der Waals
weight ion. Originally, the term was applied forces and electrophoretic mobility are being
only to synthetic polymers or copolymers but studied to determine a more scientific ap-
has become more inclusive through general proach to the selection of coagulants and co-
use. The term now includes naturally occur- agulant aids. At the present time the most de-
ring organic flocculents ;many of which func- pendable tool for checking out a system of co-
tion through hydration alone rather than agulants is the well-known jar test as illus-
through electrolytic activity. In some cases, trated in Figure 5-4.
the term is applied to combination of natural The pH of a water to be clarified is usually
organic flocculents and inorganic weighting a very critical factor. In actual practice, there
agents. is generally a very narrow pH range where the
Polyelectrolytes are further classified as an- quickest and best results are obtained with

any particular coagulant. In many cases, the verely tax the effectiveness of normal clarifica-
application of a polyelectrolyte will not only tion processes and it is here that the aids can
broaden the pH range over which satisfactory often be of most value.
flocculation will occur but also reduce the
quantity of primary coagulant required.
In many cases, the use of polyelectrolyte REFERENCES
aids has been justified on the basis of im- A. P. Black, "Basic Mechanisms of Coagulation",
Journal, Am. Water Works Assoc., Vol. 52, pp. 492-
proved quality water rather than any cost sav- 504 (1960)
ing. These polymers may also prove advan- J. M. Cohen, "Improved Jar Test Procedure",
tageous in permitting higher flow rates through Journal, Am. Water Works Assoc., Vol. 49, pp.
existing clarification equipment. This factor 1425-1431 (1957)
J. M. Cohen, G. A. Rourke and R. L. Woodward,
can be very desirable to a company which "Natural and Synthetic Polyelectrolytes as Coagu-
would otherwise be faced with a capital invest- lant Aids", Journal, Am. Water Works Assoc., Vol.
ment for new clarification equipment because 50, pp. 463-478 (1958)
of an increased water demand. Increases in C. R. Cox, "Laboratory Control of Water Purifica-
tion", pp. 163-190, Case-Shepperd-Mann Publishing
throughput up to 100% have been attained Corp., New York, N.Y. (1946)
without sacrificing the quality of the effiuent. W. A. Hardenbergh, "Water Supply and Purifica-
tion", 3rd Ed., pp. 331-367, International Textbook
In many clarification installations, the aids Co., Scranton, Pa. (1952)
are employed when influent conditions are A. A. Hirsch, "Manual for Water Plant Operators",
such that the heaviest demand is placed on pp. 56-76, Chemical Publishing Co., Brooklyn, N.Y.
the efficiency of the unit. During the run-off (1945)
"Water Quality and Treatment", 2nd Ed., pp. 131-
period following storms and during cold 162, Am. Water Works Assoc., New York, N.Y.
weather operations, surface supplies can se- (1950)


Theory Involved. . . . . . . . . . . . . . . . . . . . . . . . . . 36
Types of Filters . ......................... 38
Gravity Filters . ........................ 38
Pressure Filters . . . . . . . . . . . . . . . . . . . . . . . . . 39
Filter Operation . ......................... 40
Applications. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
Limitations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41

is the process of passing a liq- lated particle~ normally does not extend
uid containing suspended matter through
a suitable porous material in such a manner
deeper than ~-4 inches and most of the fil-
tration is secured at the surface or in the
as to effectively remove the suspended matter first one or two inches of the bed. This coagu-
from the liquid. lated mat acts as a fine filter for smaller
Filtration is employed in the treatment of particles. It is necessary in selecting the size
industrial water in order to remove or re- of filter medium to have sufficient coarseness
duce suspended solids and turbidity. The tur- so that some penetration of the top few
bidity or suspended matter may be present inches of the bed takes place. With no pene-
initially in the raw water undergoing treat- tration of the bed by coagulated material,
ment or may be the result of chemical coagu- the loss of head would increase rapidly and
lation or a precipitation process of water filter runs would be short.
treatment. For example, in lime-soda soften- It is desirable from the standpoint of filtra-
ing, calcium and magnesium salts are precip- tion to provide a size of filter medium which
itated in the softening process and final clari- will prevent floc from passing through the
fication by filtration must be done prior to filters, hold the floc as loosely as possible in
use. In general, most water treatment proc- order to permit easy washing and hold as
esses involving coagulation and precipitation large a volume of floc as possible without
are followed by filtration in order to complete clogging. The removal of turbidity is affected
the removal of undesirable impurities from not only by the sand grain size, but by the
the finished water. particle shape. Sharp, angular particles pro-
Filtration as such does not provide removal duce larger voids and do not remove as much
of dissolved solids although .it may be used fine material as rounded particles of the same
in conjunction with a softening process that equivalent diameter. From the standpoint of
does reduce the dissolved solids content of backwashing the filter medium should be of
the water treated. such size that backwashing will free it of ad-
hering floc without it being lost during back-
Sand and anthracite coal are the materials
With proper sedimentation following chemi- usually used as filter media. Bed depths of
cal treatment, the heavier coagulated parti- 15-30 inches are employed depending on the
cles will have been removed prior to filtra- type of filter used.
tion. Only the smaller and lighter particles Quartz sand, silica sand and anthracite
of floc reach the filters. When a freshly back- coal used in most gravity and pressure filters
washed filter is first placed in operation, must meet definite standards, namely effec-
many of the finely coagulated particles pene- tive size and uniformity coefficient. The ef-
trate into the filter bed through the many fective size is that size such that 10% of the
voids in the bed surface. As the particles lodge sand grains by weight are smaller and 90%
between grains of filter medium, flow is re- coarser. Thus, the effective size is the mini-
stricted. Coagulated particles then build up mum size of the bulk of the sand particles.
on the surface of the filter bed. With con- However, this does not indicate either the
tinued filtration, the rate of flow increases coarseness limits nor the degree of variation.
through the larger unclogged passages and To assure that the variation is not too great
decreases through the smaller and partially the second measurement must be made. This
clogged passages. As the flow penetrates the is the size of the_ sand grains of which 60%
bed further, the water spreads out and veloc- by weight of the sand is finer and 40%
ity decreases with lodging of coagulated par- coarser. This size divided by the effective size
ticles again at points of low velocity. is the uniformity coefficient. For example: if
Penetration of the filter medium by coagu- a screen analysis showed 10% of a sand was

finer than 0.40 mm, and 60% was finer than size of from 0.35 to 0.50 mm, as conditions
0.64 mm, the effective size would be 0.40 mm warrant, and a uniformity coefficient of not
and the uniformity coefficient would be 0.64 over 1. 75. Pressure filters are generally pro-
divided by 0.40 or 1.6. vided with filter sand with effective size of
The most desirable effective size and uni- 0.50 to 0.60 mm and the uniformity coeffi-
formity coefficient depends on conditions of cient of 1. 7 maximum. Experiments made at
operation and effluent quality requirements. the Chicago Water Department concluded
Modern practice usually calls for two grades good results can be obtained with coarser
of filter sand; for gravity filters the effective sand, with an effective size range of 0.60 to

Courtesy Roberts Filter Mfg. Co.

Figure 6-1 • Cutaway of Gravity Filter

0. 72 mm, with a uniformity coefficient of moval of accumulated suspended matter is

1.3 to 1.5. This work indicated a 26 inch bed secured by backwashing up through the filter
depth optimum. bed, expanding the bed ··and washing the
Anthracite coal prepared in similar sizes is solids to waste.
equally effective and may be used in place of The essential parts of a gravity filter, in
sand and gravel in filters. Extended filter addition to the filter medium, can be briefly
runs, less backwash water requirements and outlined as follows:
higher filtration rates have been experienced 1. The filter shell itself may be of concrete,
with anthracite coal filters. In many indus- steel or wood and may be square, rectan-
trial filtration applications, the anthracite gular or circular in shape. However, rec-
medium is preferred because no silica is added tangular units of reinforced concrete are
to the water due to high alkalinities and heat. most widely used.
The lower specific gravity permits a lesser 2. The filter medium is supported by a gravel
backwash rate. bed which serves the purpose of prevent-
Other materials such as quartz, calcite and ing the fine sand or anthracite from pass-
magnetite are sometimes used as filter media ing down into the underdrain system and
for special purposes. also acts to distribute the backwash water.
The supporting bed may vary in depth
TYPES OF FILTERS usually from 12 to 24 inches and consists
of several layers varying from 1 ~ inches
Filters are divided into two types-gravjty to Ys inch in diameter.
filters and pressure filters. 3. The underdrain system is provided for two
GRAVITY FILTERS. As the name implies, the purposes-the collection of the filtered
flow of water through these filters is obtained water after passing down through the filter
by gravity. medium and supporting gravel and also
"Slow sand filters" or "English" filters are for the distribution of backwash water to
mentioned in passing because of their his- the underside of the filter medium. It is
torical significance. These were the first form essential that the underdrain system act
of filtration employed: These filters, now es- uniformly in both operations otherwise
sentially obsolete, use beds of fine sand. filtering and backwashing rates in one
The water is filtered at very low rates, usually part of a filter would exceed those in an-
in the range of 2-10 million gallons per day other and result in inefficient operation.
per acre (0.032-0.160 gpm per sq ft). The The underdrain system may consist of
removal of suspended solids is secured by header and laterals with perforations or
straining through the pores in the surface of strainers suitably spaced. False bottom type
the sand layer and by adherence of the solid underdrain systems are also popular.
particles to the sand grains. Biological action 4. Wash water troughs are provided to col-
in these beds accounts for the good results lect backwash water. These troughs may
obtained. As backwashing of such large areas be of steel, cast iron or concrete with the
is impractical, at intervals the accumulated newer plants employing aluminum, asbes-
growth and suspended matter with some of tos-cement and fiberglass for greater re-
the sand from the top of the bed are removed sistance to corrosion. They must be of
by rakes. adequate size to handle maximum back-
In rapid sand or coal filters, the water is wash rates without flooding. Troughs are
passed downward by gravity at a relatively spaced so that the horizontal travel of
high velocity, usually at rates between 2-4 backwash water will not exceed 3 to 3~
gpm per square foot. Pretreatment by coagu- feet. Wash troughs are placed approxi-
lation and sedimentation is necessary in order mately 24 inches above the top of the filter
to permit these higher rates of filtration. Re- surface although where high backwash

Courtesy, Infilco, Incorporated

Figure 6-2 • Battery of Pressure Filters with Pilot Operated Diaphragm Valves

rates are employed, the freeboard above and manual operations previously required
the filter surface should equal the inches on standard gravity filters. Their application
of vertical rise per minute. to industrial water filtration problems is
5. Certain control devices are required to mcreasing.
insure maximum efficiency in filter opera- PRESSURE FILTERS. Pressure filters are some-
tion. Rate-of-flow controllers automatically what more widely used than gravity filters in
maintain uniform delivery of filtered industrial water conditioning. Pressure filters
water, operating from Venturi tubes in possess the advantage that they may be placed
the effluent line. Backwash rate-of-flow in the line under pressure and thus eliminate
controllers are used similarly to provide double pumping. The use of pressure filters is
proper wash conditions. Rate-of-flow and required in conjunction with hot process sof-
loss-of-head gages are also considered nec- tening to permit high temperature -operation
essary for most efficient operation. It is and to avoid thermal loss. The general de-
possible to group most gages, together sign of pressure filters is essentially the same
with the operating controls, on operating as gravity filters with respect to filter medi-
tables for convenience for either manual um, supporting bed, underdrain system and
or automatic operation. control devices. The filter shell, of course,
New concepts in automatic gravity filtra- differs from a gravity filter and wash water
tion are offered by several manufacturers troughs are not employed.
that reduce the need for various controllers Pressure filters may be of the vertical or

horizontal type. The filter shells are steel, and also permits a small amount of suspended
cylindrical in shape, with dished heads. Ver- matter to accumulate on the bed. As soon as
tical filters range in diameter from 1 to 10 feet the filter produces clear water, the unit is
with capacities from 2.4 gpm to 235 gpm returned to service.
at a filter rate of 3 gal per sq ft per minute. During operation, the suspended matter re-
Horizontal filters, usually 8 feet in diameter, moved by the filter accumulates on the surface
may be 10 to 25 feet long with capacities from and the increased frictional resistance is shown
218 gpm to 570 gpm. The general trend is by the loss-of-head gage. When loss-of-head
away from horizontal pressure filters because reaches 5 psi, it is advisable to backwash the
certain areas of the filter bed adjacent to the filter to remove the mat on the bed surface.
shell are inactive during filtration and back- While industrial filtration rates are usually 3
washing. One vendor has redesigned the hori- gal per sq ft of filter area per minute, backwash
zontal unit with cells thereby improving rates are much higher in order to lift the sus-
operations. pended matter from the surface and expand
Figure 6-3 the filter bed. Backwash rates of 12-15 gals
per sq ft per minute are employed for sand
Filter and Backwash Rates for Vertical
and rates of 8-12 gals per sq ft per minute for
Pressure Filters
anthracite. The backwashing is continued for
Diameter, Area Filter Rate• Backwash Ratet 5 to 10 minutes, then the filter is returned to
ft Sq Ft gpm gpm
service, following filtering to waste until the
3 7.1 21 106 water is clear.
4 12.6 38 189 Surface washers are of assistance in main-
5 19.6 59 295
6 28.3 85 425 taining maximum filter efficiency. These units
7 38.5 116 578 consist of horizontal arms, equipped with noz-
8 50.3 150 755
9 63.6 190 954
zles, supported just above the surface of the
10 78.5 235 1180 filter bed. High pressure backwash water sup-
plied to the washer causes rotation of the
* Filter rate of 3 gal per sq fl per minute arms and water leaving the nozzles impinges
t Backwash rate of 15 gal per sq fl per minute
on the media, and loosens the accumulated
Figure 6-4
suspended matter.
Typical Filter Bed layers and Sizes For
Pressure Filters APPLICATIONS

SAND ..... 12" layer-sand, 0.45-0.50 mm effective size Filtration is a requirement in the clarification
10" layer-sand, 0.80-1.20 mm effective size of turbid waters for most industrial uses. In
4" layer-gravel, lf.I"·Ys"
4" layer-gravel, %"·1!.1"
the preparation of domestic supplies, process
8" layer-gravel, 1"·%" water, boiler feedwater and, in fact, for almost
4" layer-gravel, 1%"·1" every use it is necessary that the water be
ANTHRAFILT 18" layer-No. 1 Anthrafllt, 0.6-0.8 mm freed of objectionable suspended solids. While
9" layer-No. 2 Anthrafllf, 3/32"-3/16"
9" layer-No. 4 Anthrafllt, 5/16"-9/16" both coagulation and sedimentation may be
4" layer-No. 6 Anthraflll, 13/16"·1 %" employed for the removal of a portion of the
suspended matter, filtration normally also is
required to complete the production of a water
of satisfactory quality.
After backwashing, and placing a filter in Filtration is also employed as a part of other
operation, the water is filtered to waste for the water treatment processes. For example, all
first few minutes. This procedure rids the of the precipitation methods of softening em-
system of possible suspended solids remaining ploy filtration as a final step in the process.
in the underdrain system after backwashing Filtration is employed following hot phosphate

softening as well as both hot and cold lime solids as in gravity filters. Pressure filters are
and lime-soda softening, except in some special more difficult to inspect and clean. The re-
cases. placement of filter medium, gravel and under-
Filtration ahead of exchange softeners and drain systems presents more of a problem.
demineralizers is standard practice. Filtration However, the most pertinent criticism of
is also a final step in certain iron removal pressure filters in comparison with gravity
methods and in the removal of oil from con- filters is the inability of the operator to ob-
densate. serve the effectiveness of the filtering and
backwashing operations. While pressure filters
are capable of providing an effluent of as
high quality as gravity filters, this inability
Filtration, without coagulation and sedimen- to maintain constant observation of the proc-
tation, is usually an unsatisfactory method for ess may result in the production of an effluent
completely removing turbidity and suspended of poorer quality than is obtained with gravity
solids. In such an application, the filter pro- filters. Consequently, gravity filters are pre-
vides only the removal of the larger particles. ferred where the conditions of the installation
Pressure filters are favored where double permit their use and where a water of highest
pumping can be avoided by placing the filter quality is desired. Any malfunctioning of a
under line pressure. Pressure filters also usually gravity filter may be quickly observed and
require less space for installation. Pressure corrected. The selection between the two types
filters are required with hot process operation must be made after a study of the conditions
and prevent thermal loss in· filtration and and problems involved in the particular instal-
backwashing. Oxygen pickup is also prevented lation.
by the use of pressure filters. REFERENCES
Disadvantages in pressure filtration are that
the appearance of the filter effluent and the G. D. Dickey, "Filtration", Reinhold Publishing
Corp., New York, N.Y. (1961)
filter medium are not under observation, back- "Filtration", The Public Works Manual and Cata-
washing effectiveness cannot be observed and log File, Public Works Publications, Ridgewood,
loss of filter medium cannot be noted. Wash N.J. (1961)
water troughs cannot be provided in pressure E. Nordell, "Water Treatment for Industrial and
Other Uses", 2nd Ed., pp. 363-387, Reinhold
filters for as effective removal of backwash Publishing Corp., New York, N.Y. (1961)

Precoat Filtration

Equipment Employed. . . . . . . . . . . . . . . . . . . . . . . 43
Theory Involved . ......................... 43
Filter Operation . . . . . . . . . . . . . . . . . . . . . . . . . . 44
Applications. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
Limitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46

IATOMA_CEous earth filtration is a proc:ss

D in whrch a filter cake or precoat of dia-
tomaceous earth is used as a filter medium.
The precoat of diatomaceous earth is formed
over a permeable base or septum, by adding
a slurry of diatomite to the filter shell and
either filtering to waste or recirculating the
slurry until the precoat is formed. The unit
is then ready for the filtering operation.
Diatomaceous earth filtration is a special
form of filtration that is employed basically
where the utmost clarity of the effluent is re-
quired or where space and weight limitations
are quite stringent.
Diatomite filtration achieved prominence
during World War II, primarily because of
the need of the armed forces for filter units for
potable water that would conserve space and
weight, which would be portable and rela-
tively simple to operate. During the war, these
filter units were also applied to the removal
of oil from condensate on shipboard, where
space and weight requirements must also be
kept at a minimum.

Courtesy Johns Manville Products Corp.
The principle of diatomaceous earth filtration Figure 7-1 • Photomicrographs (X100) of Fine,
is relatively simple. A permeable base or sep- Medium and Coarse Particle Size Filter Aids
tum of some type is required to support the
diatomite filter cake. The septum must be of THEORY INVOLVED
such a nature that it will prevent the passage
of the filter aid, must be permeable to the Chemically, diatomaceous earth is silica
flow of filtered water and must be of sufficient (Si0 2 ). It is produced from mineral deposits
strength to withstand the pressure drop. formed by diatoms (a form of plant life re-
Among the materials used as a supporting lated to algae) . Diatomaceous earth is really
base are filter cloths, porous stone tubes, wire composed of billions of tiny fossils, resulting
screens, wire-wound tubes, and porous paper from the silica "skeletons" left by diatoms in
filter pads. This supporting medium is first prehistoric times. Diatomaceous earth deposits
coated with a slurry of diatomaceous earth have been found as much as 1400 feet in
which then acts as the filter medium. The thickness.
water is filtered in passing through the diato- The structure of diatomaceous earth is due
mite coating. During the filter run, additional to the shape of the tiny plants from which it
diatomite slurry is usually added. When a originated. These plants, or diatoms, are so
high back pressure drop develops, due to the small that they must be examined under a
accumulation of suspended matter removed in microscope. Individual figures are of many
filtration, the filter coating is sloughed off by shapes such as flowers, discs and needles. Over
backwash, and the filter returned to service 10,000 types have been identified. Commer-
after again precoating. Coagulants, with at- cial filter aids are produced from the crude
tendant pH control, are not required. material by a gentle milling process which

separates the diatoms from one another. The shell is usually of steel. Diatomite for the pre-
finished product is a fine powder, consisting coat is added as a slurry and the raw water
of rigid particles so tiny that hundreds could pumped through the filter to waste or recircu-
· be placed on a pinpoint. lated until the diatomite evenly precoats the
. When a precoat of diatomaceous earth is filter elements. Only a few minutes are re-
formed on the septum, the irregular, but sym- quired for this operation and the filter is then
metrical, shape of each diatom forms an in- ready for service.
compressible filter medium consisting of ap- As suspended matter is removed during the
proximately 10% solids and 90% voids. In filtering cycle, the pressure drop across the
effect, a screen is formed with extremely tiny filter element increases and rate of flow de-
openings-so fine, in fact, that even most creases. The filter is then backwashed by re-
types of bacteria are strained out. The object versing the flow through the unit, removing
·of the septum is to provide support for the the filter cake from the elements by the re-
precoat. The openings in the septum are not verse flow of water. Spent diatomite is drained
sufficiently small in size to prevent the passage from the filter shell.
of individual diatomite particles. Instead, the
One system uses a variation of this proce-
septum supports the mat or filter cake formed
dure in that the elements are backwashed
· by the interlacing of thousands of diatomite
· particles over each opening. Some diatomite from filtered water storage to waste, ending
particles pass through the septum during the with the unit filled with filtered water.
precoat operation, but once formation of the Other designs may employ compressed air
precoat is complete, the interlocked mass of to remove the filter cake from the elements
diatomite particles prevents further passage. at conclusion of a filter run. In another
Commercial diatomite filter-aids are pro- method, the air trapped and compressed in the
duced in several different grades offering a filter dome is utilized to remove the filter cake
range of filtration rates and clarity of filter from the elements, in conjunction with back-
effluents. In addition to applications for water wash by filtered water.
filtration, diatomaceous earth filter-aids are Leaf type filters, which are employed fre-
employed widely in the filtration of process quently in the removal of oil from condensate,
liquids. employ no backwash water. At the conclusion
While diatomite is the filter aid primarily of the filter run, the unit is drained and the
employed with filters of this type, there are filter cake is manually peeled off the filter
also available various grades of purified wood leaves.
cellulose which find application in the filtra- In some operations, the precoat is the only
tion of high pH solutions where silica pickup filter aid employed. More frequently, however,
is to be avoided. Other media used include additional diatomite is continuously added
perlite, asbestos and carbon. to the raw water during the entire filter run.
It has been found that where a precoat only
FILTER OPERATION is employed relatively short filter runs will
be experienced. Suspended matter will tend
The primary clifference among the several to plug the openings in the precoat, acting to
different types of precoat filters lies in the con- form an impervious mat. Pressure drop in-
struction of the filter element. These con- creases rapidly resulting in short filter runs. By
struction differences affect the manner of filter adding additional diatomite continuously dur-
operation. ing filtration, the filter cake is maintained in
In filters where wire-wound tubes are em- a more porous condition. Lesser pressure drop
ployed, these are usually mounted vertically and longer filter runs result. The diatomite
on a removable plate and inserted as a unit added during filtration is referred to as body
in the filter shell. In the larger sizes, the filter feed to distinguish it from the initial precoat.

Because diatomaceous earth is an abrasive

material, some special designs of feeding ap-
paratus have been _de~elofed. The specific
gravity of the matenal IS high and the mate-
rial will settle out of suspension and tend to
pack at points of low velocity. Feeding equip-
ment must be designed with these considera-
tions in mind.
The amount of diatomite required to form
the precoat and also the amount necessary for
body feed will vary widely. Factors influeJ;lC-
ing the diatomite requirements include the
grade of media employed, the characteristics
of the influent water and the design of the
filter. However, experience has indicated pre-
. coat requirements in the ;range of 1.0-2.0
ounces per square foot of filter area. Body
dosages are directly influenced by influent
turbidity and filtration rates and consequently
are subject to wide variation. Values varying
from 5-50 ounces per thousand gallons of
filter effluent have been reported. Courtesy Infilco, Incorporated

Figure 7-2 • Industrial Type of Diatomaceous

APPLICATIONS Earth Filter Unit

Because of the savings in both space and Health Service Standards, even without chlo-
weight of diatomite filters compared with con- rination. Diatomite filters will remove cysts of
ventional sand filters, diatomite filters are amoebic dysentery and the cercariae of schis-
obviously favored for mobile use by the armed tosomia (blood flukes) , prevalent in Philip-
forces. Figure 7-3 illustrates the comparison pine waters. Consequently, the armed forces
of the U. S. Army pack and mobile units with favor diatomite filters for mobile purification
conventional pressure and sand filters. units.
Figure 7-3 Diatomite filtration is also applied to cer-
U. S. Army Pack and Mobile Units tain process industries where rigid water re-
Compared with Pressure Sand Filters quirements exist with respect to clarity and
bacteriological quality. In such applications,
Sand Filter
Mobile Pressure
Unit Sand Filter
the diatomite filter is frequently applied as a
"polishing" filter after conventional coagula-
Rated Capacity, gpm 15 15 50 50 tion, sedimentation and sand filtration.
Filter area, sq fl 3.6 7.1 10 19.6
Weight, lbs 30 2860 350 8050 In the removal of oil .from contaminated
Diameter, in 8 36 18 60 condensate, diatomite filters possess an ad-
Over-all Height, in 22 73 30 82 vantage over conventional coagulation and
Floor Space, in lOx 10 37 x49 24 X 24 6] X 78
filtration methods. No control of coagulant
and alkali feed is required since neither co-
Due to the extremely fine openings in a agulant nor pH adjustment is employed with
diatomite filter cake, a greater clarity of ef- diatomite filters. In addition, no increase re-
fluent can be obtained. Army tests showed sults in the solids content of the filtered con-
that diatomite filters can be expected to pro- densate, due to the coagulant. ·
vide filtered water approximating U. S. Public Diatomite filtration is being widely applied

to swimming pools where advantages are ex- Continuity of operation is generally con-
perienced in the somewhat simpler operating sidered a requirement for diatomite filters.
control and lesser need for coagulating chem- If the filters are operated intermittently, the
icals. filter cake will slough off when flow is stopped.
To avoid a turbid effluent, when flow is re-
LIMITATIONS sumed, it is necessary to again precoat the
filter. Intermittent operation is unsuitable to
The major factors limiting the use of dia-
diatomaceous earth filtration. In boiler feed-
tomite filtration are the need for a relatively
water conditioning particularly, it is necessary
low turbidity influent water, and the high
to guard against careless operation which per-
operating costs compared with conventional
sand filtration. mits high turbidity in the filter effluent from
disturbance of the precoat. The silica thus in-
In general, it is not advisable to employ
troduced would tend to be solubilized under
diatomite filters where raw water turbidities
exceed 200 ppm. Uneconomically short filter boiler temperature and pH conditions, pos-
runs result from high turbidity influent waters. sibly resulting in the formation of siliceous
boiler scales. For these applications, ,nonsili-
The purchase price of diatomite filters may
ceous media must be employed.
exceed the cost of pressure sand filters al-
though installation cost may be less because Some types of filter elements have shown
of the smaller space requirements. The operat- susceptibility to clogging with iron, manga-
ing cost of diatomite filters will generally ex- nese and other materials and require removal
ceed the cost of coagulation and pressure sand of the elements for chemical cleaning.
filtration, due primarily to the cost of the dia-
tomaceous earth.
Diatomite filters operate with a high pres-
sure drop across the unit, as much as 25-50 "Diatomite Filter Aids", Technical Bulletin D10,
psi. Pumping costs are increased and the in- Great Lakes Carbon Corp., Los Angeles, Calif.
stallation of diatomite filters may be limited "The Filtration of Water", Johns-Manville Corp.,
by the available pressure. NewYork,N.Y. (1961)

Chemistry of Chlorination . . . . . . . . . . . . . . . . . . . 48
Theory of Disinfecting Action . . . . . . . . . . . . . . 48
Chlorine Requirements. . . . . . . . . . . . . . . . . . . . 49
Chlorine-Ammonia Process. . . . . . . . . . . . . . . . 50
Breakpoint Process . . . . . . . . . . . . . . . . . . . . . . 50
Chlorine Dioxide. . . . . . . . . . . . . . . . . . . . . . . . . . 51
Effects of Chlorine . ........................ 51
Hypochlorination . ......................... 52
Dechlorination . ........................... 52
Feeding Equipment . . . . . . . . . . . . . . . . . . . . . . . . 53
Laboratory Control. . . . . . . . . . . . . . . . . . . . . . . . 53

HLORINATION is the process in which chlo- tion and oxidation.

C rine gas or chlorine compounds such as
hypochlorites are added to water or sewage,
The addition of chlorine to treated sewage
is widely used for bacterial control and sterili-
usually for the purpose of disinfection. zation. Further, substantial reduction in ap-
The principal use of chlorination is to safe- parent BOD has been experienced where rea-
guard water supplies by killing disease pro- sonably high dosages of chlorine were applied
ducing organisms. For this purpose chlorina- to sewage, either in the partially treated or
tion is standard practice and is one of the completely treated stage.
major reasons for the bacterial safety of our In the industrial waste field, chlorine has
public water supplies. been generally adopted for cyanide destruc-
In addition to its bactericidal qualities, chlo- tion in the presence of an alkaline salt. Other
rine and chlorine compounds are used in the uses in industrial wastes have included the
reduction and removal of objectionable tastes destruction of hydrogen sulfide and phenol.
and odors from drinking water, the oxidation Here, approximately 8.5 parts of chlorine are
of iron, manganese and hydrogen sulfide pres- required for each part of hydrogen sulfide
ent in waters thereby aiding in their removal, present and approximately 20 ppm chlorine
and the oxidation, decolorization and destruc- for each ppm of phenol. The pH of reaction
tion of organic matter. It has been stated that is important to obtain maximum efficiency.
chlorine is probably the most powerful and
flexible tool that chemical science has pro- CHEMISTRY OF CHLORINATION
vided the sanitary engineer. Further uses to
which it has become essential include its in- THEORY oF DISINFECTING AcTION. Chlorine
jection into sewage and industrial wastes prior hydrolyzes in dilute water solution as follows:
to their discharge to streams not only for the Ch + H20 = HOCI
bacterial reduction in these waste waters, but chlorine + water = hypochlorous acid
for the destruction of chemicals that may react + HCI
directly with oxygen, such as sulfides, sulfites, + hydrochloric acid
ferrous iron and others. Chlorine is widely It is the hypochlorous acid produced in this
used for the control of slime and algae in cool- hydrolysis that provides the disinfecting and
ing waters in both once-through and recircu- oxidizing properties of chlorine solutions.
lating systems. Early theories that the sterilizing properties
In the clarification process of water treat- of chlorine were due to nascent oxygen liber-
ment, chlorine is often applied to the raw ated from hypochlorous acid or to complete
water along with the coagulant. Included in oxidation of organic matter have been dis-
the benefits derived from prechlorination proved. Later work has led to the theory that
practices are improved coagulation probably the destruction of organisms results from the
due to the action of chlorine on the organic chemical reaction between hypochlorous acid
material in the water; reduction in taste, odor formed when chlorine is mixed with water
and color producing materials by oxidation; and an enzyme in the organism cell. The
reduction of biological loading by longer con- amount of HOC! is dependent on the pH of
tact with the water thereby providing a safety the solution. At pH below 6.5, the HOC! con-
factor; and oxidation of iron and manganese stitutes almost 100% of the free available
for removal by settling and filtration. chlorine. At a pH of 9.0 and above, the ionized
Chlorine has been used successfully as the hypochlorite ion accounts for practically all
activating agent for sodium silicate in the of the free available chlorine. Since it is hypo-
preparation of the coagulant aid, activated chlorous acid and not the hypochlorite ion
silica. The advantage claimed by this process that is the principal disinfectant in chlorine
is that the chlorine used for the activating solutions, the efficiency of disinfection will be
process continues to be available for steriliza- substantially greater at low pH values where

the hypochlorous acid content is greater. The bination and the resulting retarded action of
free available residual chlorine of a treated chlorine with organic matter, the maintenance
water is considered to be that portion of the of a stable residual in outlying water distri-
total residual chlorine existing in the water as bution systems is facilitated. Although this
the hypochlorous acid. combined residual is ordinarily more stable
than a free chlorine residual, due to its de-
creased bactericidal qualities it is essential to
maintain higher residuals.
The combined available residual chlorine
w of a treated water is considered to be that por-
~ tion of the total residual existing in chemical
3 combination with ammonia or organic nitro-
gen compounds. Studies on the bactericidal
0 properties of free and combined available chlo-
ill rine residuals have indicated that for complete
"' kill with similar contact periods, approximately
25 times as much combined available residual
chlorine is required as is necessary with free
available residual chlorine. By the same studies
it was apparent that to obtain complete bac-
teria kill when using combined available resid-
Figure 8-7 • Breakpoint Chlorination
ual equal to the free residual, approximately
100 times as long contact period was required.
The disinfecting ability of chlorine is re-
its bactericidal action chlorine displays its
duced when hypochlorous acid reacts with
strong oxidizing properties in other ways.
ammonia or amines to form chloramines.
Chlorine will react with and oxidize fer-
}?ormation of chloramines may be shown as
rous iron (Fe++) , manganous ·manganese
(Mn++), nitrite (N0 2 - ) and hydrogen
HOCI + NHl sulfide. Reaction with such inorganic constit-
hypochlorous acid + ammonia uents present in the water being chlorinated
NH 2CI + H20 increases the quantity of chlorine required
= monochloramine + water and adds to the chlorine demand of the water.
2HOCI + NHl In this oxidation process chlorine is reduced
hypochlorous acid + ammonia to the chloride ion which possesses no further
NHCI2 + 2H20 oxidizing properties or disinfecting power. The
= dichloramine + water reaction between hydrogen sulfide and chlo-
The pH of the treated water determines rine illustrates this point.
which product will be formed. At pH 8.5 H2S + 4Ch + 4Hz0
monochloramine is the major product. At hydrogen sulfide + chlorine + water
about pH 4.5 dichloramine is the major prod- H2S04 + SHCI
uct. Mixtures of the two prevail between these = sulfuric acid + hydrochloric acid
two values. Below pH 4.5 nitrogen trichloride Chlorine will also react with and be con-
predominates. sumed by organic matter present in the water.
Where chloramines exist, it is necessary to In some cases chlorine is consumed in direct
provide longer contact time or increased chlo- chlorination of organic compounds present in
rine residual for comparable bacterial reduc- sewage or industrial waste pollution with the
tion. With the oxidizing properties of chlorine formation of chlorinated compounds such as
being substantially reduced when in such com- chlorphenols. Products of this type may possess

undesirable tastes and odors. Chlorine may water is generally described as marginal.
also oxidize and bleach organic matter with Breakpoint chlorination is the application of
the extent of oxidation dependent on time of chlorine to water to maintain free available
contact and the residual chlorine present. In chlorine residuals. The addition of chlorine to
either case, this oxidation of inorganic and water containing ammonia or organic matter
organic compounds requires the use of higher will first produce an increased combined chlo-
amounts of chlorine to satisfy the chlorine de- rine residual. After reaching a maximum
mand before a chlorine residual for disinfec- value, increased chlorine doses result in de-
tion can be established. creased residual values. This decrease is
Chlorine demand may be described as the caused by the decomposition of the com-
difference between the amount of chlorine bined available chlorine residuals, formed
applied to a water and the amount of free, when chlorine is first added. This decrease
combined or total residual chlorine remaining in chlorine residual will not start until
at the end of a specified contact period. The the chlorine applied amounts to at least 8
<:;hlorine demand will vary on any water with times or more the amount of free ammonia
contact time, water quality, temperature and present. If this chlorine-ammonia ratio is ma-
pH. The chlorine requirement is the amount terially less than 8 to 1, say 4 to 1, chloramines
of chlorine that must be added to accomplish will be formed and the breakpoint reactions
a specific objective, i.e., bacterial kill, hydro- will not take place. The ammonia will have
gen sulfide removal or oxidation of iron and been essentially removed when the chlorine
manganese. dosage amounts to 10 times the amount of
The residual chlorine requirement in treated ammonia present. When the instantaneous
water is usually established by bacteriological chlorine demand is high, the chlorine require-
testing to determine the amount necessary for ments to obtain free available chlorine resi-
effective disinfection. Normally, a residual of duals may be 20 or more times the amount of
0.2 to 0.5 ppm is satisfactory after a minimum ammonia present.
10 minute contact period at or above 20 C. After the minimum chlorine residual or
Higher residual values may be required if the breakpoint is reached, the chlorine is present
chlorine exists as combined residual and the as free available residual and increases in di-
pH of the water exceeds 7 .0. The need for rect proportion to the chlorine applied. In
higher combined available chlorine residuals developing the free available residual, destruc-
is markedly increased in colder water. tion of taste and odors is usually achieved;
practically all of the bacteria have been killed,
CHLORINE-AMMONIA PRocEss. Ammonia or organic matter oxidized and the free residual
ammonium salts are employed in conjunction is quite stable and not readily dissipated.
with chlorine to produce a combined available Breakpoint chlorination may also be used
chlorine residual that is useful for certain to control slime and algal growths, to aid
specific problems. This process was originally coagulation, oxidize iron and manganese, to
developed for taste and odor control where remove ammonia from boiler makeup water
free chlorine reacted with organic matter in and to generally improve water quality in the
the water. Further, some minimizing of phe- treatment cycles or in the distribution system.
nolic and other type tastes by chloramines is The object of breakpoint chlorination is to
sometimes accomplished over conditions re- produce and maintain a free available chlo-
sulting from normal marginal chlorination. rine residual. Since a stable free chlorine resi-
The lower bacterial efficiency of this combined dual can exist only beyond the breakpoint,
residual limits the usefulness of the process. the process is controlled by making certain
Actually, this process has been superseded to that a definite free chlorine residual has been
a large extent by the more positive breakpoint produced and that it is adequate to carry
process for taste as well as bacterial control. throughout a plant or into the distribution
BREAKPOINT PROCEss. Normal chlorination of system as required.

l_l_L_l __ -- 1.8
- ...

::::e 4
....,.,' 5 MI,N. CONTACT 7 1.2
"\ \ i\
pH 7.6

J 0.9

::c / 1\ IJ ::::e
c.,) 2
v v \ \ v 0.6 ::::e

v '~ v

2 4 6 8 10 12
'\ - ---- -
-- - i--
18 20



Figure 8-2 • Ammonia Removal by Chlorination

The breakpoint process is of specific appli- and reportedly does not form reaction prod-
cation in the removal of ammonia from boiler ucts that cause taste intensification. The costs
makeup water. Figure 8-2 illustrates results for feeding chlorine dioxide are higher than
obtained in applying this process to Calumet for straight chlorine with the result that in
River water. In these experiments it was found order to justify its use there must be defi-
that greatest ammonia removal efficiency was nite improvements obtained. Fortunately, the
obtained with a chlorine to ammonia ratio of method for determining residual chlorine is
8: 1 and it was possible to reduce the undesir- applicable for chlorine dioxide. Chlorine di-
able ammonia concentration of the steam oxide reacts with acid orthotolidine rapidly
from the boilers to a negligible value. and therefore can be measured by the flash
test and the OTA modification.

Another satisfactory and proven method for Many industrial waters are chlorinated for
using chlorine in the treatment of water, the control of bacteria and slime producing
sewage and industrial wastes is in combina- organisms. However, in food processing in-
tion with sodium chlorite to form chlorine dustries, the chlorine may cause tastes in
dioxide. This process was introduced in 1944 canned and frozen foods or may cause cor-
and is now used in more than 150 water rosion of the metal containers. Also, it has
works plants. It has also found wide indus- been determined that excessive chlorine
trial application for various applications of affects smoothness and brightness of metals
oxidation, bleaching and sterilization. in the plating industry. Paper mills producing
Chlorine dioxide is used in public water fine paper usually establish very critical limits
supplies for taste and odor control as well as for chlorine residuals.
for sterilization. It has high oxidizing powers In normal concentrations, chlorine does

not render water abnormally corrosive to zinc The hypochlorite ions estaplish equilib-
nor does it increase the solubility of lead salts. rium with hydrogen ions, depending on pH,
Reports on irrigation water and subsequent as does gaseous chlorine. The significant dif-
studies indicate no injury to land plants when ference is in the effect on pH. Solutions of
the irrigation water contains less than 50 ppm hypochlorite contain an excess of alkali which
combined residual chlorine. Aquatic plants tends to raise the pH. In some hard waters,
are reportedly harmed by concentrations of scale problems have been experienced due to
3 ppm or more. the precipitation of calcium carbonate by
The concentration of chlorine required to hypochlorites. Usually a complex phosphate
control plankton in reservoirs varies from 0.2 can be added to overcome this difficulty. In
·to 3.0 ppm depending on the sensitivity and other cases, hypochlorite treatment, due to the
type of organisms present. There are some high alkalinity, improves the quality of soft
types that may require as high as 50 ppm for and highly corrosive waters.
ciestruction. Normally, free available chlorine Calcium hypochlorite is soluble in water al-
in water may be toxic to fish and other aquatic though when dissolved in hard water, a pre-
life at low concentrations, and in combination cipitate often results. For liquid feeding of
with cyanide,. phenols and with other sub- this material, the solutions should be prepared
stances may be even more detrimental. at approximately 1% to 2% available chlo-
There are no specific data concerning the nne.
limiting physiological tolerance for chlorine in Sodium hypochlorite, NaOCl, is available
potable water, but it is generally agreed that in several forms ranging from the familiar
the small amounts of chlorine present in chlo- grocery store bleaches to commercial laundry
rinated waters are dissipated by the reaction bleach solution. These products vary in avail-
with saliva and gastric juices as soon as the able chlorine from 3% to as high as 20%.
water is swallowed. It is reported that chlo- Most of them are quite unstable. Sodium
rinated drinking water may, in rare cases, hypochlorite may be satisfactorily fed in the
cause asthma, colitis and eczema. When af- normal 15% available chlorine strength or it
flicted patients drank distilled water in place may be diluted with water for feeding.
of chlorinated water, their conditions im-
The available chlorine, on a pound for
proved and when they returned to chlorinated
pound basis, is equally effective for disinfec-
water, the symptoms reappeared.
tion and other purposes whether obtained
from hypochlorites or from gaseous chlorine.
All hypochlorites are corrosive to some de-
In addition to gaseous chlorine, there are com- gree and must be carefully handled and cor-
mercially available calcium and sodium hypo- rosion resistant materials should be used for
chlorites. The most common forms include storage and dispensing. Satisfactory materials
sodium hypochlorite having approximately include wood, ceramic, glass, plastic or
15% available chlorine and high test calcium rubber.
hypochlorite with 70o/o available chlorine. The
hypochlorite salt most widely used for water
treatment IS the calcium hypochlorite,
Ca(OCl) 2 • Dechlorination may be accomplished by the
When this material is dissolved in water, use of activated carbon or by chemical re-
ionization occurs:
ducing agents such as sulfur dioxide, sodium
Ca(OCih = Ca++ sulfite, sodium bisulfite or sodium thiosulfate.
calcium hypochlorite = calcium ions Prolonged storage in open reservoirs is also
+ 2(0CI)- an effective means for dechlorination. The
+ hypochlorite ions practice of dechlorination may be desirable in



® @ R
0 L::J
GAS POWDER POWDER 20% 15% 10% 5% 1°/o



Courtesy Chlorination Topics, Spring 1956

Figure 8-3 • Comparative Chart Showing Amounts of Material Required
to Provide One Pound of Available Chlorine

either public supplies or for industrial uses. In The use of chlorine gas requires careful
public water supplies, the reduction or re- handling and the employment of special safety
moval of chlorinous taste may be necessary. measures since chlorine is highly irritating to
Many industrial process uses of water are ad- the eyes and to the membranes of the respira-
versely affected by high chlorine residuals and tory system. The physiological effect of various
consequently they must be removed or re- concentrations of chlorine gas are shown here-
duced. Also, it has been concluded that some with.
resins used in exchange softeners and de-.
minera,lizers are adversely affected by the ppm chlorine
presence of a residual chlorine in the water to by volume In air

be processed and consequently dechlorination Least detectable odor 3.5

ahead of these units is required. Cause throat irritation 15
Cause coughing 30
Dangerous for 30 minutes exposure 50
FEEDING EQUIPMENT Quickly fatal 1000

(from U.S. Bureau of Mines Paper 248)

Chlorine gas or solutions or hypochlorites are
in common use and are readily and accu- LABORATORY CONTROL
rately applied to public and industrial water
or sewage and industrial wastes. These ma- The success of treating water or wastes with
terials are applied either at a constant rate or chlorine is dependent upon the accuracy and
proportional to flow, dependent upon the adequacy of the control measures practiced.
specific requirements. The form in which the A number of chemical methods have been
chlorine is to be fed should be properly eval- developed for the control by measuring both
uated and the equipment supplied appropriate free and combined residual chlorine concen-
for the service. The tabulation (Fig. 8-3) trations. The iodometric titration was the
showing comparative amounts of the various initial method used. Modifications of this test
forms of chlorine should be helpful. Reliable have been developed which have improved
equipment for feeding all types of chlorine and the accuracy. The orthotolidine test has been
chlorine compounds is available from repu- very good for the purpose intended. How-
table vendors serving this specialized field. ever, the limitations of this test as well as







Courtesy Wallace & Tiernan Inc.

Figure 8-4 • V-notch Chlorinator

interference from various substances cause it reduction potential. Additional instruments

to be used with certain caution. A modifica- and methods of control are continuously being
tion of this test known as the OT A test has developed for this very important application.
been widely used although certain inherent
weaknesses continue to exist. The micro- REFERENCES
titration of free chlorine with methyl orange "Chlorine Manual", 3rd Ed., The Chlorine Insti-
or xylene cyanole has been successfully used, tute, Inc., New York, N.Y. ( 1959)
especially in cooling water systems. A test for A. E. Griffin and N. S. Chamberlin, "Some Chemi-
cal Aspects of Breakpoint Chlorination", Journal,
titrating chlorine with a reducing agent has New England Water Works Assoc., Vol. LV, pp.
been developed and is generally known as the 371-383 (1941)
amperometric method and is gaining in ap- E. J. Laubusch, "Chlorination of Water", Water &
plication and popularity. Fortunately, resid- Sewage Works, Vol. 105, pp. 411-417 (1958)
E. W. Moore, "Fundamentals of Chlorination of
ual chlorine recording can now be made Sewage and Waste", Water & Sewage Works, Vol.
with the instrumentation presently available. 98, pp. 130-136 (1951)
Some industrial waste chlorination treat- C. N. Sawyer, "Chemist~y for Sanitary Engineers",
pp. 246-256, McGraw-Hill Book Company, Inc.,
ment can be controlled by the oxidation- New York, N.Y. (1960)

Hydrogen Sulfide Ren1oval

Difficulties Encountered. . . . . . . . . . . . . . . . . . . . . 56
Methods for Removal . ..................... 56
Aeration . ............................. 56
Aeration with pH Reduction . . . . . . . . . . . . . . . 57
Chlorination. . . . . . . . . . . . . . . . . . . . . . . . . . . 58
Anion Exchange . . . . . . . . . . . . . . . . . . . . . . . . 58
Analysis for Sulfide. . . . . . . . . . . . . . . . . . . . . . . . 59
Application and Limitations. . . . . . . . . . . . . . . . . 59

YDROGEN sulfide removal involves the Figure 9-1

H use of mechanical or chemical processes
to secure the elimination of this undesirable
Percentage of Total Sulfide Present as H2S

gas from solution. pH

Hydrogen sulfide may be present in a water 5.0 .................................... 98
supply due to purely chemical or purely bi- 6.0 ..................•................. 83
ological causes. The presence of hydrogen 6.5 ..••.•.•••..••...•.••.•...••.......• 61
7.0 .................................... 33
sulfide is more common to well supplies than 7.5 ..................................... 14
to surface supplies. In well waters the hydro- 8.0 .................................... 4.8
gen sulfide results generally from the location 9.2 .................................... 0.32
such as Boulder clay areas or "sour gas" oil
fields. In other supplies, it is generally the re- a corrosive character to the water.
sult of contamination due to local atmos- Hydrogen sulfide accounts for a well oc-
pheric conditions, the presence of some types casionally yielding a water that is black due
of bacteria in the absence of oxygen, or the to ferrous sulfide. Ferrous sulfide is usually ex-
contamination of mine waters by sulfur. tremely finely divided and only a small amount
Bacteria often encountered in industrial is required to blacken the water. Excessive
waters may be responsible for the presence hydrogen sulfide concentrations exhibit an ag-
of hydrogen sulfide. The most common of the gressive tendency toward iron, steel, and cop-
sulfur bacteria are those that split hydrogen per alloys, even in the absence of oxygen.
sulfide from organic matter containing sulfur. Greensand zeolites are damaged by the pas-
Most protein matter contains sulfur, therefore sage of water containing hydrogen sulfide,
hydrogen sulfide will be produced when these Cases have been reported where the mineral
compounds are broken down by bacteria in has been so badly attacked in a few months
the absence of oxygen. One group of sulfur by water containing only 2 ppm of hydrogen
bacteria produces hydrogen sulfide by there- sulfide that the bed could not be recondi-
duction of inorganic sulfates, sulfites and sul- tioned. Special minerals for the catalytic re-
fur itself. moval of iron and manganese from water are
Hydrogen sulfide possesses a definite solu- also affected by the presence of hydrogen sul-
bility in water which is approximately three fide. With the use of these materials, manu-
times that of carbon dioxide and imparts a facturer's guarantees are based on the absence
characteristic taste and odor (rotten eggs) to of this gas.
the water. This odor becomes objectionable Another effect of hydrogen sulfide is to re-
when present in a concentration of 1 ppm act with chemicals used for the protection of
or more and can be detected when present in systems against corrosion thus increasing the
amounts less than 1 ppm. dosage required and, in some instances, in-
If the water has a high pH, the odor may creasing cost to an extent that treatment of the
be slight, since most of the sulfide may be system with positive corrosion inhibitors is not
present as alkaline sulfides rather than as economically feasible.
hydrogen sulfide. Figure 9-1 shows the per-
centage of total sulfide present as hydrogen sul- METHODS FOR REMOVAL
fide with the difference in the total sulfide con-
tent being that present as the alkaline sulfides. AERATION. The removal of hydrogen sulfide
by aeration follows Henry's law. Hydrogen sul-
DIFFICULTIES ENCOUNTERED fide is relatively soluble in water while the par-
tial pressure of this gas in the normal atmos-
Removal of hydrogen sulfide is desirable not phere is practically "zero." The establishment
only from the standpoint of odor, but because of a state of equilibrium between the water
even small concentrations of this gas impart and the air will result in the saturation of the

water with nitrogen and oxygen. The basic sulfides and the sulfide as discussed previously.
fundamental in securing removal of gases from By reducing the pH of the water, the ioniza-
water by aeration is the establishment of a tion constant is shifted and more of the total
state of equilibrium between the gases in the sulfide is converted to hydrogen sulfide in
water and the gases in the surrounding at- which form it is available for removal by
mosphere. aeration.
Reduction in hydrogen sulfide content is In recent yea~s flue gases which are high in
obtained with many different means of aera- carbon dioxide content have been employed
tion. The absorption of oxygen by the water in reducing the pH value sufficiently to liber-
during aeration also assists hydrogen sulfide ate hydrogen sulfide. The quantity of gas re-
removal by accomplishing oxidation of a part quired for pH reduction will depend upon
of the hydrogen sulfide and alkaline sulfides the alkalinity of the supply being treated.
to free sulfur. Forced draft aeration has been One type of equipment for this process con-
found desirable where there is an appreciable sists of a double stack with a water seal sepa-
hydrogen sulfide concentration to be removed. rating the upper carbonating chamber from
Experience has indicated the desirability of the lower aeration chamber. The flue gas is
considerable cross-sectional area and a tall introduced at the bottom of the carbonating
aeration tower. With such equipment hydro- chamber and mixes with the influent water
gen sulfide can be reduced to 1-2 ppm. The flowing counter-current. Most of the hydro-
amount of free sulfur precipitated may prove gen sulfide is removed at the reduced pH in
objectionable and require filtration for its the upper section. In the lower section, which
removal. is similar in design to the carbonating cham-
Two general types of aerators are employed ber, a counter-current of air provides addi-
for hydrogen sulfide removal; one a pressure tional scrubbing of the water with air alone
type aerator and the other a gravity type aer- for the removal of carbon dioxide and further
ator. The pressure type has the advantage removal of any residual hydrogen sulfide. Also,
that line pressure is not broken. The gravity it is possible that any remaining alkaline sul-
aerator requires that line pressure be broken; fides will be oxidized to sulfur. This type of
however, more efficient aeration is obtained. hydrogen sulfide degassifying apparatus is il-
At present the tendency is toward the installa- lustrated by Figure 9-2. The introduction of
tion of gravity aerators. carbon dioxide at the bottom of the carbonat-
In the use of aeration alone for the removal ing chamber permits a counter-current flow
of hydrogen sulfide, complete removal of this of gas and water through the entire length of
gas is not obtained. At higher pH values the the chamber, thus permitting maximum con-
sulfide content will be present chiefly as al- tact time between the gas and the water. The
kaline sulfides which will result in incomplete taller the tower used the greater is the con-
removal by aeration alone. This condition re- tact time provided.
sults because carbon dioxide is less soluble than Reduction in pH of the water may also be
the hydrogen sulfide and is removed by aera- obtained by the feed of a mineral acid. The
tion more rapidly than hydrogen sulfide. As feed of acid is controlled to obtain the desired
the carbon dioxide is removed, the pH rises pH reduction. This method provides a posi-
and the equilibrium between alkaline sulfides tive control over pH reduction.
and hydrogen sulfide is altered so that the re- By combining pH reduction with aeration,
action proceeds in the wrong direction for the the size of the forced draft aerator can be
most complete removal of hydrogen sulfide. considerably reduced. More effective hydro-
AERATION WITH pH REDUCTION. Removal of gen sulfide removal is also secured. However,
hydrogen sulfide by aeration is most effective with this method some precipitation of flowers
at reduced pH in the range of 4.0 to 5.0 be- of sulfur can be expected. Coagulation and
cause of the equilibrium between the alkaline filtration may be necessary if the initial total

aeration or aeration with pH adjustment may

not provide sufficient treatment. Complete re-
moval of hydrogen sulfide frolJl water can be
secured by means of chlorination or a com-
bination of aeration and chlorination. One
reaction with chlorine is as follows:

HzS + Clz =
hydrogen sulfide + chlorine =
2HCI + S
hydrochloric acid + sulfur

In this reaction, the formation of flowers

of sulfur is shown. This fine colloidal precipi-
tate should be removed by coagulation and
filtration to prevent reversion to hydrogen sul-
fide by the action of certain microorganisms.
Theoretically, 2.1 ppm of chlorine are re-
quired for the removal of each ppm of hydro-
gen sulfide.
When used in greater quantities chlorine
will further oxidize hydrogen sulfide to sul-
furic acid without the precipitation of sulfur.
This reaction is shown below:

H2 S + 4C)z + 4H 20 =
hydrogen sulfide + chlorine + water =
HzS04 + 8HCI
sulfuric acid + hydrochloric acid
Theoretically, 8.4 ppm of chlorine are re-
quired for the removal of each ppm of hydro-
gen sulfide. This high chlorine consumption
considerably increases the cost of this method
Figure 9-2 • Hydrogen Sulfide Degasi.fier of hydrogen sulfide removal. Consequently, it
is usually the more economical procedure to
sulfide content of the water is high. remove as much hydrogen sulfide as possible
by aeration with reduced pH and to use com-
CHLORINATION. While aeration alone and aer-
plete chlorine oxidation, if necessary, for re-
ation combined with pH reduction will reduce
moval of the residual hydrogen sulfide after
the hydrogen sulfide content of a water sup-
ply to a relatively low value, depending upon
the amount of gas originally present, the hy- ANION ExcHANGE. Anion adsorbers have re-
drogen sulfide content of a water can not be re- cently been developed which are applicable
duced to "zero" by either of these methods. for the removal of hydrogen sulfide from water
Depending on the use of the water, complete supplies. Highly basic resins have been pre-
removal of hydrogen sulfide may be required pared which, when regenerated with salt and/
as even small quantities produce an offensive or sodium bicarbonate depending upon the
odor and result in increasing the corrosive pH of the water, exchange their ions for sul-
characteristics of the water. Consequently, fides, sulfates, etc. This reaction is illustrated

by the following equation in which "R" repre- acetate solution ( 220 g Zn ( C 2 Ha0 2 ) 2 .2H2 0
sents the resin matrix: per liter of distilled water) and 1N sodium
R4N.OH + H2S hydroxide solution (40 g NaOH per liter of
basic exchange resin + hydrogen sulfide = distilled water) . To fix a sample, place 1 ml
R4N.HS + H20 of zinc acetate in a 16 oz bottle. Collect the
sulfide resin + water sample (approximately 16 oz) in this bottle
with a minimum of agitation. Add 1 ml so-
In accomplishing this removal a low cost
dium hydroxide, stopper and mix. The fixed
regenerant is employed, line pressure is not
sample should be analyzed within 24 hours
broken, excess oxygen is not added to the
for maximum accuracy. The method of analy-
water, and a minimum of equipment is re-
sis usually employed is colorimetric, based
upon the reaction between paraaminodi-
The equipment employed is a pressure type methylaniline, ferric chloride and the sulfide
softener shell designed for counter-current ion to produce methylene blue.
flow utilizing highly basic anion exchange
resins such that dealkalization will also be ob- APPLICATION AND LIMITATIONS
tained and the chloride content of the water
will be increased. By counter-current flow is In general, the method or combination of
meant the up flow of water during the service methods employed for hydrogen sulfide re-
run and the down flow of water and regen- moval involves individual study. The method
erant for rinsing and regeneration. of removal is determined by such factors as
At present the use of the highly basic anion the total sulfide content of the water, alkalin-
resins for hydrogen sulfide removal has been ity of the water, desired effluent sulfide con-
limited to domestic uses though it is con- centration, use or purpose for which the water
sidered that they are satisfactory for some in- will be employed, existing plant equipment,
dustrial uses. They have the limitations of low etc.
capacity, low flow rates, and iron must be ab- The methods discussed in detail above are
sent from the water. It is preferable for the the ones in general use. However, there are
influent water to be low in hardness. However, other methods employed in specific cases
the process can be employed with the use of which are too limited for general application
a dual bed to both soften the water and re- and which in general are indirect methods of
move hydrogen sulfide. It is also possible hydrogen sulfide removal. Examples of these
in some cases for existing zeolite softeners to methods are the removal of sulfide by ozona-
be converted to perform both hydrogen sulfide tion, usually accomplished as an indirect re-
removal and softening of the water. sult of the sterilization of water, bacterial ac-
tion on sulfides over long periods of time fol-
ANALYSIS FOR SULFIDE lowed by sterilization and filtration, and slow
aeration in an open reservoir followed by the
When dissolved gases are present in a water use of a coagulant, sterilization, and filtration.
supply, it usually is necessary to analyze the
water at the point of origin to obtain accurate REFERENCES
results. For example, this procedure is manda- M. E. Flentje, "Aeration", Journal, Am. Water
Works Assoc., Vol. 29, pp. 872-880 (1937)
tory if the gases are carbon dioxide or dis-
J. E. Foxworthy and H. K. Gray, "Removal of Hy-
solved oxygen. The determination of hydro- drogen Sulfide in High Concentrations from Water",
gen sulfide and total sulfide is a rather in- Journal, Am. Water Works Assoc., Vol. 50, pp. 872-
volved analysis which is not readily performed 878 (1958)
S. T. Powell and L. G. von Lossberg, "Removal of
in the field. Fortunately, samples for sulfide Hydrogen Sulfide from Well Water", Journal, Am.
can be "fixed" at the source and then trans- Water Works Assoc., Vol. 40, pp. 1277-1289 (1948)
ported to a laboratory for analysis. S. W. Walls, "Hydrogen Sulfide Problems of Small
Water Systems", Journal, Am. Water Works Assoc.,
The fixing reagents employed are 2N zinc Vol. 46, pp. 160-170 (1954)

Iron Ren1oval and
Iron Retention
Methods of Iron Removal . .................. 61
Aeration, Sedimentation and Filtration. . . . . . . . 61
Coagulation, Sedimentation and Filtration . . . . . . 62
Lime and Lime-Soda Softening . . . . . . . . . . . . . 63
Cation Exchange . . . . . . . . . . . . . . . . . . . . . . . . 63
Contact Filtration. . . . . . . . . . . . . . . . . . . . . . . 64
Iron Retention . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
Iron Bacteria . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
Manganese Removal . ...................... 67
Applications and Limitations. . . . . . . . . . . . . . . . . 67

RON removal is the process by which objec- change surfaces become fouled with iron oxide
I tionable amounts of iron are removed from
water either in one step or a series of steps.
deposits. Valves, lines, etc. will be plugged by
the precipitated iron. Ferric iron present in
Iron retention is the treatment of iron- water to be zeolite softened results in fouling
bearing water with surface active agents, of the zeolite beds and loss of efficiency.
capable of forming a complex with iron, Where iron is present in boiler feedwater it
thereby retaining iron in solution and pre- constitutes another source of potential scale
venting precipitation which would otherwise forming tendencies.
occur. In many industries, the presence of iron in
process water may be highly undesirable. In
the paper, tanning, textile, beverage and ice
manufacturing industries, for example, the
presence of iron is highly detrimental to the


Iron can be removed from water by several

different processes and combinations of proc-
esses. In fact, there are so many different
methods that can be employed for iron re-
moval that careful study is required to select
a method most suitable to any individual case.
Selection of an iron removal method will be
influenced not only by the amount of iron in
the raw water and the form in which the iron
is present, but also by the ultimate use of the
water and the degree of softening that may
Figure 10-1 • Typical Iron Deposition in Pipe Line also be desired.
Iron in water will deposit in water works Aeration is a process which consists of inti-
distribution systems increasing the frictional mately admixing water and air in some man-
resistance through pipes; water meters will ner. Where aeration is employed for iron re-
become heavily coated with iron oxide and moval, two actions occur simultaneously. First,
must be cleaned regularly in order to register carbon dioxide is removed by aeration, thus
correctly. If water containing large amounts increasing the pH of the water. Secondly, the
of iron is used in the home, it will injure the oxygen added to the water by aeration enters
color and taste of many edibles and will des- into the chemical oxidation of the iron. As a
troy the sanitary appearance of sinks and result of these two actions, precipitation of
cooking utensils. Clothes washed with such a the iron occurs as ferric hydroxide, illustrated
water will develop yellow stains. Crenothrix by the following equation.
or "iron bacteria" frequently accompany 4Fe(HC03h + 0 + 2H 0 =
2 2
waters of high iron content. These organisms ferrous bicarbonate + oxygen + water =
established in a city distribution system cause 4Fe(0Hh + 8C0 2
clogging of service mains and are responsible ferric hydroxide + carbon dioxide
for disagreeable odors. Iron is frequently present in well waters in
Where well waters containing iron are used the ferrous or unoxidized state. Increase in
in industrial cooling water operations, deposi- pH in the absence of oxygen would precipi-
tion of iron from solution will occur. Heat ex- tate ferrous hydroxide, but since ferrous hy-






Figure 10-2 • Iron Removal by Aeration, Settling and Filtration

droxide is more soluble than ferric hydroxide, ing relatively small amounts of iron and car-
complete iron removal requires oxidation of bon dioxide.
ferrous iron to the ferric state, with precipita- Under some circumstances, chlorine and
tion of the highly insoluble ferric hydroxide. other oxidants have been employed instead
Complete precipitation of iron by aeration of aeration in order to oxidize and precipitate
is not instantaneous. For this reason, it is ad- iron. Usually, such methods cannot compete
visable to employ a retention and sedimen- economically with aeration.
tation tank following the aerator to provide While the insoluble ferric hydroxide pre-
time for the precipitation and settling of ferric cipitated from an iron-bearing water may be
hydroxide, prior to filtration. Sludge settled removed by settling alone if sufficient time is
in this tank is removed periodically. provided, it is better practice to follow with
Even with an adequate supply of oxygen, filtration in order to insure complete iron re-
the precipitation of ferric hydroxide is hin- moval in a reasonable period of time. Stand-
dered by low pH values. For effective iron re- ard pressure or gravity filters are employed
moval by aeration the pH should be increased for this purpose, using either sand or anthra-
to approximately 7.5. If aeration alone is un- cite coal as the filter medium.
able to raise the pH to this point, the supple- CoAGULATION, SEDIMENTATION AND FILTRA-
mental feed of an alkali is required, following TioN. Some waters, usually surface supplies
aeration, to complete the precipitation of fer- and particularly those from swampy areas,
ric hydroxide. may contain appreciable organic matter which
Coke-tray aerators are frequently used in acts to retard or prevent the precipitation of
the removal of iron. Other types of aerators iron. Even following aeration and pH adjust-
such as cascades, cones, nozzles and towers ment, little removal of iron may be secured.
are also employed. Provision must be made to Chlorination may be of value because of its
periodically clean the unit of the precipitated destructive action on the organic present, but
iron. Figure 10-2 shows a flow sheet for iron may be costly and not fully effective. Where
removal by aeration, sedimentation and fil- such waters are encountered, coagulation is
tration. usually the most effective measure, followed by
To avoid re-pumping, aeration may be con- sedimentation and filtration. The coagulant
ducted under pressure. Carbon dioxide is not most widely employed is alum although it is
removed and an alkali feed may be required necessary to conduct tests to determine the
to precipitate the iron. This method of iron most effective and economical coagulant for
removal is applicable only to waters contain- each particular water.

LIME AND LIME-SODA SoFTENING. Where Removal of iron by cation exchange is simi-
either lime softening or lime-soda soften- lar to and simultaneous with the removal of
ing is employed, the pH of the treated calcium and magnesium. The cation exchange
water is elevated sufficiently high to provide material may be green sand, carbonaceous or
the softening reactions and iron is precipitated resinous. When operated on the sodium cycle,
as ferric hydroxide. Aeration is usually pro- iron removal takes place as follows:
vided prior to softening to supply any oxygen Fe(HC03h + NazZ
needed for the conversion of ferrous to ferric ferrous bicarbonate +
sodium zeolite =
iron. Aeration is frequently desirable prior to FeZ + 2NaHC03
cold process softening, regardless of any iron ferrous zeolite +
sodium bicarbonate
problem, in order to remove carbon dioxide Regeneration reactions with salt are as follows:
and thereby decrease lime requirements. FeZ + 2NaCI
The aerator frequently is located over the ferrous zeolite + sodium chloride =
lime-soda softener. As the pH of the water is Na2Z + FeCI2
increased by the feeding of lime, oxidation of sodium zeolite + ferrous chloride
iron is rapid. Precipitation and sedimentation Similar softening and regeneration reac-
can take place in the same unit, followed by tions take place when the cation exchanger is
filtration as illustrated in Figure 10-3. operated on the hydrogen cycle in which case
Iron removal, of course, is readily accom- regeneration is secured with the use of sulfuric
plished under the temperature and pH condi- acid.
tions existing in hot process lime or lime-soda While iron removal by cation exchange
softening. Iron, precipitated in such units, will possesses a field of usefulness where complete
serve to aid coagulation. hardness removal is also necessary, it is not
Iron removal in conjunction with cold proc- suited to those applications where hardness
ess lime or lime-soda softening may be fol- removal is not required. It is possible, how-
lowed, on occasion, by acidification and zeolite ever, to overrun the cation exchanger from the
softening where the use of the treated water standpoint of hardness removal while con-
requires low hardness. tinuing to secure iron removal.
CATION ExcHANGE. With certain waters, cat- In order to avoid fouling the exchanger bed
ion exchangers may be employed for the re- with ferric hydroxide, it is necessary that the
moval of soluble iron from water. The cation iron content of the influent water be present
exchange process is limited essentially to well only as soluble ferrous iron. Air must be ex-
waters where the iron content is in the ferrous cluded from contact with the well water prior
state. to the exchange unit.




Figure 10-3 • Iron Removal by Lime Treatment


Cation exchange is applied only to waters of in the process of iron removal. Backwashing
limited iron content and is not recommended of the precipitated iron from the bed is also
for the removal of high iron concentrations. necessary. Filter rate is 3 gallons per square
CoNTACT FILTRATION. Contact filtration ac- foot per minute and backwash rate is a mini-
complishes oxidation and filtration of the iron mum of 8 gallons per square foot per minute.
in one step. Two types of contact filtration are Potassium permanganate is an expensive re-
employed. In one type, the material of the fil- generant chemical and manganese zeolite in-
ter bed simply acts as a catalyst for the reac- stallations are usually limited to waters of low
tion between the iron and oxygen in the water. iron content or employed as a polishing de-
In the other type, the filter bed supplies the vice, following other processes which have re-
oxygen needed for oxidation of the iron, and moved the major portion of the initial iron
this oxygen must be periodically replenished content of the water.
by use of a regenerant chemical.
Catalytic beds usually involve a filter me-
dium containing manganese dioxide. Stand- For many industrial water uses, it is not neces-
ard pressure or gravity filter shells are em- sary that the iron be removed, provided the
ployed with the 30-36 inches catalytic bed water can be stabilized to prevent or retard
supported on gravel. Flow rates are limited precipitation. Well waters, high in iron con-
to 4 gallons per square foot per minute or less tent, are frequently employed for cooling pur-
and backwash for removal of precipitated iron poses on a once-through basis. By stabilizing
is at the rate of 12-15 gallons per square foot the iron content, deposits can be prevented
per minute. Units of this type present a simple throughout the system. The cost of stabilizing
method of iron removal since no chemical the iron is frequently less than the cost of iron
treatment is required. Proper functioning of removal and, in addition, the investment in
such units, however, requires that the water equipment is avoided.
be low in organic matter, free of hydrogen Stabilization of iron can be secured with the
sulfide, possess a minimum of pH of 6.5 and use of surface active agents, such as polyphos-
contain sufficient oxygen to react with the phates and organic sequestrants. A complex
iron present. If there is inadequate oxygen· is formed between the iron present in the
content of the raw water, oxygen must be sup- water and the surface active agent added. By
plied by an aerating device. This aeration can this complex formation, precipitation of ferric
be accomplished under pressure, without need hydroxide can be delayed or, with sufficient
for repumping. treatment concentration, prevented indefi-
Manganese zeolite is an example of contact nitely. Usually, economic considerations limit
filtration where the filter bed itself supplies such stabilization to once-through systems. In
the needed oxygen for iron removal. Manga- recirculating cooling water systems, the con-
nese zeolite is green sand which has been tinuous aeration and long retention period so
treated with manganous chloride to convert it increases the quantity of stabilizing agent re-
to manganous zeolite. Then, treatment with quired as to make iron removal from the
potassium permanganate results in the pre- makeup water the more desirable practice.
cipitation of the higher oxides of manganese The quantity of surface active agent re-
on the zeolite grains. quired to retain iron in solution will vary ac-
No removal of hardness is involved with the cording to such factors as the iron content, al-
manganese zeolite process. The oxidation re- kalinity, pH and solids content of the water.
actions occur between the natural iron con- Another important factor is the temperature
tent of the water and the manganese oxides to which the water will be heated. For higher
on the zeolite grains. Regeneration of the unit temperatures, increased treatment is required
with potassium permanganate is required, as to overcome the increased tendency for pre-
the oxidizing ability of the bed is consumed cipitation. In general, the organic iron reten-

tion agents are more effective than the poly- most economical, it· is necessary that the sta-
phosphates. Another advantage of the organic bilizing agent be fed to the water prior to a
agents is that they can be used at higher point where precipitation has already occurred
concentrations, on waters of high iron and and that the amount be limited to that neces-
high hardness content, without danger of pre- sary to retain iron in solution under the con-
cipitating calcium phosphate. Under certain ditions of that individual system.
circumstances, a combination of organics and
polyphosphates represents the most advan-
tageous treatment.
;fhe retention time of the water in the sys- The iron bacteria are an unusual group of
tem is another important variable affecting microorganisms found in industrial waters,
the treatment concentrations required. If, for streams, lakes, wells and in potable water
example, a simple system of well water supplies. These organisms may occur alone or
pumped through a single heat exchanger and in different combinations with other species
then to waste is involved, then there is little of bacteria, fungi and algae.
retention time in the system. Only sufficient Although the iron bacteria have not been
treatment may be needed to retain iron in studied as completely as some of the other
solution for 10-15 minutes. On the other hand, organisms this lack of study does not indicate
in a large plant with many varied cooling that they have not caused serious difficulty in
operations, and numerous smaller lines at low industrial and fresh water systems. There are
flow rates, it may be necessary to feed suffi- many case histories regarding the effect of
ci~nt stabilizing ~gent to retain iron in solu- these iron bacteria in city and industrial water
tion for 24 hours. The cost of iron retention systems. The published reports on the effect
treatment increases as the retention time in of iron bacteria in city water systems are many
the system increases and all such data are re- and are justifiably referred to as "water ca-
quired for accurate treatment estimations. lamities." The universal distribution of the
The proper surface active agents are ef- iron bacteria is borne out by the development
fective in preventing iron precipitation re- of these water calamities in Europe and in the
gardless as to whether the iron is in the ferrous United States. Recently these microorganisms
or ferric state. It is advisable, however, to have gained recognition in industry by their
introduce the stabilizing material to the water effect upon process equipment and on the
prior to its exposure to air and prior to chlori- finished product. Usually the affected water
nation in order to permit complex formation will become slightly turbid or acquire a light
with the iron as quickly as possible. reddish tint and/or objectionable odor. As
Equipment for feeding iron retention agents these microorganisms increase in number the
can be relatively simple. Where the raw water water may become more turbid, and the color
is at a constant flow rate, a chemical solution of the water will become brick-red. Hence the
tank for one day's treatment supply can be common reference to "red water." In addition
used together with a small constant rate to discoloring the water this group of micro-
chemical pump. If the raw water flow is vari- organisms produces undesirable accumula-
able, then a proportional method of chemical tions in pipes, nozzles, spray ponds, etc. These
feed should be employed so as to permit the deposits in time will slough and plug lines,
most economical use of treatment. foul pumps, valves and/or affect the quality
Iron retention agents are available which of the finished product.
possess such powerful ability to form iron com- The growth and development of iron bac-
plexes that iron can be retained in solution teria may manifest itself in several different
practically indefinitely. Such agents also pos- ways. For example, it has been reported that
sess the power to dissolve previously precipi- these microorganisms have reduced the ef-
tated iron. However, for iron retention to be fective area of a 6 inch pipe to that of a 2 inch

pipe in a matter of weeks. The accumulation Crenothrix, Clonothrix, Gallionella, Sidero-

in pipes, ponds, nozzles and so forth may be capsa, Siderosphaera, Sideronema, Ferribac-
hard and crusty or it may be relatively light terium, Sideromonas, N aumqnniella, Ochro-
and spongy in appearance. Often other bio- bium, Siderococus, Siderobacter and Ferro-
logical slimes may be associated with iron bacillus.
bacteria and thus further reduce or restrict In general, the iron bacteria are extremely
the flow of water through pipes or entrap other difficult to culture in the laboratory using
debris. The association of other species of the usual culture media. Therefore it is not
microorganisms with the iron bacteria will in- possible to evaluate or to determine the num-
crease turbidity, increase color, give rise to ber of these microorganisms in different
objectionable odors and tastes in domestic water supplies with the usual bacteriological
water supplies and also cause sloughing of de- procedures. A satisfactory procedure for de-
posits in industrial systems which may affect termining the population index of these micro-
the process or the product being produced. organisms in water supplies is a direct micro-
The iron bacteria are usually considered scopic analysis of the water. This analysis,
as typical fresh water organisms. Recently, however, will only determine the number of
however, some of the iron bacteria have been the more obvious species of this group of
isolated from high brines, which indicates the microorganisms.
versatility of this class of organisms. These In general, the iron bacteria show prefer-
microorganisms are usually considered as ence for the lower temperatures; these organ-
aerobic, but they have also been found to grow isms have been observed to grow at tempera-
in waters with low oxygen content. Although tures which range from 0 C to 40 C. How-
the iron bacteria are widely distributed in na- ever, their optimum temperature will range
ture and are usually found in wells, cool from 6 C to 25 C.
springs and brooks they have also been isolated Hydrogen ion concentration has a great
from marine waters, and connate waters (high effect upon the growth and development of
brines) . Although they are usually found in these bacteria. Although published reports
waters which are reasonably high in iron they show that the different species have different
may be occasionally found in waters which requirements, the pH range for growth will
contain a very low amount of iron. vary from 5.5 to 8.2. The optimum pH ap-
The prinicipal distinguishing characteris- pears to be approximately 6.5. Much has been
tics between the iron bacteria and other types written regarding the sensitivity of the iron
of microorganisms is that they have the capac- bacteria to light but their general reaction to
ity to absorb and to accumulate iron and/or light appears to be doubtful. These organisms
manganese when grown in environments have been found to grow in exposed areas, in
which contain these elements. These organ- shade and in complete darkness such as in
isms deposit iron and manganese salts around pipes, wells, etc.
'their cells which results in the characteristic Obviously the growth and development of
reddish brown to black color. iron bacteria and the deposition of iron in
These autotrophic (self-sufficient) organ- their sheaths is influenced by the amount of
isms oxidize iron compounds as a source of dissolved iron in their immediate environ-
energy. However the amount of energy re- ment. The iron bacteria have been found in
leased by the oxidation of iron is relatively waters with as much as 30 ppm of iron and
small. they have also been found in waters which
On the basis of their morphological charac- contain as low as 0.1 ppm of iron. In the low
teristics alone the iron bacteria can be placed concentrations the growth and development
into several different categories. The most of the iron bacteria is dependent upon the
common iron bacteria fall in the following flow of water or where the iron supply is
genera: S phaerotilus, Leptothrix, Toxothrix, continually renewed.

The control of iron bacteria in fresh and for its precipitation. However, coagulation,
salt water systems is possible with the use of lime and lime-soda softening, cation exchange
biocides. The selection of the most appropriate and contact filtration are effective methods
biocide will, of course, depend upon the en- for removal of manganese, as well as Iron.
vironment in which the organism is growing,
the general conditions of use of the water, and APPLICATIONS "'NO LIMITATIONS
whether the water is for industrial or domestic
use. Obviously the use of control agents in There are many combinations of treatment
domestic systems must be limited to those methods that can be employed for iron re-
which have the approval of the Public Health moval. A detailed study of each individual
Service. Under the usual operating conditions plant problem is necessary in order to select
for city water supplies the most suitable agent the particular system best suited to the in-
is chlorine. However, in industrial systems it dividual case. An iron removal method can-
is possible that biocides other than chlorine not be selected solely on its efficiency with
will give satisfactory microorganism control. respect to iron, but must be correlated with
In some industrial systems, where the use of other quality requirements of the water, such
chlorine is not indicated because of precipita- as the need also for low hardness in the
tion of iron and manganese, it may be neces- treated water.
sary to use a special proprietary formulation. Where there are both potable and indus-
These proprietary formulations, although usu- trial uses involved, it may be desirable to em-
ally more expensive than chlorine, will pro- ploy two different methods of iron removal.
duce the desired effect. It is, of course, Also, in some cases, iron removal may be
essential that these industrial biocides be com- employed for potable and process water,
patible with iron retention formulations and while the larger volume of cooling water is
with corrosion inhibitors. In addition, the treated for iron retention rather than iron
biocides must not affect the quality or manu- removal. A careful study, relating all treat-
facture of the product. ment methods to the individual plant needs,
is highly advisable because of the many dif-
ferent methods that can be employed both for
iron removal and iron retention.
Manganese removal is closely associated with
iron removal. Fortunately, manganese is not REFERENCES
encountered as frequently as iron in water
S. B. Applebaum, "Iron and Manganese Removal",
sources and usually is present in smaller Proceedings, Engineers Soc. of Western Penna., pp.
quantities. However, the methods used for 49-57 ( 1955)
manganese removal are essentially the same R. B. Conlan, "Correcting the Iron Problem in
as for iron removal. In most waters of ap- Water Systems", Heating and Ventilating, Vol. 48,
pp. 77-79 (Dec. 1951)
preciable manganese content, iron is also "Water Quality and Treatment", 2nd Ed., pp. 359-
present so that the treatment must be such 368, Am. Water Works Assoc., New York, N.Y.
as to remove both undesirable constituents. ( 1950)
R. S. Wolfe, "Cultivation, Morphology and Classi-
Manganese is less readily oxidized than fication of the Iron Bacteria", Journal, Am. Water
iron by aeration and a higher pH is required Works Assoc., Vol. 50, pp. 1241-1249 (1958)

11 Chemislly Involved. . . . . . . . . . . . . . . . . . . . . . . . 69
Use of Lime and Soda Ash . . . . . . . . . . . . . . . . 69

Lillle-Soda Softening
Use of Lime . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
Use of Caustic Soda . ..................... 71
Use of Gypsum . ........................ 72
Propagating the Chemical Precipitating Reactions. . 72
Equipment Employed. . . . . . . . . . . . . . . . . . . . . . . 73
Intermittent or Batch Cold Softeners. . . . . . . . . 73
Continuous Sedimentation Type. . . . . . . . . . . . . 73
Continuous Sludge Contact Type. . . . . . . . . . . . 73
Continuous Hot Process Lime-Soda Softening. . . 74
Applications. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
Limitations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77

TIME-SODA softening is the process by which very large quantltres of water are softened,
L the calcium and magnesium salts con- such as a city plant; it is usually more eco-
stituting the hardness content of a water are nomical to employ unslaked lime-GaO-
chemically precipitated and removed through which is first slaked to form hydrated lime at
the use of lime (calcium hydroxide) and the plant in accordance with the following
soda ash (sodium carbonate) . This process reaction:
may be carried out at normal raw water {1)
temperatures, in which case it is referred to CaO + H20 =
as "cold process" or at temperatures near or unslaked lime (calcium oxide) + water =
above the boiling point, referred to as "hot Ca(OHh
process." hydrated lime (calcium hydroxide)
The lime-soda ash process is based upon Soda ash-Na 2 C03 -of 98% purity is em-
four essential factors: ployed in lime-soda softening at the usual
Proper selection of specific reagents to ef- specification of 58% Na2 0 content.
ficiently precipitate part of the impurities UsE OF LIME AND SoDA AsH. Hydrated lime
to be removed in the form of an insoluble reacts to chemically precipitate the carbonate
sludge and to convert the remainder to a hardness present in water, thereby producing
soluble form. insoluble precipitates of calcium carbonate-
Correct proportioning of the specific chem- CaCOa - and magnesium hydroxide -
ical reagents to the raw water. Mg(OH) 2-without the production of any
·Propagating the chemical precipitating re- soluble by-products. These suspended precipi-
actions by a suitable combination of coagu- tates are removed by sedimentation and fil-
lation, sedimentation, heat or sludge con- tration and thus a marked reduction in solids
tact. content is obtained.
Filtration of the chemically softened or The carbonate hardness of a water is
finished water to remove traces of turbidity composed of the carbonates and bicarbonates
or suspended matter resulting from the of calcium and magnesium. Their precipita-
process. tion by means of hydrated lime proceeds in
Lime-soda softening is employed for the accordance with the following reactions:
removal of hardness in order to minimize (2)
scale and sludge in boilers, reduce calcium Ca(HC03h + Ca(OHh =
carbonate deposition in heat exchange sys- calcium bicarbonate + calcium hydroxide =
tems and cooling water systems, and in gen- 2CaC03 + 2H 20
eral to remove hardness wherever hardness calcium carbonate + water
in water is a source of difficulty either in in- (3)
dustrial process water, boiler feedwater or Mg(HC03h + 2Ca(OHh
city water supply. magnesium bicarbonate + calcium hydroxide
Incidental to the removal of hardness in Mg(OHh + 2CaC03
the proper operation of lime-soda softening, = magnesium hydroxide + calcium carbonate
iron, free carbon dioxide and turbidity are + 2H20
removed. +water
Hydrated lime also reacts with the mag-
CHEMISTRY INVOLVED nesium salts constituting a portion of the
non-carbonate hardness of water. In these
Hydrated lime-Ca(OH) 2-of 90% purity reactions, insoluble magnesium hydroxide-
is usually employed in hot process lime-soda Mg(OH)2-is produced. This precipitate is
softening, or in cold process softening where removed in softening process by sedimenta-
the quantity of water softened is not excep- tion and filtration. However, a soluble by-
tionally large. In cold process systems where product of calcium sulfate, calcium chloride,

etc., is formed as illustrated below: tation, a reduction in solids content will re-
(4) sult. If, on the other hand, the hardness
MgS04 + Ca(OHh content of the water is chi~fly in the non-
magnesium .sulfate + calcium hydroxide = carbonate form a marked reduction in the
Mg(OHh + CaS04 solids content of this supply through lime-
magnesium hydroxide + calcium sulfate soda softening will not occur.
(51 It will be noted that two different forms
MgC12 + Ca(OHh = of precipitates are produced in the lime-soda
magnesium chloride+ calcium hydroxide = process. These are calcium carbonate-CaC0 3
Mg(OHh + CaCh -and magnesium hydroxide-Mg(OH) 2 • In
magnesium hydroxide + calcium chloride contrast with the crystalline nature of cal-
Such soluble by-products of calcium sul- cium carbonate, magnesium hydroxide is
fate, calcium chloride, etc., must be removed gelatinous in character and may serve as a
in the softening process, since these salts, to- coagulating agent aiding the settling of these
.. gether with the calcium sulfate and the cal- precipitates in the sedimentation tank. Mag-
cium chloride naturally present in a water, nesium oxide is frequently employed as a
constitute the calcium non-carbonate hard- coagulant in both hot and cold process lime-
ness. soda softening. In addition, the magnesium
hydroxide precipitate, formed through the
Calcium sulfate, calcium chloride, etc.,
use of this coagulant or formed from the
whether produced as by-products in the
precipitation of magnesium naturally present
above reactions or naturally present in a
in the raw supply, possesses the property of
water, are chemically precipitated as calcium
adsorbing soluble silica from solution, thereby
carbonate-CaC0 3-by the use of soda ash
effecting its removal.
(sodium carbonate) as follows:
(6) UsE oF LIME. For certain applications, com-
CaS04 + Na 2C03 plete treatment with lime and soda ash is
calcium sulfate + sodium carbonate = not required since it may be advisable to re-
CaC03 + Na2S04 duce only the calcium or the alkalinity of
calcium carbonate + sodium sulfate the raw water. Under such circumstances,
(7) soda ash may not be needed and the use of
CaCh + Na 2C03 = lime alone may suffice. In the hot lime-hot
+ sodium carbonate =
calcium chloride ion exchange process lirne only is used to re-
CaC03 + 2NaCI duce alkalinity and calcium content prior to
calcium carbonate + sodium chloride the ion exchangers.
In the above reactions while the calcium Other applications of lime softening are in
content is precipitated as insoluble calcium city water plants where the objective is the
carbonate, soluble by-products in the form of reduction of hardness at the most economical
sodium sulfate and sodium chloride are pro- cost, in the beverage industry where high
duced. While the use of soda ash in lime-soda alkalinity is a prime objection and in the
softening effects removal of the hardness, a treatment of cooling tower makeup water
decrease in solids content is not obtained be- where calcium bicarbonate hardness may be
cause of the production of soluble sodium the chief scale forming factor. In these ap-
salts. The previously given reactions with plications, cold process lime softening is em-
lime, however, do result in a reduction of the ployed and the use of a coagulant with the
solids content. In the application of lime- lime is usually recommended.
soda softening to a given water supply, there- The degree to which hardness reduction
fore, it is evident that if the hardness content can be secured with lime alone depends on
of that water is chiefly in the carbonate the bicarbonate and calcium content of the
form, which requires lime for its precipi- water. Non-carbonate calcium hardness re-

quires soda ash for its removal and cannot even under these conditions, but the objec-
be removed though the use of lime. Non- tives of the treatment have been secured.
carbonate magnesium hardness, as shown by Calcium has been reduced to a relatively low
equations (4) and (5) can be reduced by value with a minimum of chemical treatment
lime, but only at the expense of adding ad- cost.
ditional non-carbonate calcium hardness. The third column illustrates still further
Therefore, in softening with lime alone, use is reduction in calcium, but to secure this re-
made only of the reactions shown in equa- duction it has been necessary to add soda ash
tions ( 2) and ( 3 ) . to further force calcium precipitation. The
Lime softening and selective calcium sof- additional calcium removal has been achieved
tening are not necessarily synonymous. Selec- at the expense of slightly higher softener ef-
tive calcium softening, as the name implies, fluent alkalinity.
is used where it is desired only to reduce the Complete lime-soda softening, including
calcium content without using additional reduction in magnesium, is shown by the
chemical for removal of magnesium. Lime fourth column. To secure this reduction in
softening will reduce the carbonate calcium magnesium it has been necessary to again in-
hardness, but soda ash is required for re- crease the softener effluent alkalinity and to
moval of non-carbonate calcium hardness. employ a considerable increase in lime and
Therefore, if it is desired to precipitate cal- soda ash.
cium as completely as possible, it is sometimes Each of the chemical balances illus.trated
necessary to also employ soda ash in selective is the most desirable for certain specific uses
calcium softening. of the treated water. Other variations are
Fig. 11-1 also possible. For example it may be desired
Cold Process Softening Balances to secure maximum alkalinity reduction, or
to secure maximum magnesium removal even
Lime Calcium at the expense of additional alkalinity in the
Softening Softening Lime-
for With Lime Soda Ash effluent.
Raw Calcium and So.ften-
Water Removal Soda Ash ing UsE oF CAUSTIC SoDA (somuM HYDROXIDE).
Under some circumstances, caustic soda (so-
Total Hardness as CaC03 . 160 80 69 52 dium hydroxide) can be substituted in place
Calcium as CaC03 ...... 110 35 24 24
Magnesium as CaC03 ... 50 45 45 28 of lime and soda ash if the carbonate and
"P" Alkalinity as CaC03 . 0 17 22 60 non-carbonate hardness of the raw water are
"M" Alkalinity as CaC03 . 130 50 60 82 in the correct proportions. The reaction of
Sulfate as S04 ......... 48 48 48 48 caustic soda with calcium bicarbonate hard-
Chloride as Cl .......... 16 16 16 16
ness is illustrated below:
Chemical Treatment Requirements (8)
Ca(HC0 3)z + 2NaOH ~
lime . . . . . . . . . . . . . . . . . . 98 98 156 calcium bicarbonate+ sodium hydroxide =
22 64
Soda Ash ............. .
CaC03 + Na2C03 +
calcium carbonate + sodium carbonate +
All values in parts per million.
Figure 11-1 illustrates the chemical bal- water
ances obtained in softening the same water Soda ash (sodium carbonate) is produced
with different objectives. The second column as a by-product in this reaction which be-
shows anticipated softener effluent balances comes available for the precipitation of cal-
where only lime is used in the softening proc- cium non-carbonate hardness as shown by
ess and where it is desired to remove only the equations (6) and (7). The use of caustic
calcium and not the magnesium. A small soda for softening is rather limited because
amount of magnesium will be precipitated of the necessity for a definite relationship

between the carbonate and non-carbonate PROPAGATING THE CHEMICAL

hardness in the raw water and the lack of PRECIPITATING REACTIONS
flexibility in caustic soda treatment if the
raw water is subject to material fluctuations. The chemical precipitating reactions are con-
sidered to be instantaneous. This fact is not
evident in the practical operation of softeners
alkaline waters containing appreciable quan-
because all of the insoluble reaction products
tities of sodium bicarbonate-NaHC0 3-are
(sludge) are not immediately precipitated in
characteristic of certain sections of the coun-
a form permitting rapid and efficient sedi-
try. This type of water possesses a total alka-
mentation or separation from the water. A
linity considerably greater than the total
portion of the sludge formed is comprised of
hardness. To condition this type of water
relatively large floes or crystals, readily sepa-
(particularly for boiler makeup) lime and
rated by sedimentation. The remainder of
gypsum are employed. These waters do not
the sludge will be of smaller particle size and
require the use of soda ash. Lime reacts to
will not readily settle. Some of these particles
precipitate the carbonate hardness as in
are present as colloids and are difficult to
equations ( 2) and ( 3) . In the presence of
separate from the water. It is necessary to
sodium bicarbonate, lime reacts to precipi-
coagulate these smaller particles into agglom-
tate calcium carbonate and to form sodium
erates capable of producing a sludge of the
carbonate as shown:
required characteristics.
2NaHC0 3 + Ca(OHh If a sufficient amount of magnesium is
sodium bicarbonate + calcium hydroxide = naturally present in the raw water, in most
CaC03 + Na2C03 + cases satisfactory coagulation results because
calcium carbonate + sodium carbonate + of the coagulating properties of the magne-
2H20 sium hydroxide precipitated in the softening
water process. In the absence of sufficient natural
For the removal of the sodium carbonate coagulation, recirculation of softener sludge
thus formed, gypsum (calcium sulfate) is em- or the addition of coagulants such as mag-
ployed to precipitate calcium carbonate as nesium oxide, sodium aluminate, or ferric sul-
follows: fate should be employed.
(10) Temperature is an important factor di-
NazC03 + CaS04 rectly affecting the rate of sedimentation
sodium carbonate + calcium sulfate =
which is in turn governed by the density of
CaC03 + Na S042
the water and the time required to effect
calcium carbonate + sodium sulfate
optimum particle size. At higher tempera-
It will be noted that the use of gypsum for tures, the density of the water is less and
the removal of sodium bicarbonate does not more rapid settling of the sludge is permitted
decrease the solids content of the water but in the absence of convection currents. Agglom-
instead replaces the sodium bicarbonate with eration of smaller particles into larger size
an equivalent amount of sodium sulfate. In particles capable of rapid settling is acceler-
most cases this sodium sulfate is less objec- ated by heat. For example, conventional cold
tionable than the sodium bicarbonate. process lime-soda softening usually requires
In a few rare occasions barium salts have a sedimentation time of 4 hours. At the tem-
been employed in order to precipitate and perature of hot process operation ( approx.
remove sodium sulfate in the form of in- 212 F), a one hour sedimentation period is
soluble barium sulfate. Barium salts are rela- sufficient. For consistent results, however>a
tively expensive and poisonous. Because of constant temperature must be maintained.
these limitations, this process is rarely em- Recirculation of sludge provides a contact
ployed. medium for proper precipitation. The use of

lime and soda ash in the cold results in the vantages are the long sedimentation period
formation of a super-saturated solution, or, required, the considerable manual operations
in other words, the water contains more cal- necessary and the fact that the effluent is
cium carbonate in solution than would nor- super-saturated with calcium carbonate and
mally be dissolved in that water at a given is unstable, particularly with increase in tem-
temperature. If, however, particles of cal- perature.
cium carbonate sludge are brought into con- CoNTINUous SEDIMENTATION TYPE LIME

tact with water super-saturated with calcium AND SoDA SoFTENERS. Continuous sedimen-
carbonate, crystallization of calcium carbon- tation type units may be of basin design simi-
ate from solution will occur on these par- lar to coagulation and sedimentation basins.
ticles. Units of this type are frequently employed
EQUIPMENT EMPLOYED for city water softening. The raw water is
thoroughly and continuously mixed with the
Lime-soda softeners may be generally classi- proportioned lime, soda ash and coagulant in
fied as follows: a flash mixer consisting of essentially a pro-
Cold Process Softening. peller type agitator confined in a relatively
a. Intermittent or Batch. small chamber. Effluent from the mixer flows
b. Continuous-Sedimentation Type. through a flocculator basin, employed to in-
c. Continuous-Sludge Contact Type. sure formation of proper par~icle size by the
Continuous Hot Process Softening. action of slowly rotating paddles. The water
INTERMITTENT OR BATCH CoLD PRocEss then flows to a basin designed to permit ef-
SoFTENERS. Intermittent softeners are oper- fective sedimentation of the sludge for ap-
ated by the fill and draw method. The sys- proximately a four hour period. Various types
tem consists of two or more vertical tanks. of sludge collection systems are employed.
One tank is employed to supply softened Advantages of this system are its ready
water while the other tank is being cleaned, adaptation to any cold process softening
filled, treated and settled. In some cases only problem, the minimum of maintenance and
one tank is provided if sufficient storage ca- supervision required and the fact that the
pacity is available to permit repeated soften- system may easily be converted to cope with
ing cycles without interruption of a soft water taste and odor control by pre-chlorination or
supply. super-chlorination. Disadvantages are the
The system is operated on a definite cycle. large space requirements, relatively high con-
The calculated quantities of lime, soda ash struction cost and the fact that the effluent
and coagulant are made into a slurry in a must be recarbonated to achieve stability.
chemical mixing tank and added to the tank Chemical requirements are comparable to
while it is being filled with raw water. The other cold process methods discussed.
full tank is agitated for a period varying from CoNTINuous SLUDGE CoNTACT TYPE. Im-
15 minutes to one hour, agitation stopped proved results in cold process lime-soda sof-
and the water permitted to settle for as long tening have been made possible by the
a time as available. The clarified water is development of sludge contact units. Basi-
drawn off by means of a swingpipe floating at cally, these units operate on the principle of
the surface and discharged to filters usually crystallization from a super-saturated solution
of the gravity type. by providing a contact medium for proper
The intermittent or batch softener is nor- precipitation of the sludge. The raw water
mally employed for small plants. It comprises is intimately mixed with previously precipi-
the original softener design and is rarely tated sludge and with the lime, soda ash and
recommended for new construction. Advan- coagulant. Initial precipitation of calcium
tages of this system are its low initial cost and and magnesium occurs and joins the slurry
comparatively simple operation. Its disad- pool of previously formed precipitates of cal-

cium and magnesium by contact. In this specified effects appreciable savings in equip-
manner, precipitation and equilibrium are ment cost, construction and floor space when
quickly established. The newly formed solids compared with other cold process methods.
do not separate individually as small parti- Chemical requirements are low and com-
cles, instead the fresh precipitate deposits on parable to the requirements of hot process
the slurry already present by accretion. The softeners. The hardness of the finished water
nature of this sludge permits ready separation is lower than can be obtained with other cold
of the treated water from the slurry. A defi- process units employing a given amount of
nite line of separation exists between the chemical.
slurry bed and the clarified water with the Sedimentation chambers are eliminated in-
treated effluent virtually filtered through a asmuch as all sludge particles are kept in
bed of its own sludge. The entire volume of suspension due to velocity of upward flow.
the tank is slowly but constantly circulated A gradual decrease in velocity permits a
by an agitator. definite point of separation of sludge and
One design of sludge contact unit is shown clarified water. The sludge bed is maintained
by Figure 11-2. Softeners of the sludge con- at definite density by a constant sludge bleed
tact type are rated in capacity much the same and therefore expensive sludge collection
as filters. Units are available to produce from systems are eliminated. Another design of this
1.0 to 3.0 gallons per square foot of surface type is shown in Figure 11-3.
area per minute, dependent on the design. CoNTINuous HoT PRoCEss LIME-SooA SoF-
Turbidities of the finished water of 5.0-10.0 TENING. The hot process softener is normally
units (without filtration) are not uncommon. employed as a specific external treatment for
This is much lower than obtained by most boiler feedwater. The units are designed to
other methods of cold process softening. The operate at temperatures of 212 F or above
maximum retention time of 1 hour usually and employ live or exhaust steam as the




Courtesy -Pfaudler Permutit Inc.

Figure 11-2 • Sludge Contact Softener









Courtesy Graver Water Conditioning Co.

Figure 11-3 • Sludge Contact Softener

source of heat. The design of one type of hot blown down to remove the precipitates and
process unit is shown in Figure 11-4. Other the clarified and softened water is drawn off
types of different design are procurable, but through the internal take-off and subjected
the softening principles are similar in all. to pressure filtration. Since silica is quite
While varying slightly, dependent upon the soluble in hot alkaline solutions, anthracite
particular. design of different manufacturers, or some other non-siliceous material is nor-
in a hot process softener the cold raw water mally specified as a filter medium. Filters are
enters through a float controlled regulating backwashed with hot softened water from
valve to maintain a given water level in the the sedimentation chamber and backwash
sedimentation tank. The flow of raw water water is returned to this chamber. No ap-
to the softener actuates a proportioning de- preciable thermal loss is entailed in back-
vice to continuously feed lime, soda ash and washing, therefore, and both chemicals and
coagulant. The raw water is heated to or · treated water are conserved.
above 21'2 F in a heater which may be of a Hot process softening possesses the advan-
tray, jet or deaerating design. Chemical re- tage of treating large quantities of water in
actions to form calcium carbonate and mag- a relatively small compact unit. A retention
nesium hydroxide are practically instanta- time of 1 hour is normally specified. Maxi-
neous and due to the effect of the higher mum efficiency of chemical treatment is ob-
temperature sedimentation takes place rap- tained by virtue of the high operating tem-
idly. The sludge collecting cone is periodically perature. Economies in chemical treatment









Courtesy Cochrane Division-Crane Company

Figure 11-4 • Downflow Type Hot Process Softener Figure 11-5 • Up.fiow Sludge Blanket
Type Hot Process Softener

can frequently be effected by recirculation of secured in the sludge blanket aids the effi-
softener sludge. ciency of silica removal. Softening and tur-
In recent years, the sludge contact princi- bidity removal are also favored by this design.
ple, originally developed in cold process work, The height of the sludge blanket can be de-
has been applied to hot process softening. termined by use of sampling cocks at differ-
Sludge blanket hot process softeners are par- ent positions on the sedimentation tank. A
ticularly applicable where silica reduction is disadvantage of the sludge blanket design is
required. Sludge blanket units differ from the difficulty in maintaining a uniform sludge
conventional designs in that, as shown by blanket, without carryover, under varying
Figure 11-5, the raw water, mixed with the load conditions.
added chemicals, is delivered to the bottom The hot process lime-soda softener can be
of the tank by means of a central downtake. used to combine in one complete unit the
Flow is reversed at this point and the water functions of:
rises slowly up through a blanket of previ- Softening by removal of calcium and mag-
ously formed sludge. The intimate contact nesium salts.

Silica removal through the use of magne- LIMITATIONS

sium salts.
Complete de aeration of the boiler feed- The lime-soda softening process does not re-
water. duce the hardness of a softened water to as
Filtration for the removal of suspended low a value as does the ion exchange soften-
matter. ing ~rocess. Hardness can be reduced to ap-
proximately 10 to 30 pprn as calcium car-
The effluent from the hot process softener
bonate dependent upon the temperature of
is low in residual hardness and is relatively
operation and the alkalinity maintained on
stable. Marked increases in temperatures such
t?e softener efflmmt. Consequently, \vhere
as in economizers, closed heaters, etc., will
hme-soda softened makeup water is em-
cause some after-precipitation. Under these
ployed for boiler feedwater, internal chemi-
conditions, stabilization of the softener effiu-
~al. treatment requirements are higher than
~nt by acidification or surface active agents
1f 10n exchange softened makeup \Vater is
lS necessary to prevent objectionable deposits.
employed. Because hardness can be reduced
to no lmvcr than 10 to 30 ppm this process
APPLICATIONS of softening offers little value where raw
water hardness is less than 50 ppm.
Lime-soda softening is the type of treatment
most suitable to the softening of high hard- Particularly in cold process softening, the
ness supplies, particularly those of predomi- use of special coagulants may be required to
nately carbonate hardness. Marked reduction obtain the proper softening and hardness re-
in solids content results where the raw water moval and to avoid formation of complex
contains high carbonate hardness. soluble forms of hardness.
Iron or suspended matter in the raw water On waters of high non-carbonate hardness
can be removed by lime-soda softening and there will be no decrease in solids content ef-
pre-filtration of such supplies is not required. fected by the lime and soda process and on
\Vhere the raw water contains appreciable raw waters containing only a small amount
color and dissolved organic matter, through of natural alkalinity, an increase in solids
the .use of proper coagulant it is frequently may result.
possible to effect a marked reduction in the . C?ntrol of lime-soda softening, in general,
organic content. IS slightly more difficult than the control of

Lime softening alone, without the use of ion exchange softening and more frequent
soda ash, can be employed advantageously testing of the softener effluent may be re-
quired with more frequent adjustment of
~here it is desired to reduce only the calcium
bicarbonate content. In the cold, typical ap- control.
plications are in the treatment of makeup REFERENCES
water for cooling systems, preparation of
V. J. Calise, "Factors in Hot-Process Softener De-
?rocess wate~ for the beverage industry and sign and Operation", Combustion, VoL 22, pp. 49-
m the softenmg of city supplies. In the hot, 55 (Aug. 1950)
lime softening provides an efficient method A. A. Kalinzke and J. M. Kahn "Applying Steam
Lift Circulation to Hot-Process S~ftcncrs" Proceed-
for reducing alkalinity and solids prior to ings, Midwest Power Con£., Vol. XII pp. 71-78
hot ion exchange units. (1950) '
Hot process lime-soda softening provides E. Nordell, "Water Treatment for Industrial and
Other Uses", 2nd Ed., pp. 489-54 7, Reinhold Pub-
the ideal conditions necessary for the most lishing Co., New York, N.Y. ( 1961)
economical and efficient removal of silica S. T. Powell, "Water ~onditioning for Industry",
from boiler· feedwater. In cases where raw pp. 84-145, McGraw-H1ll Book Co. Inc. New York
N.Y. (1954) ' ' '
water hardness is relatively low, magnesium
J. D. Yoder, "The Sludge Blanket Hot Process
oxide and lime may be used primarily for the Softener", Engineers Soc. of Western
value in silica removal rather than softening. Penna., pp. 65-80 (

Hot Process
Phosphate Softening

Chemical Reactions. . . . . . . . . . . . . . . . . . . . . 79
Equipment Requirements . ................... 80
Applications and Limitations . . . . . . . . . . . . . . . . 80

OT phosphate softening is a process acid itself may be used depending upon the
H whereby the calcium and magnesium
salts constituting the hardness of water are
characteristics of the water to be treated. For
control of chemical treatment, an excess of
chemically precipitated and removed through 5-10 ppm phosphate is maintained in the
the use of phosphate, normally disodium softener effluent. A pH of approximately 9. 7
phosphate, in conjunction with caustic soda. is required for precipitation of calcium, but
The calcium hardness is precipitated as tri- removal of magnesium and silica requires a
calcium phosphate while the magnesium slightly higher pH approximating 10.1. The
hardness is precipitated as magnesium hy- softening reactions may be represented as
droxide. The softening operation is carried follows:
out in the hot at temperatures of 212 F or (1)
above. Since both the calcium and magne- 3Ca(HC03h + 6NaOH 3CaC03 +
sium precipitates formed are very insoluble, a calcium sodium calcium
water of practically "zero" hardness is pro- bicarbonate+ hydroxide carbonate +
duced by this softening process. JNa2C03 + 6H20
Hot process phosphate softening is em- sodium + t
carbonate wa er
ployed as a specific form of external treat-
ment for boiler feedwater-particularly for (2)
high pressure boilers. 3CaC03 + 2Na3P04 = Ca3(P04h +
calcium + trisodium tricalcium +
This process is applicable to very soft carbonate phosphate = phosphate
waters of the order of hardness not exceed-
ing 60 ppm. It is part'icularly well adapted sodium
to turbid waters low in hardness where coag- carbonate
ulation and softening can be brought about It can be noted from equation ( 1) that
in a single apparatus. calcium bicarbonate is first converted to cal-
Where waters are high in hardness this cium carbonate which is fairly insoluble as
system may also find application by employ- discussed under lime-soda softening. In the
ing two stage softening. Thus, a greater por- second equation the calcium carbonate is
tion of the hardness may be removed econom- converted to the more insoluble tricalcium
ically by lime-soda softening as the first phosphate. The magnesium hardness is pre-
stage and the remaining hardness may be re- cipitated as the hydroxide in accordance with
moved by subsequent phosphate softening in the following equation:
the second stage. (3)
Soft waters high in silica, likewise are Mg(HC03h + 4NaOH = Mg(OHh +
suited to this process. Most efficient silica re- magnesium + sodium = magnesium +
duction is accomplished in the hot phosphate bicarbonate hydroxide hydroxide
softener by magnesium sulfate. It is therefore 2Na2C03 + lH20
possible to decrease silica to a low value and sodium + water
at the same time accomplish softening to a carbonate
"zero" hardness. The tricalcium phosphate and magnesium
Where a reduction in alkalinity is desired, hydroxide precipitates which form are re-
particularly to insure a low carbon dioxide moved in the softening process by means of
content in the steam, phosphoric acid and/or sedimentation and filtration.
sulfuric acid can be used prior to softening. From the above reactions it can be noted
that a soluble by-product of sodium carbon-
CHEMICAL REACTIONS ate alkalinity results, increasing the natural
alkalinity from one equivalent on the un-
Any of the various phosphates, from the al- treated water to two equivalents on the
kaline trisodium phosphate to phosphoric treated water. Where the natural alkalinity

is already high this condition is undesirable although a retention period of only 45 min-
and may be offset by the use of an acid utes or less has been used with phosphate
phosphate or by use of sulfuric acid. Alka- softening compared with the 1 hour required
linity reduction by means of phosphoric acid for hot process lime-soda softening.
is illustrated as follows:
Ca(HC03h + H3P04 CaHP04
calcium phosphoric_ calcium
bicarbonate+ acid -acid phosphate
+ 2C02 + 2H20
+ d'IOXI'd.e + water
The carbon dioxide thus formed may be
removed by deaeration, thereby minimizing
the carbon dioxide content of the steam.
Following deaeration, the soluble calcium
acid phosphate is precipitated by the addi-
tion of sodium hydroxide. This reaction is Figure 12-1 • Single Tank Two Stage Softener-Hot
illustrated as follows: Lime-Soda Followed by Hot Phosphate
While a calcite filter medium can be used
3CaHP04 + 3NaOH = Ca3(P04h in conjunction with lime-soda softening, re-
calcium sodium tricalcium
acid phosphate+ hydroxide= phosphate action of calcite with a phosphate softener
effluent will result in cementing the entire
+ Na3P04 + JH20 filter bed. Anthracite coal is the most desir-
+ trisodium + water
phosphate able filter medium with phosphate softening.
All of the preceding reactions illustrate The two stage water softeners may take
the removal of carbonate hardness. The re- one of several forms as, for example, the in-
moval of non-carbonate hardness proceeds in corporation of two stages in a single sedimen-
accordance with the following reactions: tation tank as illustrated by Figure 12-1.
3CaS04 + 2Na3P04 = Ca 3(P0 4h +
calcium+ trisodium _ tricalcium + Hot process phosphate softening is primarily
sulfate phosphate- phosphate
useful in the treatment, for boiler feedwater
JNa2S04 purposes, of turbid waters that are low in
sodium hardness and alkalinity. It is possible with
hot phosphate softening to combine turbidity
removal, softening and deaeration in a single
MgCh + 2NaOH = Mg(OHh unit. Where silica removal is also necessary,
magnesium sodium _magnesium+
chloride + hydroxide - hydroxide hot phosphate softening provides a single
2NaCI unit to accomplish all of these objectives.
sodium Through the use of phosphoric or sulfuric
chloride acid, reduction can be secured in potential
carbon dioxide content of the steam.
EQUIPMENT REQUIREMENTS The effluent from a hot process phosphate
softener, while quite low in hardness, will
In general the equipment employed for hot cause after-precipitation when subjected to
process phosphate softening is very similar to marked increase in temperature, such as en-
that used in hot process lime-soda softening, countered in economizers, closed heaters, etc.

Stabilization of the softener effluent by acid Until the development of the styrene base
or the application of organic surface active ion exchange resins, the combination of hot
agents is necessary to prevent objectionable lime-soda softening, followed by hot phos-
deposits. phate softening, was considered · the best
The fine particles of calcium phosphate method for producing a low hardness boiler
formed in the sedimentation tank do not feedwater from a hot process treatment
settle as rapidly as the precipitates formed in plant. However, as discussed in the chapter
lime-soda softening. Greater difficulty is en- on hot lime-hot ion exchange softening, this
countered in maintaining a low turbidity new process offers several advantages over
softener effluent with hot phosphate soften- the hot lime-soda hot phosphate process.
ing. The presence in the raw water of or- With the hot ion exchange method, more
ganic matter, which would tend to retard the complete hardness removal can be secured
precipitation reactions, may result in rela- than with hot lime-soda hot phosphate sof-
tively poor hardness removal and plugged tening, together with a lower carbon dioxide
filters. content of the steam.

The following is a list of the common commercial forms of phosphate that may be used
for hot phosphate softening. In selecting the type of phosphate to be employed there are two
important factors to consider: (a) cost per pound P 2 0 5 and (b) alkalinity of the water to
be treated. For example, although anhydrous monosodium phosphate has a greater P 2 0s
content than anhydrous disodium phosphate, the cost is higher and it is generally more eco-
nomical to purchase the latter phosphate. However, if the alkalinity of the water to be treated
is very high, it may be desirable to use anhydrous monosodium phosphate, which is acidic, in
spite of its greater cost so as to reduce the alkalinity of the treated water. The general state-
ment can be made that for the majority of cases anhydrous disodium phosphate is the most
satisfactory phosphate to be employed. However, each specific application should be studied
to determine this point.

Fig. 12-2
Supplementary Data-Hot Phosphate Softener

Lbs CaC03 NaOH Required
Removed to Convert 100
Per 100 tbs lbs of the Phos-
Name of Phosphate Chemical Formula %P20s of the Phosphate phate to Na3 P04*

Monosodium Phosphate, Anhydrous ........•.. NaH 2 P0 4 58% 122 lbs. 671bs.

Monosodium Phosphate, Monohydrate ......... NaH 2P0 4 .H 20 51% 108 lbs. 58 lbs.
Disodium Phosphate, Anhydrous .............. Na 2HP0 4 48•/. 101 lbs. 28 lbs.
Disodium Phosphate, Crystalline .............. Na 2 HP0 4.12H 2 0 19% 40 lbs. 11 lbs.
Trisodium Phosphate, Monohydrate ........... Na 3 P04.H 20 39% 82 lbs. 0 lbs.
Trisodium Phosphate, Crystalline .............. Na 3P04 .12H 20 19% 40 lbs. 0 lbs.
Phosphoric Acid, 75% Solution ..........•••. H3 P04 54% 114 lbs. 921bs.

* This value is the theoretical NaOH requirement for each phosphate used. However, the natural alkalinity or acid-
ity of the water to be treated will vary the total NaOH requirement.


E. W. Ellis and H. E. Carlson, "Lime-Soda- gineers Soc. of Western Penna., pp. 79-91 (1941)
Phosphate Feedwater Treatment", The Petroleum F. N. Kemmer, "Conditioning Makeup by High-
Engineer, Vol. 26, pp. C29-32 (June 1954) Temp., Sodium Ion-Exchange Excess Calcium, Hot
C. E. Joos, "Ramifications of the Hot Process Lime-Zeolite Process", Combustion, Vol. 21, pp.
Method of Water Conditioning", Proceedings, En- 59-62 (April 1950)

Silica Ren1oval By
Magnesiun1 Salts
Factors Controlling Silica Removal . ........... 83
Temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
Retention Time . . . . . . . . . . . . . . . . . . . . . . . . . 83
pH .................................. 83
Sludge Recirculation . . . . . . . . . . . . . . . . . . . . . 83
Various Magnesium Compounds . ........... 84
Mechanism of Reaction. . . . . . . . . . . . . . . . . . . 85
Equipment Required. . . . . . . . . . . . . . . . . . . . . . . 86
Applications. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
Limitations.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87

HE removal of soluble silica from water The removal of silica from boiler feed
T in the hot is accomplished with the use
of magnesium compounds under relatively
water is accomplished in order to prevent or
to minimize silicate boiler scales such as cal-
simple controlled conditions of temperature, cium and magnesium silicate and also to pre-
retention time, pH and sludge recirculation. vent the formation of the complex sodium-
Hot process silica removal is carried out alumino-silicate scales such as analcite,
employing a conventional hot process lime or Na 2 0.Al 2 0 3 .4Si0 2 .2H 2 0. For the prevention
lime-soda softener. It is advantageous to thus of siliceous turbine blade deposits, it may be
combine in one unit the functions of soften- advisable to limit the silica content of the
ing and silica removal. boiler water thus involving the removal of
The magnesium compound employed may silica from the boiler feedwater.
be Epsom Salt (magnesium sulfate), dolo-
mitic lime, calcined magnesite, magnesium FACTORS CONTROLLING SILICA REMOVAL
carbonate or magnesium oxide. Since mag-
Effect of Temperature-The higher the tem-
nesium oxide possesses the greatest silica re-
perature of operation, the more efficient is
moval efficiency and does not increase solids
the removal of silica from the solution. Fig-
content, it is usually employed.
ure 13-1 illustrates the increase in silica re-
moval efficiency using magnesium oxide with
100 increase in temperature up to 95 C, the ap-
proximate temperature of operation of hot
process softeners.
Effect of Retention Time-At 95 C almost
complete silica removal is accomplished by
80 magnesium oxide in 15 minutes retention
time. At lower temperatures there is a grad-
> ual increase in the quantity of silica removed
~ 10 with increased retention. Figure 13-2 illus-
a: trates this effect of retention. Since hot proc-

~ 60
ess lime-soda softeners are designed for one
iii hour retention, it is evident that complete
<( silica removal can readily be accomplished
~ 50 in this period of time.
a:0 Effect of pH-The influence of pH on sil-
ica removal by magnesium oxide is illustrated
by Figure 13-3. It can be noted that the op-
u timum pH, in each of the three cases shown,
30 is approximately 10.1. This pH value is
readily obtained with either hot or cold proc-
ess lime-soda softening. The greater the per-
centage of silica removed the less important
is the control of pH.
ON ALL TESTS Effect of Sludge Recirculation-During the
silica removal process, if the partially spent
magnesium oxide sludge is brought back in
contact with fresh incoming water high in
silica, a considerable increase in silica re-
moval efficiency can be achieved. A portion
Figure 13-1 • Effect of Temperature of the sludge collecting in the sludge cone of

> 90
w 80
(.) 70
CJ) 60

0 8 TEMP= 95C
z e TEMP= 23C
180 240


Figure 13-2 • Effect of Retention Time

the softener is recirculated by means of a ash needed to precipitate the magnesium or

small pump to the upper section of the sedi- calcium compounds thus introduced. How-
mentation tank. Here it comes in contact with ever, the major disadvantage of their use is
the raw makeup water and, through addi- the increase in solids content of the boiler
tional silica adsorption, considerably reduces feed water.
the quantity of fresh magnesium oxide that Dolomitic lime can be used to provide
must be employed. It is possible by sludge economical silica removal where the lime
recirculation to reduce magnesium require- requirements for softening will introduce suf-
ments by as much as 60%. The increased ficient magnesium oxide to accomplish the
silica removal efficiency is shown by Figure desired degree of silica removal. Where mag-
13-4. There is no change in the chemical nesium oxide requirements for silica removal
balances maintained on the softener efHuent are relatively high and lime requirements
and recirculation affects the characteristics low, if sufficient dolomitic lime were fed to
of the softener effiuent only with respect to a secure the desired silica removal, it would be
reduced silica content. necessary to feed additional soda ash to re-
Effect of Various Magnesium Compounds act with the excess lime thus introduced. In
-A number of different magnesium com- such cases, dolomitic lime can be used in an
pounds have been investigated for use in the amount sufficient to supply lime requirements
removal of silica from water. A number of and additional magnesium oxide fed to secure
these compounds such as Epsom Salt (mag- the necessary silica removal.
nesium sulfate) and dolomitic lime may con- Magnesium oxide has the advantage of
siderably increase the solids content of the causing no increase in the solids content of
treated water. They may also require large the treated water. In addition, no lime or
increases in the quantity of lime and soda soda ash is required to effect its precipitation.

100 combining weights. The removal of silica

from the solution by magnesium compounds,
however, is not a stoichiometric reaction but
proceeds by adsorption from the solution.
Figure 13-6 illustrates silica removal plotted
against the qv.antity of magnesium oxide
used on a water supply high in silica. These
0 70
data were obtained from actual operating
w results on a hot process lime-soda softening
w system rated at 270,000 pounds per hour and

conditioning water for 450 psi boilers. It can
be noted that increased quantities of magne-
50 sium oxide do not effect proportionate in-
creases in the quantity of silica removed. In
u. 40
other words, as larger amounts of silica are
0 removed, larger and larger quantities of mag-
~ nesium oxide are required to effect removal
of the last traces of silica. This action is
~ 30

characteristic of adsorption reactions.

The empirical equation of Freundlich
closely fits adsorption data. This equation
may be expressed as:
9.0 9.4 9.8 10.2 10.6 ll.O
Figure 13-3 • Effect of pH

While different forms of magnesium oxide
vary in their silica removal efficiencies, 0
specially prepared adsorptive forms of mag- ~ 0.4
nesium oxide, have been found to possess the 1-
greatest silica removal efficiency and to be ~
the most economically feasible reagent. a::
Figure 13-5 illustrates the silica removal 0 0.3
capacity of a number of different magnesium >
compounds under the same conditions. Be- ::;;
cause silica removal by magnesium com- a::
pounds is an adsorption reaction, those ma- ~ 0.2
terials which effected over 80% silica :::i
removal in this test possess very high silica en
removal capacities.
Mechanism of Silica Removal Reactions-
In the chemical reactions usually encountered
in water conditioning there is a stoichiometric
relation or constant reacting value. For ex- 6 B 10 12 14 16 IB 20 22 24 26
ample, a definite quantity of soda ash will SILICA REMOVED, PPM
always react with a definite amount of cal-
cium sulfate in accordance with the law of Figure 13-4 • Effect of Sludge Recirculation









--'- --'- I

0 10 20 30 40 50 60 70 80 90 100
Figure 13-5 • Comparison of Various Forms of Magnesium Oxide

where K and 1/n are constants. feeding equipment is required. Magnesium

x = quantity of silica adsorbed oxide can simply be added to the same mix-
m = quantity of adsorbent ing tank where the lime and soda ash are fed.
C = residual concentration of adsorbed
material at equilibrium.
A straight line results from plotting the

logarithm of ~ against the logarithm of C

rn '
when the reaction proceeds by adsorption.
Silica removal data -thus shows the reaction
to be one of adsorption similar to the re-
moval of color by activated carbon. 0


For the most efficient removal of silica by

magnesium compounds, a combination of the 18

following factors is required:

Temperature close to 212 F. 12

Retention time, 15 minutes to one hour.

pH value approximately 10.1.
Recirculation of sludge.
These conditions are most readily secured o~--------~2S~--------5~0--------~75~
through the use of a conventional hot process
softener which provides the ideal unit for the Figure 13-6 • Field Tests on Silica Removal
silica removal process. No special chemical with Sludge Recirculation

Recirculation of sludge is simply provided by When magnesium oxide is employed in

tapping the sludge cone and circulating a conjunction with hot process lime-soda sof-
portion of the sludge to the top of the sedi- tening, the additional advantages secured in
mentation tank by means of a small centrif- addition to the above are:
ugal pump. No additional capital expenditure is re-
While silica removal can be secured in con- quired since, both softening and silica re-
junction with hot lime or hot lime-soda sof- moval are accomplished in the one opera-
tening, the use of lime and soda ash is not tion, requiring no additional equipment.
necessary for the successful operation of the Quantities of lime and soda ash are not
silica removal process. On a water low in increased by the use of magnesium oxide,
hardness, a conventional hot process softener unlike the use of Epsom Salt, dolomitic
can be employed for silica removal feeding lime or iron salts which require lime and
only caustic soda and magnesium oxide. soda ash for their precipitation.
No increase in the hardness or alkalinity
of the softened water is occasioned by the
use of magnesium oxide.
Hot process silica removal by magnesium
compounds can be successfully applied to any LIMITATIONS
type of water. The use of magnesium com-
pounds is advantageous with the hot lime-hot In order to efficiently accomplish the removal
zeolite process, providing removal of silica of silica by magnesium compounds and to
to as low as 1 ppm in conjunction with hard- comply with the conditions of temperature,
ness removal to the range of 0.5-2.0 ppm. retention time and sludge recirculation, a hot
Magnesium oxide is preferred as the -silica re- process softener complete with chemical pro-
moval agent in hot lime or hot lime-soda portioner, heater, sedimentation tank, filters
softening, while more efficient results are and backwash facilities is required.
secured in hot phosphate softening by the use Where such equipment is already installed,
of magnesium sulfate. silica removal can be very simply applied.
Silica removal is particularly desirable in Where a hot process softener is not already
conditioning water for use in high pressure installed in a plant, its installation is necessary
boilers both from the standpoint of reducing before silica removal by magnesium com-
scale deposits in the boilers and also to re- pounds can be accomplished.
duce siliceous deposits on turbine blading.
When the magnesium compound employed REFERENCES
is magnesium oxide, the additional advan- L. D. Betz, C. A. Noll and J. J. Maguire, "Re-
tages secured are : moval of Silica from Water by Hot Process", Ind.
No increase in the solids content of the and Eng. Chem., Vol. 32, pp. 1323-1329 (1940)
L. D. Betz, C. A. Noll and J. J. Maguire, "Adsorp-
treated water. tion of Soluble Silica from Water", Ind. and Eng.
More silica is removed per part of reagent Chem., Vol .33, pp. 814-820 (1941)
employed than in any other method. C. A. Noll and J. J. Maguire (to Betz Laborato-
Magnesium oxide is itself an excellent coag- ries, Inc.) U.S. Patent 2,307,466 (Jan. 5, 1943)
J. D. Yoder, "Removing Silica from Boiler Feed
ulant and eliminates any need for other Water", Proceedings, Midwest Power Conf., Vol.
coagulants. XI, pp. 164-170 (1949)

Sodiulll Zeolite Softening

Zeolite Exchange Materials. . . . . . . . . . . . . . . . . 89

Equipment Employed. . . . . . . . . . . . . . . . . . . . . . . 91
Applications. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
Limitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93

term zeolite is applied to insoluble, damental reaction in sodium zeolite softening:

solid materials which have the property (1) Ca++ + Na2Z CaZ + 2Na+
of exchanging various ions with which they calcium + sodium = calcium + sodium
come in contact. The sodium zeolite soften- ion zeolite zeolite ion
ing process, also referred to as cation ex- When this reaction has proceeded to the
change on the sodium cycle, exchanges so- point where the zeolite will release no more
dium ions for all cations of two or more sodium in exchange for calcium, it becomes
positive charges, when operating on the sof- necessary to regenerate the zeolite bed. This
tening cycle. When water containing the regeneration is accomplished by washing with
hardness salts of calcium and magnesium is a strong salt solution (sodium chloride). The
passed through the cation exchange bed, the high concentration of the sodium ion in the
sodium ions of the bed replace the hardness salt solution reverses the above reaction re-
ions with a resultant effiuent water of close sulting in regeneration of the zeolite bed.
to "zero" hardness. Sodium zeolite softening This regeneration reaction may be written as
is, by far, the most common of the exchang~ follows:
reactions involved in water conditioning and (2) CaZ + 2NaCI = Na2Z + CaCI2
is generally the one referred to vvhen the calcium sodium sodium calcium
term "zeolite softening" is used without zeolite + chloride = zeolite + chloride
further qualification.
The calcium placed in solution by regen-
The primary purpose of sodium zeolite sof-
eration is run to waste as calcium chloride.
tening is the removal of the scale forming
The fundamental chemical principle of mass
ions of calcium and magnesium, replacing
action determines if the reaction proceeds as
these ions with an equivalent amount of
in equation ( 1) or equation (2). In the sof-
sodium ions. By removal of the calcium and
tening cycle, where the calcium ions pre-
magnesium ions, which normally constitute
dominate, the reaction proceeds as in ( 1) .
the hardness content of a water, this process
In the regeneration reaction, where the so-
is useful in the preparation of water for boiler
dium ions predominate, the reaction proceeds
feedwater, laundry use, various industrial
as in ( 2) . The net result is that sodium chlo-
processes and in the reduction of the hard-
ride (common salt) has been consumed and
ness content of city supplies.
hardness has been removed from solution.
ZEOLITE EXCHANGE MATERIALS As it is necessary to have an excess of salt
present for regeneration, an equivalent
Natural zeolite minerals possess a complex amount of hardness is not removed for a
chemical structure and usually consist of so- given amount of salt consumed.
dium-aluminum-silicate. Other cation ex- The capacity of zeolites, that is the amount
change materials are made synthetically and of hardness that can be removed before re-
have as a base such chemicals as sulfonated generation, will vary greatly. The synthetic
coal (carbonaceous) , phenolic resins and zeolites generally have a higher exchange
polystyrene resins. Regardless of how the capacity than the natural zeolites, but are
zeolites are produced, all operating on the more expensive per unit volume. The capac-
sodium cycle contain a complex molecule ities for synthetic zeolites range between 6000
with sodium attached. Accordingly, it is pos- grains per cubic foot and 32,000 grains per
sible to represent sodium zeolite in a chemical cubic foot. Natural zeolite capacities vary
equation by using the symbol Na 2 Z where Na between 2500 and 5000 grains per cubic foot,
stands for the sodium and Z for the complex depending on the processing. Greensand zeo-
zeolite radical. For simplification in the fol- lite is more resistant to waters of low pH
lowing reactions, the calcium ( Ca ++) in solu- than the synthetic aluminum-silicate zeolites.
tion will be considered as representing all the Greensand, carbonaceous and resinous zeo-
hardness. The following is therefore the fun- lites are less likely to increase the silica con-

tent of the softened water and, consequently,

are better applied to the softening of waters
of low silica content. Synthetic zeolites of the
sodium-aluminum-silicate type may impart
considerable silica to waters that are low in
initial silica content.
Generally, synthetic zeolites are employed
in waters of high hardness where the ex-
change value of this material is of importance
in permitting the use of a smaller unit. For
low hardness waters, greensand~the natural
zeolite~is still occasionally employed.

The effluent from a zeolite softener will

rise quite rapidly in hardness as it approaches
the limit of its capacity. In some cases, where
minimum hardness is essential, the softener
will be regenerated before it reaches this
point. In other cases, the hardness is allowed
to rise only to a predetermined level before
regeneration. As the concentration of anions
such as chloride and sulfate increase in the
Courtesy Rohm & Haas Company '"'·ater to be softened, there is the tendency
Figure 14-1 • Microscopic View of Resin for hardness leakage to take place. With the
Beads (20-50 mesh) of a Sulfonated older types of zeolite minerals and synthetics,
Styrene-Divinyl Benzene Cation Exchanger the amount of hardness leakage could not be
controlled since this leakage was a function
of the influent ion concentration. However,
with the advent of the polystyrene type res-
ins, the amount of hardness leakage can be
accurately controlled by adjusting the salt
regeneration level. For example, in a normal
water supply where the total hardness and
sodim11 salts were in the range of 700-800
ppm, using a carbonaceous zeolite the mini-
mum effluent hardness was found to be in the
range of 15 ppm with a salt regeneration rate
of 3.5 pounds of salt per pound of hardness
removed. Usmg a polystyrene resin at the
same salt rate produced an average effluent
hardness of 8 ppm. By doubling the salt rate,
the effluent hardness was reduced 2 ppm.
Figure 14-3 also illustrates the effect on
capacity at different salt regeneration levels.
From this table it can be readily understood
that sodium zeolite softeners, using a poly-
styrene base resin, can be designed for low
Courtesy Rohm & Haas Company capital investment and high operating cost or
Figure 14-2 • Microscopic View of Beads (20-50 mesh) higher capital investment with lower operat-
of a Carboxylic Exchanger Resin ing cost for the same number of gallons of

water with constant characteristics. The ap- EQUIPMENT EMPLOYED

plication of this type of resin is also advan-
tageous when used with waters having wide Natural zeolite minerals have a sandy nature
seasonal variations in hardness since all that while the synthetic zeolites are in the form of
is necessary is the adjustment of the salt rate small crystals or beads. A zeolite softener is,
to obtain softening cycles of predetermined therefore, very similar in construction and
length. appearance to ,a sand filter excepting that it
Fig. 14-3 is filled with a bed of zeolite instead of sand.
Figure 14-4 illustrates one type of zeolite
Effect of Salt Regeneration on Exchange
Capacity softener.
Although there are different modifications,
Salt Requirement Polystyrene Resin in general the softener consists of a steel
lbs sall/kilograin Exchange Capacity
hardness removed grains/cu ft shell, the bottom of which contains graded
sizes of gravel or anthracite. Above the
0.285 ............................. 18,000
0.323 ............................. 21,000 graded base is located the zeolite material,
0.385 ............................. 24,500 usually 24 inches or more in depth. In the
0.440 ............................. 27,000
0.490 ............................. 29,000
majority of softeners, the water enters the
top of the tank and passes down through the



Courtesy Cochrane Division-Crane Company

Figure 14-4 • Zeolite Softener with Single Control Valve and Header Lateral Underdrain

zeolite. The size of the zeolite bed will de- turning a valve or valves. Semi-automatic
pend upon the exchange capacity of the zeo- operation entails pushing a button for each
lite, the hardness of the water being treated, regeneration step to actuate motorized valves.
the amount of water to be softened between The motive force can be electric, pneumatic
regenerations, and the flow rate. The bed or hydraulic. Completely automatic operation
must also be of sufficient size to allow proper requires no attendance. The usual method
contact time between the water and the zeo- consists of having a water meter preset so
lite and flow rate is usually regulated at 5-8 that when a certain number of gallons of
gallons per square foot per minute. In addi- water have been softened, an electrical con-
tion, it is necessary to provide a tank to hold tact is made in the meter. This electrical con-
the salt solution for regeneration, proper con- tact actuates a system of counters and timers
trol valves and interconnecting piping. which in turn actuates the valves controlling
In regenerating the unit, it is first back- the various regeneration steps.
washed to clean the bed. In backwashing,
water is first passed up through the bed and APPLICATIONS
sent to waste until it runs clear. The rate of
backwash should be sufficient to lift and ex- Zeolite softeners have a definite application
pand the bed to thoroughly clean it, but not in any process where soft water is an advan-
high enough to wash the zeolite out of the tage. Laundries find a decided use for these
unit. The required rate is a function of the softeners to reduce their soap consumption.
type of zeolite, water temperature and free- Softening offers an added advantage in im-
board in the tank above the zeolite bed and proving the appearance of the wash as dis-
the usual range is between 4 and 10 gallons colorations, due to soap curds-those insoluble
per square foot of bed area per minute. A particles of soap formed by the reaction of
clear water, free from suspended matter, soap with the hardness in the water-are
should always be available for backwashing. eliminated. Dye houses and other textile
A strong salt solution is next passed into the plants where finishing is accomplished also
softener at the top. Rinse water is next added find these softeners beneficial for the same
to the top of the softener to force the salt reasons, eliminating many causes for uneven
water or brine through the zeolite and out dyeing.
the bottom to waste. At first, this water will In certain districts where the water sup-
contain large quantities of calcium and mag- plies are quite hard, many small zeolite sof-
nesium chlorides together with unreacted teners are used in the home. In some cities
sodium chloride. As the bed becomes regen- softeners of the zeolite type have been in-
erated, the calcium and magnesium chlorides stalled to supply the public with softened
in the effluent are eliminated, and the excess water.
salt is gradually washed out by continuing to In boiler feedwater conditioning, the use
add further· quantities of fresh water, which of zeolite softening is frequently attractive
acts as a rinse. Rinsing is continued until all because of the simplicity of operation and
of the ex<;ess salt is washed from the bed and control. An outstanding advantage of zeolite
the unit is then returned to service. Many softening is the low hardness obtainable in
units are provided with a master valve which the treated water. Marked reduction in the
controls the various operations by merely scaling properties of a water is secured by
turning a hand wheel in steps as indicated at proper zeolite softening.
points on the rim of the valve. With the use of resinous and carbonaceous
Softeners can be manufactured for man- exchangers, it is possible to operate on the
ual, semi-automatic or completely automatic sodium cycle and to secure the advantages of
operation. In manual operation the regen- a low hardness effluent without danger of
eration steps are accomplished by manually silica pickup from the zeolite bed. Even

though such exchange materials are some- should be removed prior to softening. In such
what more expensive than greensand zeolite, cases, the additional expense of an iron re-
this increased installation cost is frequently moval system is required.
justified by salt savings and by avoiding any In boiler water conditioning, zeolite has a
silica increase while softening. marked advantage over lime-soda softening
insofar as it more completely removes hard-
LIMITATIONS ness. Whereas lime-soda softening may reduce
the hardness to as low as 10-30 ppm, depend-
Sodium zeolite softening should not be con- ing on the water, zeolite softening will show
sidered as a cure-all simply because it re- "zero" hardness or very close to "zero." Never-
moves scale forming solids from the water. theless, scale formation can take place in a
Sodium zeolite softeners require very little boiler using water softened by the zeolite proc-
maintenance and are quite simple to operate. ess. Invariably, some internal treatment is
Plant operators can be trained to operate necessary in a boiler following sodium zeolite
them properly in a very brief_ period because softening in order to prevent scale formation.
of this simplicity; however, many zeolite sof- Naturally, the quantity of internal treatment
teners have been installed to operate on required after zeolite is less than that required
. waters entirely unsuited for their use. There after lime-soda softening. As the amount of
are many types of waters and conditions of precipitated sludge formed in the boiler by the
operation for which sodium zeolite softening internal treatment depends on the original
is the proper solution; however, there have feedwater hardness, it follows that the amount
been numerous installations whereby the ap- of suspended matter in a boiler following the
plication of zeolite softening alone has in- use of zeolite softening will be small. Accord-
creased cost of operation and added several ingly, there is less chance of the sludge baking
problems which could be considered more on to the surface to form a scale than with
serious than those initially present. lime-soda softening or when using internal
Turbid waters should not be passed through treatment only.
a zeolite softener as the accumulated deposits One of the principal disadvantages of so-
will coat the zeolite grains and reduce effi- dium zeolite softening is that the hardness is
ciency. Water for a zeolite softener should removed without a reduction in alkalinity or
have a turbidity of less than 10 units. City total solids. This defect can be seen from equa-
supplies are generally filtered and accordingly tion ( 1). Calcium is removed, but is replaced
are suitable for zeolite softening. Most well by two ions of sodium, weighing slightly more
water-s are relatively free from turbidity as than the original calcium. With zeolite soft-
passage through the earth usually filters out ened water, since the hardness has been re-
suspended matter. Surface waters generally moved and the calcium exchanged for sodium,
tend to be turbid and in most cases must be the entire alkalinity is free to concentrate in
filtered prior to zeolite softening. Filtration, the boiler. Unless the natural alkalinity is
to be properly effective, should be preceded relatively low in comparison with total solids,
by coagulation. With these requirements, the zeolite softening results in high boiler water
zeolite system becomes more complex due to alkalinity and high carbon dioxide concentra-
the necessity of a coagulation and filtration tion in the steam. This condition often re-
system, increasing capital cost, space require- quires high rates of blowdown to maintain
ments and operating cost. In certain cases, fil- boiler water alkalinities within reasonable
tration will be omitted, particularly in small bounds. High alkalinities are objectionable as
installations, with the understanding that re- they tend to promote both carryover and
duction in efficiency and life of the zeolite are caustic metal embrittlement ( intercrystalline
to be expected. cracking), unless special precautions are taken
Ferric iron is detrimental to zeolite and with respect to chemical treatment of the

boiler water. investigation of these factors. Zeolite soften-

Where silica is a problem, there is the defi- ing may be coupled with other methods to
nite possibility of silica being picked up from provide a satisfactory solution to a problem.
the zeolite by the water, thereby adding to Zeolites fill a very important place in water
the problem. This undesirable condition can conditioning, but should be used only with a
be overcome by using a nonsiliceous zeolite full knowledge of their limitations. Entirely
material. too many installations have been made in the
Sodiuni. zeolite softened water is normally mistaken belief that sodium zeolite softening
corrosive, with the severity of the corrosive offered a final solution to a problem. Espe-
characteristics dependent on the amount of cially in boiler feedwater conditioning, the
dissolved gases, such as carbon dioxide and proper solution to a problem is rarely achieved
oxygen. In large cold water distribution sys- by the use of sodium zeolite softening alone.
tems, and ~n domestic hot water systems, seri- However, in combination with other external
ous corroswn may result following softening and internal treatment methods, zeolites per-
unless these corrosive gases are reduced to low form a very useful water conditioning func-
values by adjusting pH, deaeration or apply- tion.
ing positive corrosion inhibitors. REFERENCES
Many of the disadvantages outlined here
for zeolite can be overcome by methods in- L. F. Collins, "A Study of Contemporary Zeolites",
Journal, Am. Water Works Assoc., Vol. 29, pp.
volving other types of zeolite operating on 1472-1514 (1937)
other than the sodium cycles. R. Kunin, "Ion Exchange Resins", 2nd Ed., John
Whether zeolite should or should not be Wiley & Sons, New York, N.Y. (1958)
S. T. Powell, "Water Conditioning for Industry"
used depends on many factors and any indi- pp. 146-165, McGraw-Hill Book Co., Inc., Ne._,;
vidual case can be judged only following an York, N.Y. (1954)

Cation Exchange by
Hydrogen Zeolite
Principles of Cation Exchange . . . . . . . . . . . . . . . . . . 96
Hydrogen and Sodium Zeolite Units in Parallel. . . . 96
Hydrogen Zeolite with Alkali Neutralization . . . . . 97
Hydrogen Zeolite with Raw Water Neutralization. 98
Equipment Arrangements. . . . . . . . . . . . . . . . . . . . . . 98
Hydrogen and Sodium Zeolite Units in Parallel. . . . 98
Hydrogen :(eolite with Alkali Neutralization . . . . . 99
Hydrogen Zeolite with Raw Water Neutralization. 99
Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
Limitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 701

YDROGEN zeolite is the name given to a ready for use. Sulfuric acid is generally used
H group of non-siliceous organic materials,
either natural or synthetic, which have the
for the regeneration process due to its low
cost, although hydrochloric acid also is fre-
property when operated in the acid cycle, to quently used, particularly when high calcium
exchange hydrogen for cations such as cal- water results in the precipitation of calcium
cium, magnesium, sodium, etc. Upon exhaus- sulfate in the zeolite bed.
tion of the hydrogen zeolite bed, it is regen- The following equations illustrate the re-
erated with acid. Non-siliceous or carbona- actions which take place when sulfuric acid
ceous zeolite may also be operated in the is used for regeneration:
sodium cycle by regenerating with sodium
chloride (salt) and- when used for sodium (5)
zeolite softening possesses some advantages
over the siliceous zeolites.
ea IZ + H2S04 = H2Z + leaMg
An example of natural hydrogen zeolites lNa2 Na2
-are the various lignites found in the Dakotas;
From reaction ( 1) it is evident that all of
while sulphonated coal, coke, and synthetic
the bicarbonates are converted into carbonic
resms are examples of the man-made prod-
uct. acid (H2 C0 3 ). This acid is very unstable in
water solution and can easily be removed by
aeration or degasification.
Thus, in addition to completely removing
When water containing calcium, magne- the hardness the bicarbonate alkalinity is also
sium and sodium ions is passed through a removed without substituting other salts in
hydrogen zeolite these ions are exchanged for its place and accordingly it is possible by this
hydrogen, and the bicarbonate, sulfate, ni- process to reduce the total solids. The re-
trate and chloride radicals are converted to moval of bicarbonate alkalinity without sub-
their respective acids, viz., carbonic acid stitution of any other ion is the outstanding
(H2C0 3 ), sulfuric acid (H 2S04 ) , nitric acid advantage in the use of hydrogen zeolite.
(HNOa) and hydrochloric acid (HCl) as Is Since these materials are non-siliceous in
shown by the following equations: nature no pick up of silica is experienced.
( 1) Furthermore, with this type of softener it is

possible to reduce to any predetermined
value not only the alkalinity due to calcium
and magnesium bicarbonates, but also the
alkalinity due to sodium bicarbonate.
While it is possible to rid the system of

I ~~ 1so4
carbonic acid by degasification, sulfuric, ni-
tric and hydrochloric acids cannot be re-
moved by this method. These acids must be

I ~~ I I
neutrali:;r:ed, otherwise severe acidic corrosion

el2 + H2Z = ~~~
Z+ 2Hel
of the metal in the system would result.
Three general methods for the neutraliza-
(4) tion of these acids are used.
PARALLEL. A pre-determined percentage of
raw water is passed through a sodium zeolite
When the hydrogen zeolite becomes ex- which is operating in parallel with the hydro-
hausted it is backwashed and regenerated gen zeolite so as to obtain the desired alka-
with acid. After rinsing, the bed is again linity in the mixed effluent. This is accom-

plished with simple automatic rate-of-flow 4 ppm of S04 as S0 4 X 1.04 =

controls. 4.16 ppm of H2 S04 as CaC0 3
The correct amount of water to pass 2 ppm of N0 3 as N0 3 X 0.81 =
through the hydrogen zeolite unit depends on 1 .62 ppm of HN0 3 as CaC03
the factors: 6 ppm of Cl as Cl X 1.41 =
8.46 ppm of HCI as CaC03
The alkalinity of the influent ( Ar) .
The free mineral acidity (FMA) of the Free mineral acidity =
14.24 ppm as CaC03
hydrogen zeolite effluent (which depends
upon the sulfate + nitrate + chloride con- Substituting these values in the formula,
tent of the influent) . we have
The desired alkalinity in the mixed efflu- 34- 15
ent (Am). % H2Z treated = 100 X 39.5
34 + 14.24
The proper proportion can readily be cal-
% Na 2Z treated = 100 - 39.5 = 60.5
culated according to the formula:
Percent H2 Z treated= 100 X - - - - - ZATION. Neutralization of the acidity of the
Ar + FMA hydrogen zeolite effluent is accomplished with
The use of this formula will be clarified by any suitable alkali, such as caustic soda
an example. Consider a raw water of the (NaOH), soda ash (Na 2 C0 3 ), etc. This
following characteristics: method is economical where the FMA in the
effluent is low or where the total volume of
Total Hardness as CaC03 = 36
water is small so that the cost of the chem-
Sulfate as S04 4
Nitrate as N03 = 2 ical used is negligible.
Chloride as Cl = 6 With the water used in the example above,
Methyl Orange Alkalinity as CaC03 = 34 the quantity of caustic soda required to es-
Calcium as CaCOl = 24 tablish a residual alkalinity of 15 ppm, as
Magnesium as CaC03 = 12 CaC0 3 is determined as follows:
Carbon Dioxide as C0 2 = 15
Add an amount of caustic soda equal to
It is desired to maintain an alkalinity of the FMA present plus enough to give the
15 ppm as CaC03 in the mixed effluent. desired alkalinity in the effluent.
We are given Ar = 34 ppm as CaCOl The amount of caustic soda required can
We want Am = 15 ppm as CaCOl
readily be calculated according to the for-
All that is required before applying the Inula:
formula is the value for FMA. Calculation
of the FMA value is based upon the fact that ppm NaOH = (0.8 X FMA) + (0.8 X Am)
the mineral acidity value is equal to the sum (The factor 0.8 in the equation is em-
of the acids of the sulfates, nitrates, and ployed to convert the FMA as CaC0 3 and
chlorides present in the water. In addition alkalinity as CaC0 3 to NaOH).
to converting the sulfates, nitrates and chlo- We have calculated FMA = 14.24 ppm, as
rides to their respective acids, they must also CaC03
be converted to a common denominator ·we are given Am = 15 ppm as CaC03
so that they can be added together; the com- ppm NaOH = (0.8 X 14.24) + (0.8 X 15)
mon denominator CaC0 3 is chosen because = 23.4
Ar and Am are already in terms of CaCOa. The advantage of this alkali neutralization
The constants necessary for these conversions method is that only a hydrogen zeolite unit
are as follows: is required. The additional investment cost
S04 - 1.04; N0 3 - 0.81; Cl - 1.41 of a sodium zeolite unit is avoided. Chemical
Therefore, the FMA value can be calcu- treatment cost is higher than with the hy-
lated: drogen-sodium zeolite units in parallel. Ac-

cordingly, the double unit process is usually generation level for the exchange capacity
employed for all systems except those where desired.
the total sulfate, nitrate and chloride content When the carbonaceous z~olite is operated
is very low or where the installation is only in the acid cycle, acid requirements for re-
of small capacity. generation will average 0.25 pound of 66°
HYDROGEN ZEOLITE-RAW WATER NEUTRA- Baume sulfuric acid per kilograin of cation
LIZATION. Neutralization of acidity of the hy- exchange, or 1. 75 pounds of acid per million
drogen zeolite effluent is accomplished by di- pounds of water treated for each part per
lution with the raw unsoftened water itself. million of cation exchanged.
The elimination of a sodium zeolite unit is
an advantage secured also by this system.
However, the hardness of the mixed effluent EQUIPMENT ARRANGEMENTS
is increased in direct proportion with the
quantity of raw water used for neutralization. HYDROGEN AND SomuM ZEOLITE UNITS IN
. This method is usually feasible only on a PARALLEL (Fig. 15-1) . Major equipment in-
water of low total hardness and high sodium cludes:
bicarbonate alkalinity. Rubber-lined steel pressure tank to con-
A non-siliceous resinous or carbonaceous tain hydrogen zeolite, with single valve
zeolite, operating in the sodium cycle, is re- control.
generated by passing a strong sodium brine Steel pressure tank to contain sodium zeo-
solution through the bed in the same manner lite, with single valve control.
as in an ordinary zeolite. The sodium salt Acid regeneration tank for hydrogen zeo-
may be sodium chloride, sodium sulfate or lite.
sodium nitrate, but sodium chloride is gen- Salt regeneration tank for sodium zeolite.
erally used because of its lower cost. This Control for automatically maintaining de-
type of zeolite softener requires the same sired ratio of flow to the hydrogen and so-
amount of sodium chloride for regeneration dium units.
as the conventional zeolite. Polystyrene type Closed wood degasifier with trays which
resins may be employed on the sodium or break up treated water into series of small
hydrogen cycle. The salt or acid requirements streams. Air from blower then comes into
for the sodium or hydrogen cycle may be intimate counter-current contact with the
quite variable and will depend on the re- water and scrubs out the carbon dioxide.





Figure 15-1 • Hydrogen and Sodium Zeolite Units in Parallel



r---'"--_., ATMOSPHERE




Figure 15-2 • Hydrogen Zeolite with Alkali Neutrali;;ation

HYDROGEN ZEOLITE WITH ALKALI NEUTRALI- counter-current contact with the water
ZATION (Fig. 15-2). Major Equipment In- and scrubs out the carbon dioxide.
cludes: A neutralizing feeder.
Rubber-lined steel pressure tank to con-
tain hydrogen zeolite, with single valve HYDROGEN ZEOLITE WITH RAW WATER NEU-
control. TRALIZATION (Fig. 15-3). Major Equipment
Acid regeneration tank for hydrogen zeo- Includes:
lite. Rubber-lined steel pressure tank to con-
Closed wood degasifier with trays which tain hydrogen zeolite, with single control
break up treated water into series of small valve.
streams. Acid regeneration tank for hydrogen zeo-
Air from blower then comes into intimate lite.



FOR <Xl2





Figure 15-3 • Hydrogen Zeolite with Raw Water Neutrali;;ation


Figure 75-4 • Laboratory Testing of Ion Exchange Resins


Control for automatically maintaining de- The cost of chemicals must be considered.
sired ratio of raw water and hydrogen In one locality, acid and salt may be expen-
zeolite flow. sive while in another they may be inexpen-
Closed wood degasifier with trays which sive. Installation, maintenance, depreciation,
break up treated water into series of small etc., all must be taken into consideration.
streams. Air from blower then comes into An engineer, specializing in water treatment,
intimate counter-current contact with the should carefully evaluate the many variables
water and scrubs out the carbon dioxide. and make the final recommendations for the
equipment which will provide optimum re-
APPLICATIONS sults with a minimum of cost.
The properly blended and degasified hy-
Cation exchange by hydrogen zeolite is drogen zeolite-sodium zeolite effluent is low
utilized in boiler feed water conditioning in hardness and alkalinity, but is saturated
principally on a water low in hardness and with oxygen and is consequently of a cor-
high in bicarbonate alkalinity. The removal rosive nature. Corrosion resistant pipe or
of sodium and bicarbonate ions by the hydro- pipe specially painted or lined may be re-
gen zeolite process constitutes the outstand- quired to handle the water and deliver it to
ing advantage of this process inasmuch as the feed water heater.
this removal is secured without the substitu-
In general the hydrogen zeolite method for
tion of other ions. It is thereby possible to
removing calcium and magnesium bicarbo-
effect marked reduction in the solids content
nate is more expensive than the lime-soda
of a high bicarbonate water.
process, but provides an effluent of lower
The hydrogen zeolite process finds exten- hardness. The operating personnel must be
sive use in the manufacture of raw-water ice carefully trained inasmuch as the handling
and in the carbonated beverage industry. To of sulfuric acid is somewhat hazardous and
produce clear raw-water ice rapidly and to the control tests necessary are very important
prevent a brittle product, the water must be for proper operation.
low in alkalinity. In the carbonated beverage
This type of softening for treatment of
industry, low alkalinity is necessary because
boiler makeup water is not applicable on
many soft drinks are prepared with acid ex-
water of high hardness and low alkalinity or
tracts that require certain pH values to ob-
on a water high in silica. In cases where the
tain the desired flavor.
raw water silica concentration is high, the
concentration of silica becomes the con-
LIMITATIONS trolling factor with respect to boiler concen-
Before definite recommendation for this tration control. In such cases, hot process sof-
type equipment can be given it is necessary tening methods in which silica is easily
to make a thorough plant study inasmuch as removed are more suited for use in the
operating cost is composed of several factors. preparation of boiler feedwater.
If the water is muddy, sedimentation, co-
agulation and/or filtration will be necessary REFERENCES
because a clear influent must be delivered to
R. F. Goudey, "Removal of Salts from Water",
the hydrogen zeolite unit otherwise the sus- Journal, Am. Water Works Assoc., Vol. 32, pp.
pended matter will coat the grain surfaces 435-455 ( 1940)
and materially reduce efficiency. Further- R. Kunin, "Ion Exchange Resins", 2nd Ed., John
more, the influent should be free from ferric Wiley & Sons, New York, N.Y. (1958)
E. Nordell, "Water Treatment for Industrial and
iron because iron will also be precipitated on Other Uses", 2nd Ed., pp. 445-458, Reinhold Pub-
the grain surfaces. lishing Corp., New York, N.Y. (1961)


Ion Exchange Resins. . . . . . . . . . . . . . . . . . . . . . 103

Demineralization Equipment. . . . . . . . . . . . . . . . 105
Applications and Limitations . ......... 0 0 0 0 0 • 108

EMINERALIZATION is the process of re- did not gain wide acceptance because of the
D moving the mineral salts from water
by ion exchange. Deionization is generally
problems involved in feeding the fluoride and
in disposing of the waste water from the re-
considered synonymous with demineralization. generation operation. The waste water has
Only those substances which ionize in water a high fluoride content and is toxic in high
can be removed by this process. With most concentration. With the development of the
industrial water sources, it is possible to use highly basic anion exchange resins, the use
demineralization to remove mineral salts to of fluoride is no longer necessary. The highly
the same extent or to a greater extent than basic anion resins are capable of removing
would be possible with distillation. Deminer- silica from demineralized water and the proc-
alization involves two ion exchange reactions. ess now competes directly with distillation as
The cations such as calcium, magnesium and a source of makeup water for high pressure
sodium are removed in a hydrogen cation ex- boilers. In many cases, demineralization is
changer. The process and equipment are the more economical than distillation and can be
same as those used for hydrogen zeolite sof- operated in such a manner that the makeup
tening. The anions, such as sulfate and chlo- water will be of superior quality to distilled
ride, are removed in an anion exchanger. water.
Various arrangements of equipment are em-
ployed dependent upon the requirements of
the installation. Demineralization is carried out with the use
Demineralization and distillation are the of two types of materials. The cation ex-
two most commonly used methods for pro- change material replaces the metallic cations
ducing large quantities of water with a low with hydrogen ions. This process is identical
solids content. In comparison with the cost to hydrogen zeolite softening. Sulfonated coal
of producing distilled water, demineraliza- is sometimes used as the cation exchange
tion by ion exchange will frequently be more material for economy in the use of acid, but
economical for many industries. For this rea- it is more common to use one of the organic
son, the process of demineralization has found cation exchange resins in which the active
wide acceptance among industries requiring group is a sulfonic. acid group. The remainder
process water of low solids content. of the resin is of such a nature that it is in-
Prior to the development of the highly soluble, chemically and physically stable, and
basic anion exchange resins, demineralization permeable to diffusion of ionsfrom the water.
of a water by ion exchange did not lower the Ion exchange is an equilibrium reaction
silica content. Without silica removal, de- which is reversible. In the hydrogen cation
mineralization possessed limited application exchange resins, the active group shows a
for the preparation of boiler feedwater. greater affinity for other cations in preference
Water of quality comparable to that of dis- to hydrogen ion. This process is reversed
tilled water is required for high pressure when the resin is regenerated but the reverse
boiler operation. The presence of silica in process takes place only when the hydrogen
such water is most undesirable, since the ion is present in high concentration. The
silica may permit the formation of boiler tube position of equilibrium depends upon the
deposits or vaporize with the steam and cause relative concentrations of cations present in
turbine blade deposits. It is therefore neces- the water.
sary for demineralized water to have a very During the exchange operation, the princi-
low silica content in order to be useful for pal reaction is indicated by the following
boiler feedwater purposes. equations. In these equations, the symbol
With the weakly basic anion exchange res- R · S0 3 is used to represent the portion of the
ins, it is possible to use fluorides to obtain silica cation exchange material which does not go
removal with demineralization. This process into solution.

(1) Strongly basic anion exchanger:

R·S0 3H + NaCI = R·S03Na + (4)
hydrogen + sodium sodium +
exchange resin chloride resin
R4N ·OH + H2Si03 = R4N ·HSi03
basic + silicic silicate
HCI exchange resin acid resin
hydrochloric + H20
acid +water
(2) In the operation of a strongly basic anion
2R·S03H + CaS04 = (R·S03)zCa exchanger with downward flow, it is believed
hydrogen + calcium = calcium
that the anions of the strong acids will be
exchange resin sulfate resin
removed at the top of the bed with silicate
+ H2S04 representing one of the last anions to be re-
+ sulfuric
acid moved. The top of the bed is exhausted first
and the portion of the bed which is effective
!he ani~n. exchange resins are organic
in removing silica is gradually displaced
·· re~1.ns contammg active amine or quaternary
downward. Eventually the bottom layer in
am1?e gr~ups. The anion exchange resins are
the bed will no longer effectively remove
ava1lable m two types known as weakly basic
silica and the effluent will show increasing
and strongly basic anion exchangers. The
concentrations of silica. When this condition
weakly basic type will remove sulfate chlo-
is reached the bed will still remove other an-
ride and nitrate, but will not remove silica or
ions, with the result that silica break-through
carbon dioxide. The weakly basic resins can
will take place before there is a significant rise
~e. use.d to advantage where the presence of
in conductivity of the effluent.
s1~1ca m the demineralized water is not sig-
m?cant. The .s~rongly basic anion exchangers The purity of demineralized water and
wlll remove s1hca, sulfide and carbon dioxide distilled water is frequently expressed in terms
as well as the other common anions. of specific resistance or specific conductance.
One micromho of specific conductance will
The weakly basic anion exchangers can
normally indicate a dissolved solids content
remove silica only with the feed of sufficient
of 0.5-0.6 ppm. Completely pure water would
fluoride to convert the silicate to a fluosili-
cate. Silicate acts as the anion of a weak acid have a ~pecific conductance of approximately
0.04 m1cromhos due to a small quantity of
~nd is not removed by the weakly basic an-
lOll exchanger. The fluosilicate acts as an
hydrogen ion and hydroxide ion. These ions
anion of a strong acid and can be removed result from the dissociation of a small portion
by a weakly basic anion exchanger. This of the water molecules. Water of such purity
pr?cess is not applied to the preparation of Fig. 16-1
bmler makeup water at the present time. Quality of Low Solids Water from
The strongly basic anion exchangers will Various Sources
(Laboratory Scale Tests)*
remove the anions of both strong acids and
weak acids. For this reason, direct silica re- Specific
moval is possible. The following equations micromhos
can be used to illustrate the manner in which
the anion exchange resins operate. Calculated Maximum Quality ............... 0.038
Single Distilled Water (in glass) . . . . . . . . . . . . . 2.0
Weakly basic anion exchanger: Triple Distilled Water (in glass) . . . . . . . . . . . . . 1.0
Monobed Demineralization ................. 0.055
(3) 28 Distillations in Quartz .................. 0.044
RNH 3·0H + HCI = RNH 3·CI * '' Amberlite Mono bed Deionizotion,'' Rohm and Haas Company
Philadelphia, Po., 1950. '
basic + hydrochloric = chloride
exchange resin acid resin
+ H20 can not be produced by methods which are
+water commercially available at the present time.

The purity of various types of demineralized ployed only when the bicarbonate alkalinity
and distilled water are outlined in Figure of the raw water is low. Otherwise, the cost
16-1. of removing carbon dioxide in the anion ex-
changer becomes a sizeable factor.
DEMINERALIZATION EQUIPMENT An alternate arrangement is to place an
aerator between the hydrogen cation ex-
The equipment employed for demineraliza- changer and the highly basic anion ex-

tion is similar to that employed for hydrogen changer, as illustrated in Figure 16-3. This
zeolite softening. The piping and exchanger arrangement is suitable for water with high
shells are generally lined with synthetic rub- bicarbonate alkalinity, since the carbon di-
ber or plastic materials. Small fittings which oxide is liberated to the atmosphere in the
can not be conveniently protected by coatings aerator. In place of an aerator, it is possible
are constructed of Monel, plastic, or stainless to install a vacuum deaerator which will re-
steel. move nearly all dissolved gases such as oxy-
The aeration towers which are used for gen, in addition to the carbon dioxide.
removal of carbon dioxide are frequently of Another arrangement is to pass the raw
wooden construction with any exposed metal water into the hydrogen cation exchanger,
parts of Monel, plastic, or stainless steel. Ce- then into a weakly basic anion exchanger fol-
ramic materials are also being employed. lowed by an aerator and then a highly basic
Final effluent water lines and storage tanks anion exchanger. This system is illustrated in
should be protected with a suitable paint or Figure 16-4. It is necessary to regenerate
lining to prevent corrosion. Such corrosion the highly basic anion exchanger with caustic
will arise from the dissolved oxygen content soda, but the weakly basic anion exchanger
of the demineralized water and would con- may be regenerated with soda ash or with
taminate the water with iron oxide. the waste regenerant from the highly basic
Demineralization with silica removal may anion unit. As a consequence, the operating
be carried out in several steps. The simplest cost for chemical regeneration is lower with
arrangement is illustrated in Figure 16-2 this type of arrangement when the raw
in which the raw water passes through a water has relatively high sulfate or chloride
cation exchanger directly to a highly basic content.
anion exchanger. This arrangement is em- Maximum purity of demineralized water


l 1



~ 1 T

Figure 16-2 • Two-Bed System for Demineralization and Silica Removal



1 l




~ T T

Figure 16-3 • Two-Bed System with Aerator for Demineralization and Silica Remoual

can be obtained by placing a number of ex- types of resins are intermixed. The resins are
changer units in series. A similar result can of different density and during backwashing
be obtained by mixed bed demineralization. will separate into two layers. Each layer is
This arrangement uses a mixture of cation regenerated and rinsed. After rinsing the
and highly basic anion resins in the same resins are remixed by air agitation and the
shell. Figure 16-5 illustrates the internal ar- exchange operation is resumed.
rangement of the resin beds during regenera- The theoretical purity of the effluent water
tion. from a mixed bed demineralizer is very high.
During the exchange operation the two The equipment requires a complicated ar-


j_ _L


T 1

Figure 16-4 • Three-Bed System with Aerator for Demineralization and Silica Remoual

RAW WATER INFLUENT to growth of bacteriological slime. Depending

RINSE INFLUENT upon characteristics of the water source, the
BACKWASH EFFLUENT influent water to the demineralizing system
CAUSTIC SODA REGENERANT may require a preliminary treatment of co-
AIR agulation, filtration and possibly chlorination.
I EFFLU~ NT However, excess chlorination will cause de-
_...... -......... terioration of the cation resin so that some-
times dechlorination is required. The usual
methods of dechlorination involve the feed-
ing of sodium sulfite or passing the water
through a bed of activated carbon ahead of
1- the cation exchanger. In some cases, cold
z process softening is applicable prior to the
ct:: demineralizer, particularly if there are other
1-z plant water requirements for which demin-
ANION EXCHANGE zW eralized water is not necessary.
3ct:: The operation of equipment is similar to

... 1..1..0
the operation of zeolite softening equipment.
Operating water flow will usually be in the
range of 5-8 gallons per minute per square
foot of bed surface. A minimum bed depth
....J of 30 inches is usually recommended and
RESIN :::> the bed depth will frequently be much
greater in order to obtain the desired ex-
change capacity for the unit. Operating flow
__.... rates are· sometimes expressed in terms of
f J
gallons per minute per cubic foot of bed. In a
moist condition, the exchange resins will
weigh approximately 30-50 pounds per cubic
SPENT REGENERANT foot. The cation exchange resins are heavier
BACKWASH INFLUENT than the anion exchange resins.
Figure 16-5 • Arrangement for Mixed Bed Backwash rates of 4-8 gallons per minute
Demineralization per square foot are frequently employed, de-
rangement of valves in order to permit using pending upon density of the resin and tem-
caustic soda in one regeneration step and acid perature of the backwash water. Higher back-
during the other step. There is also likely to wash rates are required for the heavier resins
be a variation in capacity of the unit, unless and for higher water temperature, due to
remixing of the resins after regeneration is reduced viscosity of the water. Backwash rate
thorough and uniform. is normally adjusted to provide between 50
Regardless of the relative arrangements of and 100% expansion of the bed. In some
the exchangers. it is essential that the influent plants, where the influent water is not com-
water be free of suspended solids. Silt and pletely free of suspended solids, it has been
organic matter are particularly objectionable, found advantageous to install filter bed agita-
since these materials can deposit in the ex- tors to obtain maximum cleaning of the resin
changer beds and reduce the capacity of the during backwash.
units by coating the beads of exchanger resin Regenerant flow rates are usually low and
with films which either prevent or retard the close to approximately one gallon per minute
movement of anions and cations through the per square foot. With strongly basic anion
resins. Deposits of organic matter may lead exchangers, maximum removal of silica in a

given period of time will be obtained if the gen cation exchanger will have an effluent pH
caustic soda solution is heated to 90-100 F, or in the acid range and a certain value for free
slightly higher. mineral acidity. Operation of ,the cation unit
With the cation exchangers, the concen- can be followed to some degree by means of
tration of the regenerant solution may be in a continuous pH recorder. Sodium will break
the range of 1-10% for hydrochloric acid and through the hydrogen cation unit a short
1-5% for sulfuric acid. When using sulfuric time before there will be a significant increase
acid to regenerate a resin cation exchanger in pH. For maximum purity of the effluent
it is sometimes necessary to start at a low water from the system, it is necessary to check
concentration of acid, which is increased in the hydrogen cation exchanger effluent for
steps if the influent water to the unit con- sodium content when the end of the expected
tains an appreciable calcium content in com- run approaches. If maximum purity is not
parison to the other cations present. This essential, determination of pH and free min-
procedure is necessary to avoid precipitation eral acidity are used as the control tests.
of calcium sulfate upon the resin. Hydro- The quality of the effluent from the
chloric acid does not require step-wise re- strongly basic anion exchanger can be checked
generation procedure. It is more difficult to by means of a conductivity recorder. Silica
handle than sulfuric acid and for this reason and carbon dioxide will break through the
its use is avoided. unit somewhat ahead of other anions. For
In the regeneration of strongly basic anion minimum silica in the effluent water, silica
exchangers, the strength of the caustic soda determinations must be conducted as the end
solution is in the range of 2-4%. With weakly of the expected run approaches.
basic anion exchangers, the strength of the
sodium carbonate regenerant will normally
be in the range of 2-4%. Ammonia can also
be used as a regenerant if silica removal is Demineralization with silica removal has
not required. proved to be an economical method for pro-
Following introduction of the regenerant viding makeup water for high pressure boiler
solution, the unit is rinsed to remove excess operation. In many cases, it is a more eco-
regenerant. Rinsing will generally require 50 nomical method of makeup water prepara-
to 150 gallons of water per cubic foot of tion than distillation using single effect evap-
resin. About 5 to 10% of the influent water orators. Demineralization either with or
will be wasted from the system as a result of without silica removal has proved to be a
backwash and rinse operations. It should be useful and economical process for many in-
noted that the hydrogen cation exchanger dustries. The use of demineralized water has
can be rinsed with raw water, but the anion been applied to the manufacture of chemical
exchanger must be rinsed with effluent from and pharmaceutical products. Demineralized
the hydrogen cation exchanger. Rinse water water is used in some textile processes, in the
from the anion unit can be recycled to the silvering of glass, and in some electro-plating
raw water supply when the conductivity of processes. It has also been used widely in
the rinse water decreases to the conductivity various manufacturing processes where qual-
of the raw water. supply. A small portion of ity of the finished product may be affected
the initial rinse water from the hydrogen by the presence of mineral salts in water
cation exchanger can be reclaimed and used coming in contact with the product.
at the beginning of the next regeneration At the present time, hydrogen cation ex-
cycle, since this water will have an appreci- changers are available which can tolerate
able acid content. operating temperature as high as 250 F.
Control of demineralization plants requires Strongly basic anion exchange materials must
determination of several factors. The hydro- generally be operated at a water temperature

not exceeding 105 F. As a consequence, it is more precise methods of operation control.

necessary to carry out demineralizing proc- These changes will permit production of de-
esses at relatively low temperatures at the mineralized water with greater purity and at
present time. When the raw water source is lower cost.
very high in dissolved solids, evaporation will
frequently prove to be the most economical
method of obtaining water which will be REFERENCES
nearly free of mineral salts. If the raw water
S. B. Applebaum, "Experiences With Silica Re-
is contaminated with substances of an or- moval Demineralizer Plants", Proceedings, Mid-
ganic nature, distilfation may be required for west Power Conf., Vol. XIII, pp. 275-289 (1951)
purification. Coagulation, chlorination and H. W. Frazer, "Special Operating and Regenerat-
filtration will frequently reduce organic con- ing Techniques for Ion Exchange", Proceedings,
Engineers Soc. of Western Penna., pp. 83-96
tent of water to very low values. However, (1958)
not all organic materials can be completely N. W. Frisch and R. Kunin, "Long-Term Operat-
removed by such methods. Many organic ing Characteristics of Anion Exchange Resins",
Ind. and Eng. Chem., Vol. 49, pp. 1365-1372
substances can not be removed by deminerali- (1957)
zation, since they do not ionize in water. M. E. Gilwood, C. Calmon and A. H. Greer, "Sil-
Several years ago the fluosilicic acid process ica Removal Characteristics of Highly Basic Anion
Exchangers", Proceedings, Engineers Soc. of West-
for silica removal was the only one which ern Penna., pp. 119-130 (1951)
was available for practical application to J. Harlow, V. J. Calise and M. Lane, "Boiler
the preparation of boiler makeup water Feedwater Treatment For A High-Pressure, High-
Makeup, Power And Steam Producing Plant",
with demineralization. This process is now Proceedings, Midwest Power Conf., Vol. XIII, pp.
obsolete, having been superseded by a more 290-303 ( 1951)
efficient and economical process. Since the R. Kunin, "Ion Exchange Resins", 2nd Ed., John
Wiley and Sons, Inc., New York, N. Y. ( 1958)
present methods of demineralization are com-
F. C. Nachod and J. Schubert, "Ion Exchange
paratively new and being widely applied, it Technology", Academic Press, New York, N. Y.
can be expected that considerable progress ( 1956)
will be made within the next few years. Such R. I. Smith and H. D. Reppin, "Four Years Of
Operating Experience With A 3200-GPM Demin-
progress will most likely result in more effi- eralizer", Proceedings, Am. Power Conf., Vol.
cient use of regenerant chemicals and in XXIII, pp. 635-649 ( 1961)

Dealkalization by
Chloride-Anion Exchange

Use for Dealkali::;ation. . . . . . . . . . . . . . . . . . . . 111

Chemistry Involved. . . . . . . . . . . . . . . . . . . . . . . 111
Equipment Required. . . . . . . . . . . . . . . . . . . . . . 113
Abplications. . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
Limitations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113


D change is a process whereby a sodium
zeolite softened water is passed through a The basic usefulness of this new process lies
second exchange unit which removes anions in its ability to reduce the alkalinity of the
such as sulfate, nitrate, carbonate and bi- treated water without the use of acid.
carbonate. These anions are replaced by The alkalini~y of a water is unchanged in
chloride. Salt is used to regenerate the chlo- passage through a sodium zeolite softener.
ride-anion exchanger. Since carbonate and Concentration of feedwater alkalinity in the
bicarbonate ions are exchanged for chloride boiler may develop high boiler water alka-
in this process, alkalinity reduction is achieved linities, with possible danger of foaming and
without the use of acid. carryover of solids with the steam. In addi-
It is theoretically possible to employ un- tion, decomposition of alkalinity under boiler
softened water as influent to the chloride- water temperatures will produce carbon di-
anion exchanger. However, at this stage in the oxide in the steam. Carbon dioxide, thus
development of this process, it is the con- formed, is a major source of steam and co.n-
sensus among equipment manufacturers that densate line corrosion.
it is desirable to precede the dealkalizer with a These disadvantages to sodium zeolite sof-
sodium zeolite softener. With unsoftened tening can be minimized by acid treatment
water there is the danger of precipitating of the zeolite effluent or by employing hy-
calcium carbonate and magnesium hydroxide drogen zeolite softening. Both of these treat-
in the dealkalizer bed. In addition, the chlo- ment methods involve the use of acid, which
ride-anion exchange bed is even more critical is regarded with disfavor in small plants,
with respect to suspended matter than the without adequate personnel.
conventional sodium zeolite cation exchanger. While not as efficient or economical as
Because the anion exchange resin is lighter acid treatment or hydrogen zeolite softening,
than the cation exchanger, it can be back- dealkalization by chloride-anion exchange
washed only at low rates, insufficient to re- possesses the advantage of securing similar
move suspended matter. The use of a sodium alkalinity reduction without the necessity of
zeolite, preceding the anion exchange, serves handling acid.
as an additional safeguard for the removal
of suspended matter. Even for the sodium CHEMISTRY INVOLVED
zeolite unit, turbidity of the influent water
should not exceed 10 ppm. Dealkalization by chloride-anion exchange



1-::-'::. I:1
~ -~1

Figure 17-1 • Sodium Zeolite-Chloride Anion Exchange


makes use of highly basic anion exchange of chloride. Unlike demineralization, solids
resins employed in demineralization. In de- are not removed by· dealkalization, but con-
mineralization, a raw water is first passed verted to the sodium chloride form. Reac-
through a cation exchanger, where cations tions in the chloride-anion exchangers are as
are exchanged for hydrogen ions. Calcium, follows:
magnesium and sodium are removed and the (HC0 3) - + R·CI R·HC0 3
cation exchanger effluent contaiils sulfuric bicarbonate + chloride = bicarbonate
acid, hydrochloric acid, carbonic acid, etc., ion anion anion
which are then adsorbed in passing through resin resin
the second unit, the anion exchanger. The + Cl-
cation unit is regenerated with acid and the +
anion unit with alkali. This demineralization ion
process results in the complete removal of (504)= + 2R·CI = R2·S04 + 2CI-
both cations and anions, producing an efflu- sulfate chloride sulfate chloride
ent comparable to distilled water. ion + anion = anion + ion
resin resin
Chloride-anion exchange differs from de-
mineralization in that cations and anions are Of course, there is no advantage secured
not removed from solution, but are ex- in exchanging sulfate for chloride. Nitrate is
changed for an equivalent amount of other also exchanged for chloride in this process.
ions of less undesirable characteristics. In de- Where the sulfate and nitrate content of a
alkalization, two separate units are employed. water is high, the cost of the dealkalization
The first unit is a conventional sodium zeo- process is increased by the exchange of these
lite softener, regenerated with sodium chlo- ions, from which exchange no advantage is
ride, which exchanges calcium or magnesium secured.
ions for sodium ions. The reactions are as However, as indicated in the preceding re-
follows: actions, bicarbonate is removed by passage
Ca++ + Na 2Z = CaZ + 2Na+ of a sodium zeolite softened water through
calcium + sodium = calcium + sodium the chloride-anion exchanger and this re-
ion zeolite zeolite ion moval of alkalinity is the chief objective of
Mg ++ + Na2Z = MgZ the process~namely removal of alkalinity
magnesium + sodium = magnesium without the use of acid.
ion zeolite zeolite The chemical reactions of regenerations
+ 2Na+ with this process are as follows:
ion 1st Unit~Sodium Zeolite
Of course, as previously mentioned, anions CaZ + 2NaCI = Na2Z +
CaCI 2
are not affected by passage through a sodium calcium + sodium = sodium + calcium
zeolite unit. Alkalinity, sulfate, chloride, etc., zeolite chloride zeolite chloride
pass through unchanged. 2nd Unit~Chloride-Anion Exchanger
In the second unit, the exchange material R·HC0 3 +
is a highly basic resin anion exchange ma- bicarbonate + sodium = chloride
terial, also regenerated with sodium chloride. anion chloride anion
In passage through this unit the sodium salts resin resin
present in the influent, such as sodium bicar- +NaHC03
bonate and sodium sulfate, are converted to + sodium
anequivalent amount of sodium chloride. In bicarbonate
other words, the dealkalization unit removes The exchange capacity of the dealkaliza-
the anions present in the influent by exchang- tion unit is affected by several variables. Data
ing these anions with an equivalent amount have been presented to the effect that ca-

pacity for the total exchangeable anions (bi- should not exceed five gallons per square foot
carbonate, carbonate, sulfate and nitrate) per minute.
will decrease with an increasing chloride con- Salt regeneration rates range between 3.0-
tent in the raw water. The total anion 4.0 pounds of salt per cubic foot of resin. It
exchange capacity (including the raw water is desirable to employ softened water in pre-
chloride) is not affected by variation of the paring regenerant brine. Rinse and regener-
percentage of the total anions which are ant rates of 0.5 gallons per minute per cubic
present in the chloride form. However, since foot of resin are recommended.
the purpose of the process is the removal of
bicarbonate alkalinity, the capacity of the APPLICATIONS
resin bed for bicarbonate removal will de-
crease with increasing chloride concentrations The major usefulness of dealkalization by
of the raw water. ion exchange is in relatively small plants
The process is most efficient when the where the hazard of acid handling is a con-
anions are low with the exception of bicarbon- trolling factor in deciding the type of ex-
ate. There is no advantage in exchanging sul- ternal treatment process to be installed.
fate and nitrate with chloride, and yet regen- Where neutralizing amines are employed, the
erant must be employed for this purpose. use of a chloride-anion exchanger may per-
Consequently, for low cost operation, it is mit economy, since any reduction in the C0 2
desirable that the sulfate, chloride and nitrate content of the steam will be reflected in a
content of the raw water be low in compar- dec.reased requirement for the neutralizing
ison with the bicarbonate. amme.
Exchange capacities of 5-10 kilograins ex- Only one regenerant chemical (salt) is re-
changeable anions (expressed as CaC0 3 ) per quired, although as mentioned previously,
cubic foot of resin have been reported. Im- improvement in the exchange capacity may
provement in exchange capacity has been be achieved through the use of caustic soda
noted when sodium hydroxide is used with along with the salt. There is no hazard in-
the sodium chloride in the regenerating solu- volved in the handling of salt and it is a
tion, in the ratio of approximately one pound cheap, readily available material. Where suf-
of caustic soda for each ten pounds of so- ficient competent personnel are not available
dium chloride. for controlling the other methods of alkalinity
At one of the first installations of the de- reduction, the chloride-anion exchange proc-
alkalization process, it was reported that the ess is definitely advantageous.
addition of sodium hydroxide to the sodium
chloride regenerant resulted in an increase LIMITATIONS
in the exchange capacity of approximately
With the installation of proper equipment
45-60%. At this plant, raw water alkalinity
and adequate personnel, there is little hazard
as calcium carbonate is 145-150 ppm and the
involved in handling acid. Direct acid treat-
dealkalization unit is taken off line and re-
ment of a sodium zeolite effluent or the use
generated when the effluent alkalinity IS m
of hydrogen-sodium zeolite softening will ac-
the range of 27-31 ppm as CaC0 3 •
complish alkalinity removal more efficiently
and economically than the chloride-anion ex-
change process. In addition, in plants of suf-
The chloride-anion exchanger is similar in ficient size to justify the installation, the hot
construction to a conventional sodium zeolite lime-hot ion exchange process will prove
unit and equipped with the same accessories more economical than chloride-anion ex-
for regeneration. A resin bed depth of 30-36 change.
inches is recommended. Operating flow rate Like other treatment methods, dealkaliza-

tion by chloride-anion exchange, reduces but quently, in a plant with high carbon dioxide
does not eliminate the co1 content of the content of the steam and satisfactory corro-
steam. Dependent on the type of equipment sion control with the filming amines, the in-
in which steam is condensed and the degree stallation of a chloride-anion exchanger to
of stratification of carbon dioxide in condens- lower the carbon dioxide in the steam would
ing equipment, severe corrosion problems can not necessarily permit a reduction in the
result even from relatively low C0 2 concen- amine rate of feed.
trations in the steam. Where filming amines
are used for control of carbon dioxide cor- REFERENCES
rosion, the amine feed rate is substantially in-
dependent of C0 1 concentration. The quan- S. B. Applebaum, "Experiences with Chloride An-
ion Exchangers for Reducing Alkalinity Without
tity of amine required is only sufficient to Acid", Proceedings, Engineers Soc. of Western
provide a non-wettable film on the metal Penna., pp. 1-8 (1952)
surfaces. Except under unusual circumstances, A. E. Kittredge, "Dealkalization of Boiler Feed
it is ,not necessary to increase the feed rate Water in Small Power Plants", Proceedings, Am.
Power Conf., Vol. XV, pp. 567-572 (1953)
of the filming amines because of higher car- E. Maher, "Dealkalization System", Industry and
bon dioxide content of the steam. Conse- Power, Vol. 63, pp. 102-103 (Sept. 1952)

Hot Lin1e-Hot Ion
Exchange Softening

Use of Process 0 0 0 0 0 0 0 0 0 o 0 o o 0 0 o 0 0 0 0 0 o o o o o 0 116

Chemical Reactions 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 o 0 o o 0 o 116
Equipment Employed 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 o 0 0 o o o 0 117
Applications 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 o 117
Limitations 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 o , 0 o 0 o 0 o 118

ITH the development of styrene base quires close attention and frequent control
W resins that can withstand waters of
high pH and high temperature, a two-stage
tests to assure proper operation at all times.
Hot lime softening alone will reduce the
softening process has been developed, em- carbonate and magnesium hardness as well
ploying hot lime softening as the first stage as the total solids of a raw supply. Only lime
followed by hot ion exchange softening as is required when the calcium exceeds the
the second stage. In many cases, this method total alkalinity. On high alkalinity waters, it
of treatment offers definite advantages over may be found necessary to apply calcium
the usual two-stage hot process softeners using chloride or calcium sulfate to reduce the al-
lime and soda ash in the first stage and phos- kalinity to the desired range. The effluent
phate in the second stage. from the lime treater, whether filtered or un-
Although the ion exchange unit will oper- filtered, is passed through a sodium ex-
ate satisfactorily on the effluent from hot changer containing the styrene base resin
process lime-soda softening, it is usual to use which removes the residual calcium and
only lime treatment in the first stage soften- magnesium salts, thereby producing a water
ing, so as to minimize softener effluent alka- with a total hardness in the range of 0.5 to
linity and thereby produce less carbon dioxide 2 ppm.
in the steam.
Hydrated lime is employed to react with the
For waters containing high concentrations of bicarbonate alkalinity of the raw water, pre-
hardness, alkalinity, total solids and silica, it cipitating calcium carbonate as shown by the
was often found necessary to resort to hot following reaction.
process lime-soda softening to reduce these
( 1)
factors to a reasonable range for boiler feed-
Ca(OH)z + Ca(HC03lz = 2CaC03
water purposes, even for relatively low boiler hydrated + calcium = calcium
pressures. At higher boiler pressures, hot lime bicarbonate carbonate
lime-soda softening alone is not completely
satisfactory due to the relatively high softener
+ 2H20
effluent hardness, resulting in boiler sludge To reduce silica, the natural magnesium
accumulations even with proper internal of the raw supply can be precipitated as mag-
treatment. To overcome this difficulty, hot nesium hydroxide which adsorbs silica from
phosphate softening following hot lime-soda solution. Precipitation of magnesium requires
softening has often been employed to reduce a slight excess of lime and this reaction is
the makeup water hardness to approximately shown by the following equation.
2-3 ppm.
Due to variations in composition common Ca(OHlz + MgCh = Mg(OHlz
to many water supplies, it has often been hydrated + magnesium = magnesium
found difficult to control chemical balances lime chloride hydroxide
in the effluent from a hot lime-soda softener. + CaCh
When such variations occur, boiler water + calcium
balances are upset, resulting in poor internal chloride
conditions. With two-stage hot lime soda-hot If the quantity of natural magnesium in
phosphate softening, any variations in the the raw supply is not sufficient to reduce the
primary softener are immediately reflected in silica to the desired level, it will then be
a poor effluent from the secondary softener. necessary to employ an activated magnesium
From the operating standpoint, either single oxide or dolomitic lime depending on the
stage hot lime-soda softening or two-stage type water to be treated. No soda ash is
hot lime soda-hot phosphate softening re- employed to reduce the noncarbonate hard-

ness since this is automatically removed by posits. With the use of hot phosphate soften-
the ion exchange softener. The reactions oc- ing, difficulties with adherent boiler sludge,
curring in the hot exchange softener follow- principally in the form of magnesium phos-
ing the hot lime softener are identical to those phate are encountered. The use of the hot
outlined in the chapter on sodium zeolite ion exchange process reduces this factor to
softening. a minimum.
The normal hot lime-soda process requires
EQUIPMENT EMPLOYED maintaining an excess of carbonate alkalinity
in the softener effluent to reduce calcium
The equipment employed in this process is hardness to a minimum. The high carbonate
similar to that described in the chapters on alkalinity of the softener effluent increases the
hot lime-soda softening and sodium zeolite carbon dioxide content of the steam, render-
softening with only minor modifications. Fig- ing the resulting condensate more corrosive.
ure 18-1 illustrates a typical hot lime-hot ion In addition, a high blowdown rate often is
exchange softening process. necessary to restrict boiler water alkalinity
within safe limits. With hot lime softening,
APPLICATIONS the carbonate alkalinity of the softener efflu-
ent can be maintained at a minimum and, as
This system of treatment is desirable for high a result, the carbon dioxide content of the
pressure boilers ( 450 psi-1000 psi) since it steam will be lower than results with lime-
produces a water of low total hardness, low soda softening. Blowdown for alkalinity con-
silica and low alkalinity. trol can be maintained at a minimum.
Low total hardness is essential for high Following hot lime-soda softening or hot
pressure operation because of the suspended lime-soda, hot phosphate softening, calcium
solids limitation in preventing adherent de- carbonate or calcium phosphate deposits will




Figure 18-1 • Flow Sheet of Typical Hot Lime-Hot Ion Exchange Process

Figure 18-2-Typical Softener Effluent Analyses

Hot Lime-Soda, Hot Phosphate Hot Lime, Hot lon Exchange
lsi Stage 2nd Stage lsi Stage , 2nd Stage
Raw Water (Lime-Soda) (Phosphate) (Lime) (Exchanga Resin)

Total Hardness ........ as CaC03, ppm ... . 150 18 0-2 70 0-2

Calcium ............. as CaC03, ppm ... . 120 15 0-1 67 0-1
Magnesium ........... as CaC03, ppm ... . 30 3 0-1 3 0-1
"P" Alkalinity ........ as CaC0 3 , ppm ... . 0 40 30 10 10
"M" Alkalinity ........ as CaC03, ppm ... . 100 60 50 20 20
Silica ................ as Si02, ppm ..... . 15

Hot Lime-Soda, Hot Lime-

Treatment Requirements Hot Phosphate ion Exchange

lime, lbs/1000 gal .................................................. . 1.0 0.9

Soda Ash, lbs/1000 gal, ........... , ........................•......... 0.8
Phosphate (D.S.P.), lbs/1000 gal ....................................... . 0.18
Salt, lbs/1000 gal. ......................................... , ........ . 2.0

*Can be reduced to approximately 1 ppm with use of magnesium oxide or dolomitic lime.

occur in stage heaters or economizers unless be installed to supplement an existing hot

the pH is reduced by the addition of an process softener, the question of turbidity of
acidic material or dilution with appreciable the influent to the ion exchange softener is
returned condensate. Following hot lime-hot of primary importance. Excessive turbidities
ion exchange softening, there is no danger of in the influent to the exchange softener will
such deposits because the calcium and mag- tend to reduce capacity and form deposits
nesium salts are practically "zero." In fact, within the unit. When such plugging occurs,
with the use of certain types of surface active it becomes necessary to acid wash the ion
agents, the required amount of internal boiler exchanger to remove the deposits.
water treatment may be fed directly to the Even though a properly designed lime
feedline without danger of deposition. The treater will normally produce a low turbidity
internal chemical feed system is thus simpli- water, there are occasions when high tur-
fied since individual high pressure chemical bidities can occur, such as: quick load
feed pumps for each boiler are unnecessary. changes, overload conditions, mechanical
Silica deposits or silica carryover with the failure of softener controls, improper chemi-
steam are important factors to be considered cal balances, and rapid changes in the raw
with high pressure boiler operation. The silica supply. To prevent excessive turbidities from
concentration in the boiler feedwater can be interfering with the hot exchange softening
accurately controlled in the first stage sof- process, filters are normally recommended.
tener and there is no pick-up of silica as the If the exchange units are used as filters,
water passes through the hot ion exchange the flow rate in gallons per square foot per
softener. Figure 18-2 illustrates typical efflu- minute must be low to obtain proper filtra-
ent analyses for a hot lime-soda, hot phos- tion. Lack of filters requires a larger volume
phate softening process and also a hot lime- of the costly resin than for the softening
hot ion exchange softening process. alone since proper softening can be obtained
at flow rates considerably higher than the
LIMITATIONS acceptable filtering rate. It will also be neces-
sary to backwash the hot exchange units at
Whether an entirely new system of hot lime- more frequent intervals than they require re-
hot ion exchange softening is being installed, generation in order to avoid break-through of
or whether a hot ion exchange system is to hardness, due to channeled beds. In the ab-

sence of filters, it can be generally stated that suitable for process use and consequently the
more operating attention is required. While installation of a hot lime softener is generally
some installations operating without filters not economical. In such cases, cold ion ex-
have reported little difficulty, it is the opinion change softening is generally used to com-
of the manufacturers that it is safer to have plete the treatment for boiler makeup water.
filters installed. From the chemical cost viewpoint, it is
When hot exchange softening is being necessary to c~lculate the exact lime and salt
considered to supplement an existing hot requirements and the lime-soda requirements
process softener, there are many items which in each case in order to determine if any
require detailed study. One consideration advantage exists. Also, the cost of internal
should be the turbidity in the effluent from treatment following the various types of ex-
the first stage softener and its effect on the ternal softening systems should be correlated
exchange resin in the event that the installed in order to determine any economic advan-
filters were converted to hot exchange sof- tage. ·
tening units. Like any other system of water treatment,
The hot exchange resins must be back- a hot lime-hot ion exchange system has its
washed during the softening phase to loosen limitations which should be thoroughly
the bed, due to the long softening runs re- studied and evaluated in order to assure that
sulting from the low influent hardness and the system is practical and economical to
high exchange capacity. Higher backwash meet the requirements of the individual
rates are normally required to properly ex- boiler system.
pand the bed, because of the lower density
of the hot backwash water. The quantity of REFERENCES
backwash water and rinse water per unit of
time should be accurately calculated since S. B. Applebaum, "Hot Lime Treatment Followed
by Sodium Zeolite", Proceedings, Engineers Soc. of
this water must come from the hot lime sof- Western Penna., pp. 127-145 (1950)
tener tank and therefore may impose a seri- W. S. Butler, "An Inspection Report After Eighteen
ous overload condition. Months with Hot Lime Zeolite", Proceedings, En-
gineers Society of Western Penna., pp. 1-13 (1951)
In plants operating at lower boiler pres- V. J. Calise, "Hot Zeolite Softening", Power En-
sures with a reasonable quantity of return gineering, Vol. 54, pp. 54-59 (Sept. 1950)
condensate, the installation of a hot lime-hot J. E. Harden and G. R. Hull, "Operating Experi-
ion exchange system represents a consider- ences with a Large Hot Lime-Zeolite System for
1500 PSI Boilers", Proceedings, Am. Power Con£.,
able capital investment that may not be Vol. XIX, pp. 672-684 (1957)
economically justified. B. E. Varon and S. B. Applebaum, "An Application
of Hot Lime Zeolite to Moderate High-Pressure
In many industries, the entire water sup- Boiler Operations", Proceedings, Am. Power Con£.,
ply must be lime treated to produce a water Vol. XIX, pp. 660-671 (1957)

Oil Ren1oval

Difficulties Due to Oil. . . . . . . . . . . . . . . . . . . . . 121

Oil Removal Methods. . . . . . . . . . . . . . . . . . . . . 122
Oil Removal from Exhaust Steam. . . . . . . . . . 122
Oil Removal from Liquid Condensate. . . . . . . 122
Analysis for Oil. . . . . . . . . . . . . . . . . . . . . . . . . 126

IL removal is a step in the preparation Where exhaust steam is used for process
0 of water used for process or boiler feed
by which this contamination is removed by
work involving heat exchange, it is desirable
that the steam be free of oil in order that
either mechanical or chemical means. oily films may not develop on the heat ex-
Oil which is employed for steam cylinder change surfaces of radiators, calendering
lubrication is usually mineral oil containing rolls, heater coils, etc. The presence of oil
a small percentage of a compounding ma- decreases efficiency by impeding the flow of
terial or fatty oil. The use of a compounded heat, requiring higher back pressures on the
oil provides better lubrication by emulsifying engine exhaust in order to maintain sufficient
with the moisture in the steam and prevent- heat flow.
ing the oil film from being washed off the Compounded oils may cause foaming of
lubricated surfaces. While lubrication is aided the boiler water. Animal fatty oils employed
by the compounding oil, the emulsion so for compounding purposes are saponifiable in
formed makes the removal of oil from the the presence of normal boiler water alkalin-
condensed steam more difficult. ities, thus forming soap and developing a
foaming condition of the boiler water. The
impure steam resulting from such a boiler
DIFFICULTIES DUE TO OIL water may foul superheater tubes and turbine
blading, destroy cylinder lubrication, clog
The presence of oil in boiler feed water usu-
lines and traps, etc.
ally originates from three sources:
Due to the difficulties that may result from
Exhaust or condensed steam from recipro-
the presence of oil in boiler water, the pre-
cating engines, pumps, etc.
vention of oil contamination of the boiler
Steam or water used in certain process
feedwater is recognized as highly important.
heating equipment.
The American Boiler Manufacturers Associa-
Industrial wastes.
tion in its standard guarantee on steam pur-
The removal of oil from boiler feedwater ity formerly specified that the total quantity
is desirable and necessary because of the un- of oil or grease, or substances which are ex-
predictable and objectionable effects it may tractable either by sulfuric ether or by chlo-
have on boiler operation. roform, shall not exceed 7 ppm in the boiler
The presence of even thin films of oil on water when the sample being tested is acidi-
the boiler heating surfaces will seriously re- fied to 1% hydrochloric acid, or 7 ppm in the
tard heat transfer. Oil acts as a heat insula- feed water when the sample being tested is
tor and preveB.ts the rapid transmission of first concentrated at low temperature and
heat from the metal to the water. The re- pressure to the same ppm total solids as the
sultant increase in metal temperature may boiler water.
be sufficient to cause blistering and loss of Because of the importance of avoiding the
boiler tubes and plates, particularly at high introduction of oil into the boiler feedwater
boiler ratings. system, the decision must be made as to
The combination of oil with boiler water whether oily condensate or exhaust steam
sludge may result in the formation of a should be discharged to waste or oil removal
pasty mass. Pellets of greasy sludge may form, equipment installed. Since considerable heat
which may be oval or spherical in appear- will be contained in the oil contaminated
ance, or the entire mass may be deposited on steam or condensate, and since this steam or
the boiler surface. Circulation and heat trans- condensate will usually be of a high degree
fer may be seriously impaired, resulting in of purity with respect to mineral solids and
starvation of boiler tubes or overheating due otherwise desirable for boiler feed purposes,
to the heat insulating effect of the oily the installation of oil removal equipment is
sludge. usually justified.

OIL REMOVAL METHODS rator reservoir. Frequent inspection of trap

operation should be made. If the trap is not
Oil removal equipment may be divided into functioning properly the reservoir will fill
two types dependent upon the field of ap- with the oil-water mixture and overflow, in
plication. Baffle or centrifugal type separators which case the oil separator ceases to func-
can be used for the removal of oil from tion. Separator efficiency is dependent on the
steam, while various adsorbents, filters and proper removal of the oil-water mixture from
chemical coagulation can be employed for the reservoir.
the removal of oil from liquid condensate. Separators operating on the centrifugal
OrL REMOVAL FRoM ExHAUST STEAM. Baffle principle are also employed. Receiver and
type separators are used to separate oil from purifier type separators operate on the same
steam. Steam containing oil impinges on a principle as baffle type separators but may
baffle and the condensed moisture and oil are have more than one baffle, may be filled with
thrown out on the baffle and drain down the steel wool, and may be drained from more
vertical ribs into the reservoir of the separa- than one compartment.
tor. From this point the oil-water mixture is Mechanical separators of this type are, of
drained or trapped to waste. The direction course, of no value for the removal of oil
of steam flow is altered horizontally and from liquid condensate, but are of definite
passes around the baffle. Figure 19-1 illus- usefulness in the removal of oil from ex-
trates a unit of this type. haust steam. It is usual for such separators to
For the proper operation of such separa- be installed in the exhaust lines from engines,
tors, it is important that the oil-water mix- particularly if the exhaust steam is to be used
ture be continuously removed from the sepa- for direct process work or in heat exchange
operations. Open type feedwater heaters are
customarily provided with baffle type separa-
tors on the exhaust steam inlet.
Baffle type oil separators cannot be em-
ployed with superheated steam, but must be
used with saturated steam since the action
of the separator depends on condensation
taking place on the baffles. If oil is present in
superheated steam, a de-superheater may be
used to provide saturated steam or feedwater
may be sprayed into the superheated steam
prior to the separator.
Free Oil. When oil is present as a mechanical
mixture in water, direct filtration may be
employed. The oil bearing water is passed
through various devices, using for a filter
medium some coarse material as loofa
sponges, burlap, terry cloth, coke or excelsior.
Equipment using the above materials as fil-
tering media operates on the principle that
the semi-porous mass, when thoroughly wet
with water, will permit water to pass through
bm will retain the oil. Hot wells may be
Courtesy Cochrane Division-Crane Company filled with sponges or coke designed to re-
Figure 19-1 • Baffie Type Oil Separator tain on its surface the oil present in the

water. Sponges are usually changed fre- ployed do not differ from those used in con-
quently and reinstalled after boiling with a junction with hot or cold process lime-soda
caustic soda or soda ash solution for removal or hot phosphate softening. The optimum
of the adsorbed oil. Coke beds are usually re- filtration rate for an oil filter is two gallons
placed with fresh coke when inspection re- per square foot of filter area per minute, this
veals the bed to be fouled with oil. rate applying to the standard pressure type
Cloth strainers must also be removed fre- filter. As the mat of coagulated oil and ad-
quently for cleaning in an alkaline solution sorbing floc accumulates on the filter bed,
and may be provided in the double unit type the resistance to flow increases until a point
with provision for one unit out of operation is reached where the filter must be taken out
at a time. of service and backwashed. The frequency
As with skimming devices, units of this type of this operation will depend upon the load
are effective only in the removal of free oil imposed upon the filter. The usual practice
from the water. Emulsified oil is too finely is to backwash when the loss in head through
dispersed to be removed by such straining ac- the filter reaches 4 to 5 psi. For backwashing
tion. In some cases, the labor and handling purposes, hot water should be specified since
required is undesirable. Since oil contamina- the use of cold water tends to. "set" the co-
tion from reciprocating engines is usually in agulated oil retained on the filter bed.
the emulsified form, units of this type have a A caustic scouring tank is sometimes pro-
restricted field of application. vided for intermittent hot scouring of the
Emulsified Oil. Several methods are avail- filter for the removal of oil which has not
able which will satisfactorily break the oil- been removed by the normal backwashing
water emulsion, causing the oil particles to operations. The hot caustic solution remains
coalesce and be readily removed by the ad- in contact with the bed for approximately
sorbing medium. The usual systems involve one hour, dissolving and saponifying the oil.
the use of aluminum and iron salts for the The frequency of caustic scouring is depen-
purpose of adsorbing the oil. While ferric dent upon the quantity and the type of oil
sulfate has been used successfully, aluminum contamination of the condensate.
sulfate is the more commonly used material In certain cases where the volume of con-
for this purpose. Caustic soda or soda ash is densate is large and a high degree of oil
fed in addition to the coagulant in order to contamination exists, it is found advanta-
produce the floc of aluminum hydroxide or geous to employ flocculating basins in the oil
ferric hydroxide which coagulates and en- removal process. The chemistry and me-
meshes the oil so that it can be removed by chanics of this operation are fundamentally
filtration. The pH of the treated water is the same as previously described under co-
regulated at a point that insures the most agulation with the flocculation and coagula-
efficient coagulation. tion of the oil being accomplished in basins
In one type of unit, the coagulant and al- equipped for chemical feed and agitation. A
kali are fed to the filter influent in propor- retention time of one hour is usually pro-
tion to the flow of condensate. Conventional vided. If free oil is present in the condensate
pressure type filters are usually employed together with emulsified oil, mechanical
with the filter medium of anthracite coal. skimmers may be used to remove as much oil
Sand or gravel as a filter is undesirable due as possible prior to chemical treatment in the
to the pickup of silica from the filter bed by flocculating basin.
the hot condensate. Calcite filter medium is A variation of the above system involves
also undesirable because some of the filter the application of a certain amount of pre-
bed would dissolve in the condensate, intro- formed floc to the filter bed to form an initial
ducing calcium hardness. adsorbing mat, followed by the continuous
In design and operation the filters em- feed of a small amount of the preformed floc


------±-- EFFLUENT




Courtesy Infilco, Incorporated

Figure 19-2 • Preformed Floc Type Oil Removal Plant

Courtesy R. P. Adams Co., Inc.

Figure 19-3 • Diatomaceous Earth Filtration System

to the condensate passing to the filters. In specially designed pumps are available to
this instance, the coagulant and necessary handle such material. At the end of the filter
quantity of alkali are added in the predeter- run, newer designs of the filters permit the
mined quantity to a chemical mixing tank sloughing off of the filter cake by backwash
where the floc is formed and aged. Figure or air pressure with the filter mat dropping
19-2 illustrates a system of this type. into a hopper bottom and discharging to
This method of preparation and injection waste. After thorough rinsing, the unit is
of the floc eliminates the need for proportion- again ready for the precoat application.
ing the chemicals and results in appreciable In the event improper operation causes
savings in initial equipment costs. The filters the septa to become fouled, they are readily
used with this type system are usually the removed for chemical or mechanical clean-
standard pressure type although some manu- ing. The diatomite filters operate with con-
facturers supply units equipped with rotary siderably higher pressure drops than do
surface washers or rake agitators. These standard pressure filters. They may operate
special features are usually employed where with head losses as high as 50 psi before re-
the oil content of the condensate is high and quiring backwashing. This operation is pos-
difficulty in removing the oil and floc mat sible since the rigid support of the septa pre-
from the filter medium is expected. vents channeling and the filtrate will remain
During World War II, the need for im- clear even though the throughput decreases
proved oil removal methods was a primary as the pressure drop increases.
one in the marine field. This need spurred In oil removal practice, it is usual to em-
the application of diatomaceous earth as a ploy a precoat of filter-aid approximating
filter-aid and several types of filters were de- one to two ounces per square foot of filter
signed specifically for the purpose of oil re- area. The body coat will vary depending
moval on shipboard. Development has con- upon the quantity of oil to be removed. It
tinued with units of this type being designed is necessary to employ a body coat whenever
for stationary plant application. Most of the oil is present to any extent in order to main-
latest designs have eliminated former ob- tain the porosity of the filter cake and pre-
jections and provide easy removal and dis- vent separated oil from clogging the inter-
posal of the filter cake in a backwashing stices. In general, the length of run obtained
operation that does not require removal of on a diatomite filter depends upon the filtra-
the head of the filter or the individual septa tion rate, amount of filter-aid used and the
under ordinary circumstances. amount of foreign matter, such as oil, present
in the water.
The diatomite filters employ a permeable
base or septum to support the filter cake. A suitable grade of filter-aid specially
These septa include various types of filter processed for adsorption of oil is supplied by
cloths and papers, porous stone tubes, wire several manufacturers and adsorbs approxi-
cloth and wire wound tubes. The usual pro- mately one pound of oil for each pound of
cedure involves application of a precoat on material used. An application of this type
the septa with the filter-aid recirculated filter has produced a filtered condensate of
through the unit for several minutes to form practically "zero" oil content with an in-
a porous mat. The unit is then operated to fluent oil content as high as 50 ppm.
waste for several minutes to make certain Operation of these filters requires reason-
the effluent is clear before cutting it into the able care and adherence to operating instruc-
system. The usual practice in oil removal tions. However, their operation is no more
systems is to follow the precoat with a complicated than that of standard filters.
"body" coat or continuous feed of the filter- Chemical coagulation and filtration or use
aid into the condensate passing to the filter. of diatomite adsorption and filtration are the
Even though diatomaceous earth is abrasive, only certain means for the removal of ernul-

sified oil from liquid condensate. Naturally, The determination of oil or grease is based
systems of this type are a more costly initial upon refluxing through a trap for volatile
installation and require a greater degree of materials followed by solvent ,,extraction of
supervision than do baffle type separators or the remainder of the sample. Commonly
towel filters. Consequently, adsorption and used solvents are hexane, petroleum ether,
filtration systems are usually installed so as benzene, chloroform or carbon tetrachloride.
to handle only the contaminated condensate. These solvents exert selective extraction of
Oil free condensate such as turbine conden- specific greases and oily constituents. How-
sate is returned directly to the feedwater ever, nonoily materials such as phenolic com-
heater without passing through the oil re- pounds or colloidal sulfur may be selectively
moval system. By limiting the size of the extracted to varying degrees. The method of
equipment to the flow of contaminated con- sampling is of utmost importance. Solvent-
densate, the initial cost may be minimized. washed glass containers should be employed
The adsorption systems described have using a closure with an impervious liner such
certain limitations. Those systems using a as aluminum.
coagulant and alkali require rather rigid pH The definition of grease and oily matter
control and do not function well when the must be based upon the procedures used, the
pH of the condensate is variable, requiring source of contamination, the type of solvent
constant adjustment of the chemical feed. and other factors. The interpretation of the
This defect is true also for the preformed analytical results will vary accordingly. In
floc system since introduction of the pre- view of the multiplicity of the variables en-
formed floc to a condensate which does not countered, the Joint Committee on Uniform-
have a satisfactory pH value will result in ity of Methods of Water Examination has
the solution of the floc and resulting failure concluded that uniformity of methods for
of the system to coagulate and adsorb the oil. the determination of grease and oily matter
Diatomite filters eliminate the necessity for is not practicable on the basis of present
pH control, since no coagulant is employed. technical knowledge.
Since diatomite is a siliceous material, some
silica will be dissolved in the hot condensate. REFERENCES
Usually, this silica pickup is less than 1 ppm.
G. R. Bell, "Removal of Oil from Condensate by
Diatomite Filtration", Proceedings, Engineers Soc.
of Western Penna., pp. 123-137 (1952)
ANALYSIS FOR OIL V. J. Calise, "Removing Oil from Water by Floc-
culation and Filtration", Power Engineering, Vol.
The method to be used for the removal of 58, pp. 86, 87, 110, 138 (Apr. 1954)
oil and grease from water usually is predi- J. J. Maguire, "Oil Removal from Heating Sur-
faces", Southern Power & Industry, Vol. 74, pp.
cated upon a chemical analysis to determine 52, 54, 64, 67 (Aug. 1956)
the amount and type of contamination. Oil "Manual on Industrial Water and Industrial Waste
or grease is of animal, mineral or vegetable Water", 2nd Ed., pp. 400-402, Am. Soc. for Test-
origin and may be present in water as an ing Materials, Phila., Pa. ( 1959)
emulsion, free-floating or in some cases a
J. H. Richards, "Maybe It's Oil that Needs to be
Removed from Boiler Feedwater", Industry and
light petroleum fraction may be in solution. Power, Vol. 47, pp. 62-66 (Dec. 1944)


Theory Involved. . . . . . . . . . . . . . . . . . . . . . . . . 128

Removal of Carbon Dioxide and Ammonia. . . . . 130
Equipment. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131
Applications and Limitations. . . . . . . . . . . . . . . . 131
Condenser Deaeration. . . . . . . . . . . . . . . . . . . 133
Vacuum Deaeration . .................... 134

rium from being established. Therefore, the

M ECHANICAL deaeration, as applied to
boiler feedwater, is the removal of
dissolved gases such as oxygen, carbon di-
reaction will proceed to completion until
either the metal or the gas in ,the system has
oxide and ammonia by the process of raising been entirely exhausted.
the water to the saturation temperature for Ferrous hydroxide is an alkaline compound
the pressure under which the process is con- and its rate of solution depends upon the pH
ducted; employing an equipment design that of the water with which it is in contact. The
secures intimate mixture of the water and the lower the pH of the water, the more rapidly
scrubbing steam; and providing means for ferrous hydroxide goes into solution. Con-
the liberal venting from the system of the densate may contain carbon dioxide ( C0 2 )
gases released. in solution, which, by forming carbonic acid
The necessity for removing these corrosive (H 2 C03 ) , may lower the pH to a range such
gases from boiler feedwater in order to pre- as 5.6-6.5. This low pH is more conducive
vent corrosion is now fully recognized. Dis- to the solution of ferrous hydroxide. Hence,
solved oxygen will cause pitting and corrosion it is highly desirable to rid the system of car-
of the boiler feed lines, closed heaters, econ- bon dioxide so as to increase the pH of the
omizers, boilers, steam and return lines. If water.
carbon dioxide is also present, the two gases Ammonia may be extremely corrosive to
acting simultaneously may be considered copper and copper bearing alloys in the
from 10% to 40% more corrosive than would presence of moisture. Corrosion of valves,
be expected for the same quantity of the fittings, turbine blading, condenser tubes,
two gases acting individually. etc., may take place with sufficient ammonia
It has been determined that oxygen is contamination. In contact with copper, am-
relatively 5 to 10 times more corrosive than monia forms a complex copper-ammonium
equal quantities of carbon dioxide, and that ion [Cu (NH 3 ) 4 ++ ] and solution of copper
temperature is also an important factor. For and copper bearing alloys may be quite
example, a condensate in which is dissolved rapid. A source of ammonia in the system
equal molar concentrations of each gas at may be from breakdown of nitrogenous mat-
194 F is 2.0 to 2.5 times as corrosive as that ter under boiler temperatures and the evo-
same condensate at 140 F. Where corrosion lution of ammonia with the steam.
is caused largely by carbon dioxide, the
structure and composition of the steel are THEORY INVOLVED
factors in determining its life.
Dalton's Law states that the total pressure
When iron or steel is brought into contact
of a mixture of several gases is equal to the
with water, iron goes into solution. A soluble
sum of the pressures which each gas would
compound, ferrous hydroxide is formed as
follows: exert were it alone present in the volume oc-
cupied by the mixture.
Fe++ + 2H 20 = Fe(OH)z + 2H+
Henry's Law states that the concentration
ferrous ion + wa t er _ ferrous +hydrogen
- hydroxide ion of the dissolved gas in the solution is directly
proportional to the partial pressure of that
If dissolved oxygen is present in the sys-
gas in the free space above the liquid, with
tem, it will combine with the ferrous hy- the exception of those gases which chemically
droxide to form the insoluble compound, unite with the solvent. For example, carbon
ferric hydroxide, as shown:
dioxide and ammonia unite chemically with
4Fe(OH)z + 0 2 + 2Hz0 = 4Fe(0Hh water.
ferrous f- + wa t er = ferric Henry's Law may be written:
hyd rOXI'd e - oxygen h yd rOXI'd e
The oxygen, by removing the ferrous hy- -=K
droxide from the solution, prevents equilib- p

where K = solubility coefficient (constant for falling streams.

any one temperature) Oxygen similarly can be removed from
c = concentration of dissolved gas in water by passing another gas counter-flow to
solution falling streams of water. Air could not be
p = pressure of gas used for this purpose because of the high
For example, if 0.002 mol of a gas is dis- percentage of o;-;:ygen (21%) in air. Nitrogen
solved in a certain volume of water under could be employed but cost would make the
one atmosphere pressure, 0.004 mol will be process impracticable.
dissolved if the pressure is doubled. The most desirable and practicable method
Deviations from Henry's Law may be ex- of removing dissolved oxygen from boiler
pected when the dissolved gas undergoes a feedwater involves the use of the vapor of
secondary reaction with the solvent, as is the the solvent or steam as a scrubbing gas. This
case when ammonia gas is dissolved in water. choice of a scrubbing gas is logical since
Henry's Law applies only to the dissolved steam is the most economical gas that could
ammonia gas (NH 3 ) but the concentration be employed, and in the process of gas re-
in the solution is determined by analytical moval a thermodynamic advantage is secured
methods which include both the ammonia by heating the water with the steam.
gas and the ammonium hydroxide (NH40H) The mechanism of gas removal from water
formed. through the use of the scrubbing action of the
In accordance with the basic principles vapor (steam) of the solvent (water) in-
contained in Dalton's and Henry's Laws, it is volves two separate mechanisms as follows:
evident that removal of a dissolved gas from 1. Mechanical separation of gas bubbles re-
water can be effected by reducing the partial sults when saturation conditions are achieved
pressure of that gas in the surrounding at- and when the vapor pressure of the liquid
mosphere, regardless of the total pressure on solvent and the solution pressure of dissolved
the system. The simplest manner in which gases appreciably exceed the pressure im-
this can be accomplished is to bubble another posed on the system. Under these conditions,
gas through the water or to spray the water all gases becom,e insoluble. However, satu-
into a countercurrent flow of another gas, ration conditions alone do not assure com-
supplying in each case free venting of the gas plete gas removal. If only saturation condi-
from the system. In this manner, the con- tions were necessary for gas removal, there
centration of the dissolved gas in the sur- would be no need for the complicated tray
rounding atmosphere is reduced by dilution or jet design of feedwater heaters, since an
with the scrubbing gas. The partial pressure open tank, filled with water at 212 F at
of the dissolved gas in the surrounding atmos- atmospheric pressure, would accomplish com-
phere decreases and, in accordance with plete deaeration.
Henry's Law, to establish equilibrium the 2. Diffusion of gas particles to the sur-
concentration of dissolved gas in the water rounding atmosphere (without bubble forma-
decreases in direct proportion with its de- tion) results when the total internal pressure
crease in partial pressure. of the liquid solvent and dissolved gases is
Practical application of this theoretical less than or slightly more than the pressure
principle is obtained in the removal of free imposed on the system. Ammonia removal,
carbon dioxide from water by blowing air because of the extreme solubility of this gas,
countercurrent to falling streams of water as must be accomplished by diffusion rather
in a degasifier. Such a unit, as described than mechanical separation.
under cation exchange by hydrogen zeolite, In the removal of dissolved oxygen from
consists of slatted wooden trays to break up water, 90 to 95% of the initial oxygen con-
the falling water into thin streams and a low tent will leave the water immediately and
pressure blower to force air up through the without difficulty by mechanical separation

when the water is heated to the saturation nism of removal of these gases is primarily
temperature for the pressure imposed on the molecular diffusion rather than mechanical
system. The remaining 5 to 10% of the separation.
initial oxygen content must be removed When carbon dioxide is dissolved in water
through molecular diffusion. In ridding the it forms a relatively unstable compound, car-
system of this last 5 to 10% of the initial bonic acid (H 2 C03 ) • This acid ionizes into
oxygen content, it is obvious that the rapidity the bicarbonate radical (HC0 3-) and the
of oxygen removal will be a function of the hydrogen ion (H+). Further ionization of the
heater design and also of the rapidity with bicarbonate radical takes place to form the
which the surrounding atmosphere is changed carbonate radical (CO a= ) and the hydrogen
and gases vented from the system. ion ( H + ) as illustrated.
The practical considerations involved in COz + HzO P HzC03
the removal of dissolved oxygen from boiler carbon dioxide + water P carbonic acid
feed water may be briefly summarized as: HzC03 P H+ +
Heating the water to the boiling tempera- carbonic acid p hydrogen ion +
ture for the pressure under which the Heal-
process is conducted (saturation condi- bicarbonate ion
tions). Hco3- P col= +
Providing a heater design that secures in- bicarbonate ion p carbonate ion +
timate mixing of steam and water. H+
Continuously venting from the system the hydrogen ion
mixture of gases and steam. The hydrogen ion concentration or pH
controls the distribution of carbon dioxide
among these three forms. It is only in the
MONIA. Unlike oxygen, carbon dioxide and
form of carbonic acid (H2C0 3 ) that a gas
ammonia do not obey Henry's Law because
solution pressure is exerted and that carbon
they unite chemically with water. At elevated
dioxide can be removed by deaeration. Fig-
temperatures their deviation from Henry's
ure 20-1 illustrates the availability of carbon
Law decreases due to decreased chemical
dioxide for removal by deaeration and in-
combination with the solvent. The mecha-
dicates it is at the more acid pH values that
the most efficient removal of carbon dioxide
can be obtained.
When ammonia gas is dissolved in water,
pH !.s. "1.-Haco, OF roiAL co~ it forms an unstable compound, ammonium
r-.;;: l". hydroxide (NH 4 0H). Ammonium hydroxide
ionizes to form the ammonium ion (NH 4 +)
0 '\ '\. and the hydroxyl ion (OH-) as illustrated.
870 1\ r\ NH3 + HzO P NH40H
g. \ \ ammonia + water p ammonium hydroxide
,_ 0
\ NH40H P NH4 + +
0 •0 21S•F
r--' - 77•F
ammonium hydroxide + ammonium ion +
.. \ OH-
_i \ hydroxyl ion
\ The hydroxyl ion concentration controls
10 '\ l\.. the distribution of ammonia between these
......... ~ two forms. It is only in the form of ammo-
3.0· 4.0 6.0
7.0 B.O 9.0 nium hydroxide (NH4 0H) that a gas solu-
tion pressure is exerted and that ammonia.
Figure 20-1 • Per Cent C02 Available for Removal can be removed by deaeration. Figure 20-2


pH VS. '% NH40H OF TOTAL NH3 The conventional type open heater is not as

v efficient in the removal of oxygen and carbon
V; dioxide as is a deaerating unit. With deaerat-

1/ v
£ eo ing units, extraJactors of safety are incorpo-
;;t 70 rated in the design and a vent condenser is
t I employed so that a liberal amount of steam
77" F I /100°F may be vented from the heater without heat
~ 50
I II waste. Ordinary open heaters are not usually
z 40

designed to provide as intimate mixture of
steam and water and as efficient venting as is
I I required for more complete deaeration.
/ / The Standards of the Heat Exchange
~/ Institute define a deaerating heater as an
6.0 7.0 8.0 9.0 10.0 11.0 12.0
equipment for heating water by means of
steam, designed so as to insure the removal
Figure 20-2 • Per Cent NHa Available for Removal
from the water of the dissolved oxygen in
illustrates the availability of ammonia for re- excess of 0.03 cc per liter.
moval by deaeration and indicates it is at A deaerator is defined as an equipment for
the more alkaline pH values that the most heating water by means of steam and de-
efficient removal of ammonia can be ob- signed so as to insure the removal from the
tained. water of the dissolved oxygen in excess of


In accordance with the Standards of the

Heat Exchange Institute a standard deaerat-
ing unit for operation at atmospheric pressure
and above includes:
Deaerating unit.
Integral float or floats.
Piping between deaerating heater and/or
water inlet regulating valve and the vent
Water storage capacity not less than two
Weir type metering device with recording
and/ or integrating instrument.
Material of construction is usually cast iron
for working pressures up to 20 psi with rolled
steel plate for higher operating pressures.
It is standard practice to provide the
storage compartment of a deaerating unit
with a manhole, gauge glasses and fittings,
supporting feet and openings for water con-
nections conforming to the deaerating unit Courtesy Cochrane Division-Crane Company
capacity. Figure 20-3 • Tray Type Deaerating Heater

Figure 20-4 • Jet Type Deaerating Heater Courtesy Cochrane Division-Crane Company

0.005 cc per liter. "zero" oxygen feed water must be qualified

There are two principal types of deaerat- as to the specific test to be employed in
ing heaters-the tray type and the jet atomiz- measuring the oxygen content.
ing type. The tray type deaerating heater, as CoNDENSER DEAERATION. In central station
illustrated by Figure 20-3 spreads out the operation, deaeration is sometimes secured
water in thin films by means of baffles or by the use of surface condensers with de-
trays so as to insure intimate contact of the aerating hot ~ells. In such installations, a
steam with the water and to provide thorough separate deaerating feedwater heater is not
agitation. In addition, a relatively large employed. The condensate leaving the sur-
amount of steam is vented from the unit, face condenser is guaranteed not to exceed
passing to the vent condenser where the 0.03 cc per liter oxygen content. The con-
steam is condensed and returned to the denser design is such that the condensate
heater while the noncondensable gases are spills into the hot well through a steam at-
discharged to the atmosphere. The flow of mosphere, with liberated gases removed by
steam may be counterflow to the waterflow the air ejector. Because of the danger of air
or parallel flow may be employed. Satis-
factory deaeration is obtained with each de-
In the jet atomizing deaerating heater, as
illustrated in Figure 20-4, the water is sprayed
into a steam atmosphere and thereby heated
to within 2 to 3 F of the steam temperature.
This operation removes 90 to 95% of the
oxygen. The water is then delivered into a
high velocity steam jet and the water is
broken up, or atomized, by the action of the
steam jet securing complete deaeration in
this operation. Steam flow is countercurrent
to waterflow. As in a tray type heater, the
gases are passed through a vent condenser
and the condensate returned to the heater.
The atomizing deaerator, while subject to
a 0.5 psi pressure drop across the nozzle, is
especially suited to handling turbid waters
and those of high carbonate hardness which
would tend to plug the trays in a tray type
unit. The atomizing deaerator is also suited
for marine use inasmuch as the deaeration
will not be adversely affected by the roll of a
ship which would cause uneven water distri-
bution over the trays of a tray type unit.
In general it can be definitely stated that
the production of a deaerated water of a
"zero" oxygen content is not physically pos-
sible although the dissolved oxygen content
may be reduced to a value below that which
can be detected by chemical tests. Since an
absolute "zero" oxygen in a deaerator efflu- Courtesy Cochrane Division-Crane Company
ent is not possible, any specification for a Figure 20-5 • Vacuum Deaerator

leakage into the system, both on start-up and either by steam jet eductors or by vacuum
during normal operation, this system of de- pumps.
aeration is not considered as good a practice These units are usually not designed for
as the use of a deaerating feedwater heater. complete deaeration inasmuch as the size
VAcuuM DEAERATION. It is occasionally de- and cost of the apparatus increase sharply
sirable to deaerate water at temperatures be- for the removal of the last traces of oxygen.
low the atmospheric boiling point of water. Oxygen content has been reduced to 0.25 cc
The object is to render non-corrosive the per liter with the maintenance of a 28.5 inch
water in long pipe lines, for industrial proc- vacuum. Removal of small residual oxygen
ess uses or in certain power plant cycles. The concentrations can be secured economically
deaeration of cold water requires that an with the use of catalyzed sodium sulfite.
equipment design be employed which will
divide the water into small particles to fa- REFERENCES
cilitate gas removal, provide venting of the
· g<;tses from the unit and maintain a vacuum H. A. Grabowski, H. D. Ongman, W. B. Willsey
in the deaerator corresponding to the boiling and W. Nelson, "Problems in Deaeration of Boiler
Feedwater", Combustion, Vol. 26, pp. 43-48 (Mar.
pressure for the water temperature involved. 1955)
The incoming water is introduced usually A. L. Jones, "Deaerating Feedwater Heaters -
Spray Type vs Tray Type", Proceedings, Engineers
through a spray pipe and then falls over Soc. of Western Penna., pp. 41-55 (1949)
staggered wooden trays. The object of this A. E. Kittredge, "The Removal of Dissolved Gases
arrangement is to reduce the water into thin from Boiler Feedwater", Proceedings, Engineers
films and particles, aiding gas release. Water Soc. of Western Penna., pp. 105-114 (1941)
G. T. Skaperdas and H. H. Uhlig, "Corrosion of
vapor and gases released from the water are Steel by Dissolved Carbon Dioxide and Oxygen",
removed through the vacuum maintained, Ind. and Eng. Chern., Vol. 34, pp. 748-754 (1942)

Wet or Solution Type Feeders. . . . . . . . . . . . . . . 136

Shot Feed Systems .. ' .................... 136
Pot Type Feeders .. ................... 136
Water Jet Eductor... . . . . . . . . . . . . . . . . . 137
Automatic Shot Feed. . . . . . . . . . . . . . . . . . 140
Constant-Rate Feeders. . . . . . . . . . . . . . . . . . . 140
Controlled Volume Pump . .............. 140
Drip Feeder. . . . . . . . . . . . . . . . . . . . . . . . . 142
Proportional Feed Systems . ............... 142
Pot Type Proportional Feeds. . . . . . . . . . . . 142
Positive Displacement Pump .... ......... 142
Electric-Motor-Driven Pumps . .......... 142
Meter and Timer Controlled Pumps .. ..... 143
21 Thymotrol . ......................... 143

Chelllical Feed Systellls

Pneumatic Proportioning Control. . . . . . . . . 144
Hydraulic Feeders. . . . . . . . . . . . . . . . . . . . 144
Decanting Feeders . ................... 144
Double-Orifice Feeders. . . . . . . . . . . . . . . . . 146
Dry Type Feeders. . . . . . . . . . . . . . . . . . . . . . . . 148
Volumetric Feeders .. .................... 148
Gravimetric Feeders. . . . . . . . . . . . . . . . . . . . . 148
Automated Feeding Equipment . .............. 149
Continuous Monitoring ................... 149
Boiler Systems. . . . . . . . . . . . . . . . . . . . . . . . 149
Recirculating Cooling Systems. . . . . . . . . . . . 149
Control Mode...................... . . . . 149
Sampling . ............................. 150

T HE proper introduction of chemicals is a

very important phase of the treatment of
water and sewage. A predetermined concen-
water passes through stage heaters or econo-
mizers, etc. For example, the feed of sodium
sulfite for oxygen elimination should be con-
tration of the chemicals must be established tinuous to the storage section of the heater or
and maintained to secure the desired reac- the suction side of the boiler feed pump for
tions. Thus, accurate and satisfactory treat- most effective results. A suitable chemical
ment application has required the develop- pump or similar device is required for the
ment of many types of chemical feeders. continuous feed of the sulfite solution. On the
Chemical feeding equipment can be di- other hand, the phosphates normally used for
vided into two general classifications: dry boiler scale prevention may be shot fed
feeders and wet or solution feeders. Either type through the feedlines to the boilers, provided
may be further classified on the basis of method stage heaters or economizers are not used and
of feed, that is, constant rate or proportional. the "number of boilers is restricted to two or
Dry feeders are manufactured in two designs, three.
volumetric and gravimetric. Solution feeders SHoT FEED SYsTEMS. These systems are sub-
may be either of the gravity or pressure type divided into pot type feeders, eductor sys-
and under some circumstances are used as a tems and automatic shot feed arrangements,
"shot feed" rather than constant rate or pro- in which the object of the feeding device is to
portional. rapidly introduce a unit charge of chemical
Dry chemical feeders are usually applied in treatment. Usually these systems are em-
external treatment where large quantities of ployed to avoid feedline deposits that might
chemicals must be handled. The application result from continuous chemical feed.
of this equipment is common to sewage and Pot Type Feeders. Probably the oldest meth-
trade waste treatment plants as well as water od used for the introduction of chemicals to
works and large industrial water conditioning boilers is the pot type shot feeder which
plants. In these installations the demand for consists of a pressure tank and fittings as
chemicals such as alum, lime and soda ash is shown in Figure 21-1. The chemical charge is
frequently quite large. prepared by dissolving the chemicals in a
The solution type of feeder provides greater bucket and then filling the pressure tank with
convenience and accuracy when smaller the solution. It is necessary, of course, to drain
quantities of chemicals are to be employed. water from the tank and open the air vent
Also, some types of chemicals can not be prior to pouring the chemical solution into
handled satisfactorily with a dry feeder. Solu- the tank. The illustration shows the feeder
tion feeders are used exclusively for the addi- connected across the boiler feed pump. Feed-
tion of secondary or internal treatment to water from the pressure side of the pump
boiler feedwater systems for the prevention of forces the chemical solution into the suction
scale, corrosion and carryover. The primary side of the pump. Within a few minutes, the
or external treatment of boiler makeup solution will be washed out of the pressure
water, such as softening, also usually employs tank and will be injected into the boiler
the solution type of feeder since the quantity shortly thereafter.
of chemicals required is relatively small. A system of this type is usually satisfactorily
applied when the number of "shots" per day
WET OR SOLUTION TYPE FEEDERS does not exceed three or four. If more fre-
quent "shots" are required, it is possible to
The type of boiler chemical feed system that install a large mixing tank above the pressure
should be employed in any individual instance tank, similar to the tank shown in Figure
depends primarily upon the amount and type 21-2, thus eliminating the inconvenience of
of chemicals and for what purpose they are dissolving chemicals each time an addition of
being used; type of boilers; whether the feed- treatment is necessary.

being taken from the pressure side of the

boiler feed pump, the outlet discharging with-
in the boiler drum through a separate chemi-
cal feedline. With the difference in pressure
between feed pump discharge and boiler
drum, this type of installation is quite work-
able. By this means, the necessary treatment
for each boiler can be injected at will. The
system, while practical, is rather cumber-
some in that an individual charge must be
prepared for each boiler.
As is indicated by the foregoing discussion,
shot feeders are not suitable for the continu-
ous feed of chemicals such as sodium sulfite
(for oxygen removal), caustic soda or soda
ash (for pH adjustment) or any of the surface
active organic and inorganic chemicals (for
prevention of deposits in distribution lines) .
The shot feed of precipitating chemicals
through equipment such as economizers or
PUMP DISCHARG£---._ stage heaters is not recommended, inasmuch
as the water flow through units of this nature
Figure 21-1 • Pot Type Shot Feeder is sufficiently slow to permit premature re-
action of the treatment with the incrusting
salts, forming deposits in heat transfer equip-
Pot type feeders are frequently installed, in ment. This condition is also accelerated by
the case of small boiler plants, on the suction the increase in temperature associated with
side of return condensate pumps or in a by- the operation of units of this type.
pass arrangement on the raw water line, Water Jet Eductor. A second method of shot
where raw water pressure forces the chemical feed utilizes a water jet eductor, installed as
charge into the boiler upon by-passing the shown in Figure 21-3. The eductor works on
water through the feeder. the same principle as an injector, using a
Shot feeders are convenient and economical stream of water under pressure instead of
for injecting precipitating chemicals such as steam. Water from the pressure side of the
phosphates, caustic soda and soda ash to boil- boiler feed pump passes through a pressure
ers, thus reducing the hazard of feedline de- nozzle within the eductor body, producing a
posits caused by premature reaction of these high velocity jet. The high velocity jet creates
chemicals with the incrusting salts present in a vacuum, drawing in the treating solution,
the feedwater. If the chemical treatment is and injecting it into the suction side of the
added frequently, the treatment will usually boiler feed pump. The eductor of the size
average itself among the boilers on line. How- ordinarily used has a capacity of approxi-
ever, the number of boilers treated in this mately 6.5 gpm.
manner generally should not exceed three, Although the eductor is of advantage in
otherwise it will be extremely difficult to that any amount of treatment may be in-
maintain consistent chemical balances in all jected, as compared to the restricted quantity
boilers. of the "pot" type feeder, the disadvantages of
Where a large number of boilers are in use, shot feed are inherent in this system as well.
it is relatively common practice to install a In addition, a pressure differential of at least
feeder for each boiler, the inlet to the feeder 4 to 1 between the pressure and suction sides







Figure 21-2 • Pot Type Shot Feed System Employing Treatment Storage Tank

of the pump must be available for satisfactory do not prove suitable for continued use, since
operation of the eductor. Also, it is not pos- efficiency greatly depends upon proper align-
sible to introduce the eductor discharge direct ment of internal parts and freedom from de-
to the boiler drum, in the average plant, be- posits. Chemical treatment may attack the
cause of insufficient pressure differential. metal used in construction of the internal parts
Steam injectors are sometimes used in the and contact of high-temperature steam with
same manner as eductors. Usually, injectors treating solution frequently causes deposits.

- FILLING LINE (preferably condensate)






LEAST 4 : 1


1" X 1%"

llfz" PIPE



Figure 21-3 • Shot Feed System Employing Eductor


Automatic Shot Feed. A small gear pump,

timer controlled, is frequently employed for
automatically shot feeding treatment to the
suction side of the boiler feed pump. The
timer is regulated to operate the pump for a
variable length of time up to a maximum of
ten minutes, once each hour. Water charac-
teristics would determine whether more fre-
quent operation would be possible and
whether a "shot" of as long as ten minutes
duration would be permissible.
This automatic shot feed method permits
charging of the chemical tank once per day.
Due to the frequent additions to the boilers, a
··larger number of boilers can be treated more
efficiently than in the case of the pot and
eductor feeders.
Gear pumps of this nature usually operate
most efficiently with discharge pressures of
100 psi or less.
CoNSTANT-RATE FEEDERS. A constant-rate
feeder is one which will continuously add a
treating solution at a fixed rate for any given
control setting. The rate of feed is adjustable
by manual control but is not in proportion to
changing water flow. Controlled volume
pumps and drip feeders are the most common
units falling under this general classification.
Controlled Volume Pump. The most common
example of this type of feed employs a motor-
driven chemical feed pump to inject a specific
quantity of chemicals into a system at ·a given
A chemical feed system of this nature is
used most frequently in the continuous Figure 21-4 • Package Chemical Feeding System
feed of sodium sulfite to feedwater heaters or
feedlines and for discharge of precipitating feeding system. Figure 21-5 shows the recom-
chemicals such as phosphates within the boiler mended piping for any applicable positive dis-
drum against boiler pressure. When the rate placement pump with separate chemical tank.
of flow is relatively constant, a system of this It is necessary to supply one pump or one
nature can be used to introduce treatment to pumping cylinder for each point of applica-
once-through cooling and water distribution tion of treatment. The capacity of the pumps
systems. This system is also used quite fre- used is generally in the range of 3.0-10.0
quently to insure introduction of the proper gph per cylinder. It is impossible to throttle
amount of chemicals to a circulating cooling the discharge from one cylinder satisfactorily
water system. so that several points will receive an equal
Figure 21-4 illustrates a pump of suitable amount of treatment. The discharge rate is
type, shown as a part of a complete packaged too low to permit satisfactory division of flow,




1/2" LINE TO





1/2" TO 1/4"

1/2" TO 3/8" 1/2" LINE TO



Figure 21-5 • Piping Diagram for Chemical Pump Installation

even when using various types of spt;cially trol tests are conducted with sufficient fre-
designed flow control valves. quency, so as to maintain chemical balances
Employment of this system of feed readily in the system within prescribed limits. While
approaches proportional feed where the there will be periodic variations, this type of
chemicals are injected into a large body of system will maintain chemical balances within
water, such as contained in a boiler or cir- limits, on the average. Usually, a suitable
culating cooling system. It is common practice factor of safety is incorporated when pre-
to maintain certain chemical residuals in the scribed limits are established by the water con-
water being treated. In the average plant con- sultant. During the course of operation, there-

fore, sufficient treatment is present at all times The degree of proportionality is question-
and adjustments are readily made if the con- able at all times, for there is little control over
trol tests so indicate. These adjustments in- the solution rate of the briquettes or of the
volve changing the length of stroke on the chemicals incorporated in them. This system,
pump or changing the strength of chemical while it is classified as proportional, is con-
solution being pumped. sidered only roughly so and cannot be applied
Drip Feeder. A rather crude type of constant- where appreciable accuracy of feed is re-
rate feeder, described as a drip feeder, con- quired.
sists of a suitable tank with bottom take-off A further adaptation of the pot type pro-
and control valve for governing solution dis- portional feeder is the type commonly used to
charge. By controlling the flow of treating proportion alum and alkali to water systems
solution by means of a needle valve, an almost for coagulation, or for oil removal in feed-
constant volume of solution is added over a water systems. This system also consists of a
given period of time. As the tank empties, the pot in which potash alum or ammonium alum,
solution head pressure is reduced, thus re- in lump form, or sal soda (crystalline
ducing the rate of feed. Dependent on total sodium carbonate) are placed. An orifice is
solution head pressure above the point of ap- installed in the raw water line, and a line from
plication, this reduction in feeding rate may the upstream side of the orifice leads to the
or may not be of significance. This feeder is a under-side of the tank. From the topside of
gravity type and finds application frequently the tank, a discharge line returns to the water
in the addition of acid and other chemicals to line just after the orifice. The resulting pres-
circulating cooling systems, air washers in air sure differential causes a small stream of water
conditioning systems, and addition of sulfite to flow through the pot feeder, gradually dis-
and like chemicals to some point in a feed- solving the crystalline chemicals contained
water system permitting gravity flow. therein and introducing the treating chem-
While a feeder of this type is readily fabri- icals to the water line.
cated, in actual use it requires constant at- This type of feeder is roughly proportional
tention to make certain the discharge line is and can be used in any installation where ac-
not plugged with some foreign material. curate proportioning is not necessary and
PROPORTIONAL FEED SYSTEMS. There are a where the chemicals to be fed are obtainable
considerable number of systems for achieving in large crystalline or lump form. Dry or
feed of chemicals proportional to the flow of powdered materials cannot be used in a feeder
water, involving a great number of mechan- of this type.
ical and electrical principles in their opera- Positive Displacement Type Pump. Pumps
tion. This discussion will be confined to the are available that can be driven by a ratchet
more commonly used devices and the prin- arm connected to a reciprocating pump or
ciples upon which they are based. engine. These pumps can also be driven by a
Pot Type Proportional Feeds. Probably the belt and pulley for connection to motor driven
simplest type of proportional feeder is the pot pumps which operate intermittently or under
type feeder with. a special opening in one constant head conditions. By so connecting
end to permit charging with chemicals in either type of pump, feed of treatment is ac-
briquette or lump form. A portion of the complished in proportion to the flow of water
water to be treated is allowed to pass through through the main pump. Such a system is
the tank, gradually dissolving the chemicals. quite suitable for the continuous and propor-
Feeders of this type are used for continuous tional feed of chemicals to boiler feedlines,
and proportional feed of treating chemicals once-through cooling systems, and the like.
to boiler feedlines for protection of feedlines, Electric-Motor-Driven Pumps. An electric-
to hot and cold water distribution systems and motor-driven positive-displacement pump can
various types of cooling systems. be used to provide proportional feed of chem-

icals to a well supply under certain conditions. ing upon the time cycle, the three-way valve
The chemical pump can be controlled by the permits chemical solution or condensate to
same electrical circuit that operates the well flow to the suction of the chemical pump.
pump, so that the chemical pump operates It is intended that this system prevent
only when the well pump is operating. Where plugging of the chemical feedlines by flushing
several well pumps are used, the system may periodically with condensate. However, con-
not be applicable. A system of this nature, siderable maintenance is usually involved with
when adaptable, provides accurate propor- operation of three-way valves. If the chemical
tioning of chemicals for treatment of the well solution is made up originally with condensate
water for prevention of corrosion, scale for- and the solubility of the chemicals used is not
mation and to retain iron in solution. exceeded, no difficulties will be experienced
Meter and Timer Controlled Pumps. This with plugging of chemical feedlines. In addi-
system utilizes a water or steam meter tion, the chemical feedline must be properly
equipped with an electrical contacting device located within the boiler drum, in order to
that makes contact after passage of a specific further prevent the possibility of plugging of
amount of steam or water in the lines. Upon the line.
each contact, a stop-cycle timer is started, Thymotrol. An electronic system employing
which operates a complete cycle. The timer the Thymotrol, as developed by the General
is adjusted to permit the operation of the Electric Co., has been adapted to the regula-
motor-driven pump a pre-set number of min~ tion of positive-displacement type pumps. Al-
utes for each contact. The time cycle is readily ternating current available at the plant is con-
adjustable to provide flexibility. For the appli- verted to direct current, full wave rectifica-
cation of chemical treatment to boiler water, tion being employed. A potentiometer is
the control system is applied to steam, feed- mounted in the water meter and governs the
water or makeup water meters. The system speed of the direct-current motor operating
is usually established so that the pump will the pump. The setting of the potentiometer is
operate approximately 75 per cent of the total immediately changed with any change in rate
time, allowing ample capacity for emergencies of water flow. Since the speed of a pump is
and unusual demands on the system. constant for any potentiometer setting, re-
This type of system is not suitable for pro- gardless of load, and since the quantity of
portional feed of treatment to once-through liquid pumped by any positive-displacement
systems unless the treatment is injected ahead pump is proportional to the speed of the
of a large storage tank where adequate mixing pump, accurate proportioning of chemical
can be accomplished. Also, for proportioning feed to the water flow is achieved with imme-
phosphates and other treatment chemicals di- diate compensation for changes in rate of flow.
rect to the boiler drums, it is necessary to This system has been applied in a number
supply a meter, timer, and pump for each dis- of instances requiring considerable accuracy,
charge point; this makes such an installation such as proportioning of lime and soda ash in
rather expensive. If makeup water character- the lime and soda ash softening process. The
istics are changeable or if there is a wide vari- system can be applied quite readily in the
ation in amounts of return condensate, boiler treatment of large quantities of water, such as
water balances may prove to be quite variable, once-through cooling supplies that vary in
even though treatment is added proportion- flow, and where a suitable mixing tank or
ally. reservoir is not available. Typical examples of
Another variation of this type of system application of Thymotrol proportioning con-
permits the motor-driven pump to operate sist of the addition of acids or alkalies in the
continuously. A meter controls the timer as adjustment of the pH of a treated water,
before, but in this instance the timer controls and the addition of stabilizing agents for
the functioning of a three-way valve. Depend- prevention of scale in heat exchange units

and similar equipment. and motor-driven agitator for handling slur-

Pneumatic Proportioning Control. Standard nes.
meters are available which transmit control- The chemical solution is drawn off either
ling air signals to a stroke adjusting mecha- at a constant rate, or proportionately, by
nism installed as part of a constant speed means of a swing pipe lowered by a cable
positive displacement pump. The adjustable and drum arrangement, which in turn is
stroke can be linked to a flow meter or other controlled by a small fractional-horsepower
unit, such as a pH controller-recorder. When motor geared to very low speed. Figure 21-8
flows in the line are quite variable and it is shows a feed system of this type. This par-
desired to control pH, a flow meter as well as ticular type is especially applicable for the
a pH controller are used to provide a signal to feed of corrosive solutions such as acids, alum
the pump stroke mechanism. Both variables and ferric salts. By the use of rubber tubing,
are transmitted to a ratio controller which the contact of the draw-off system with the
then sends an air signal adjusting the stroke corrosive liquid is limited to the overflow
· in relation to both flow rate and pH. Figure box, which is lowered by the control device.
21-6 shows such a controlled pump, respond- When operating as a constant-rate feeder,
ing to two variables. the decanting arm is lowered at a constant
Hydraulic Type Feeders. Figure 21-7 illus- rate by the control mechanism. This rate is
trates an economical and effective feed system manually adjustable. If proportional feed is
frequently utilized for chlorination and am- desired, a suitable water meter, equipped
moniation of water supplies, pH adjustment, with an electrical contactor, which makes
sterilization of sewage treatment plant effluent contact upon passage of a specific quantity of
and similar applications. The system may be water, is used to control the operation of the
used for the proportional feed of chemicals lowering motor. A predetermined amount of
to water lines for any purpose if the ma- solution is thereby drawn off for each contact
terial used in constructing the reagent head made by the water meter.
of the pump is selected accordingly. The discharge from the decanting arm can
Most of the standard meters may be be allowed to gravitate to the point of treat-
adapted as the control medium. The meter ment application. Frequently, however, the
does not supply the power to operate the discharge is introduced to a pump suction
pump but simply controls the inflow of pres- box. Several pumping arrangements exist,
sure water to the hydraulic cylinder. The depending upon the service requirements. In
design of the pump is such as to permit in- the case of lime and soda ash proportioning,
jection of treating solution into the pressure for example, a centrifugal pump is used
line. The pump is normally actuated by the which operates continuously, pumping the
pressure of the fluid present in the main line decanted solution to the point of discharge
but can be separately actuated by air, steam, with addition of dilution water, as necessary,
oil or water. controlled by a float in the suction box.
The hydraulic system is usually applicable A proportional-feed system of this type is
to lines operating at pressures of 100 psi or sometimes used for adding phosphate and
less. The addition of the pump control mech- other treating chemicals to boilers, boiler
anism does not affect the meter accuracy to feedlines or other water lines, using an inter-
any appreciable degree. mittently-operated piston type pump, con-
Decanting Type Feeders. This type of chemi- trolled by a float switch in the suction box.
cal feed system involves the use of a constant- In some cases, a high-capacity chemical
diameter tank fabricated of material suitable pump is employed, discharging to a common
for the chemicals to be handled. The tank header equipped with flow control valves to
will be flat-bottomed for handling solutions permit division of flow to a number of dis-
or will be equipped with a rounded bottom charge points while using only one pump.




'--:::::;;;;===~'-=-:I'~'-------===--- ----- _J



Courtesy The Foxboro Company

Figure 21-6 • Automatic pH Control with Wide Flow Variations

Courtesy B-I-F Industries, a Division of The New York Air Brake Company

Figure 21-7 • Meter-Controlled Hydraulic Type Proportional Feeder


the pressure lines shunts a portion of the

~ Electrical
Contact Head flowing water into the top of a pressure-tight
chemical feed tank, in which the chemical
solution occupies the complete volume. As
flowing water enters the top of the pressure
solution tank, an equal quantity of chemical
solution is released from the tank at a bottom
connection, entering the low pressure tap on
the downstream side of the secondary orifice.
The equipment is so designed that the
amount of chemical solution injected is rea-
sonably proportional to flow.
When recharging the feeder, chemicals
are dissolved in an upper tank and then
drained into the pressure solution tank. The
density of the chemical solution is sufficient
to support a special float installed in a gage
glass. The float will normally sink in water.
Since a definite line of demarcation exists
between the water and chemical solution at
all times, the float will show the amount of
chemical solution present.
Figure 21-8 · Decanting Type Proportional Feeder
Final control over the chemical solution
The decanting type feeder is used most is obtained by adjusting a needle valve
frequently as a proportioning feeder for line, governing the water flow into the pressure
soda ash, coagulants, acids and other chemi- solution tank. The feeder usually must be
cals applied in the external conditioning of recharged at least once daily.
water. A definite advantage is the fact that Double-orifice type feeders are frequently
only one meter and contactor is needed to used for the application of various chemicals
control the operation of several decanting to once-through cooling water and domestic
feed mechanisms. It is not generally applied water systems to condition these waters for
for addition of phosphates and other chemi- the prevention of scale and corrosion. This
cals to the boiler systems due primarily to type of feeder is also employed for propor-
availability of other equipment which is more tioning of alum and soda ash ahead of coagu-
convenient and less expensive. Also, dis- lation and oil removal systems.
charge of phosphates to feedlines is undesir- Use of these feeders is generally restricted
able and proportioning feed of treatment to in boiler feed systems by pressure limitations.
more than one boiler with one pump and one The feeder manufactured by the average
control meter again introduces the objections supplier is usually built to operate at pres-
outlined in the discussion of systems operating sures not greater than 100 psi. Also, the pres-
in an essentially similar manner. sure drop over the primary orifice is sufficient
Double-Orifice Type Feeders. The double- to cause steam binding of the boiler feed
orifice type feeder is a solution feeder which pumps if a sufficient static head does not
normally proportions with reasonable per exist between the feedwater heater and the
cent accuracy. The principle of operation suction of the feed pumps. Application of
involves passage of water through a primary these feeders for continuous feed of phos-
orifice, creating a pressure differential over phates and other precipitating chemicals to
the orifice and a flow in the lines connected boiler feedlines is not desirable because of
to the pressure taps. A secondary orifice in the danger of feedline deposits.

Fig. 21-9-Methods of Feeding for Common Treatment Chemicals


Alum Coagulation Dry Coagulation and sedimentation systems,

Color removal Pot Type proportioning prior to pressure filters for removal of
Double Orifice suspended matter and oil.

Sodium Aluminate Coagulation Dry Usually added with soda ash Ia softeners.
Decanting Used to some extent for internal boiler
Pumps water treatment.
Shot Feeders

Coagulant Aids Coagulation Pumps As primary coagulants, to chemical mix-

ing secfion.
As coagulant aid, to flume after primary
chemical mixing or to flocculation zone of

Ferric Salts Coagulation Dry Prior to coagulation and filtration systems.

Color removal Decanting
Oil removal Double Orifice

Lime pH adjustment Dry Prior to coagulation systems; to softeners;

(Hydrated) Softening Decanting to treated water lines for adjustment of
Special pumps (slurry) pH.

Soda Ash pH adjustment Pot Type Prior to pressure filters.

(Crystalline) proportioning

Soda Ash pH and alkalinity All types of wet To domestic systems, feedlines, softeners,
(Anhydrous) adjustment and dry feeders coagulation and filtration systems; boilers.

Caustic Soda pH adjustment All types of solution To softeners; oil removal systems; domes-
Alkalinity adjustment feeders tic water systems; boilers.

Acid Feed pH adjustment Decanting Treated water lines, prior to degassifiers

(H2S04) (Sulfuric Acid) Reduction of Pumps or deaerating heaters; H3P04 to phos-
(H3P04) (Phosphoric Acid) alkalinity phate softeners (for both softening and
(NaHS04) alkalinity reduction); tower sump or basin.

Surface Active Prevent calcium All types of Treated water lines.

Phosphates carbonate deposits proportional
Eliminate "red water" solution feeders

Orthophosphates Prevent scale in Pumps Added continuously to boiler drums; shot

(Monosodium Phosphate) boilers Shot feeders fed to drums or boiler feedline.
(Disodium Phosphate)
(Trisodium Phosphate)

Chelating Agents Prevent scale in Pumps Added continuously to feedline, on pres-

boilers and sure side of feed pumps.
Sodium Sulfite Prevent corrosion All types continuous Storage section of deaerating healer;
due to oxygen in solution feeders suction or pressure side of boiler feed
boilers, feedlines, pumps.

Sodium Nitrate Inhibition of All types of solution Any point in boiler feedlines or direct to
embr'ittlement feeders boilers.


Sodium and Potassium Corrosion inhibitor All types of To brine systems and various circulating
Chromotes solution feeders cooling and hot water systems.

Reactive Colloids Coagulation All types of To boiler feedlines; circulating cooling

(Sodium Mannuronate) Particle absorption solution feeders systems.
Protective Colloids and adsorption

Amines and Related Prevention of Shot Feeders Application depends upon material used.
Organic Compounds return line Pumps Some materials may be added Ia boilers
corrosion and volatilize with steam; others are
added to steam line direct, requiring

DRY.TYPE FEEDERS tency of the chemical being fed. If the

Dry-type feeders are generally classified as density or weight of the chemicals per cubic
volumetric type and gravimetric type; and foot is subject to any great variation, then
they are usually warranted in the larger-size this same variation will appear in the rate of
plants where the quantities of chemicals that chemical feed. Certain chemicals tend to
must be added to a system exceed approxi- "arch" or "hang" in the feed hopper, re-
mately 10 lb. per hr. quiring agitators or vibrators to be installed
in the hoppers. Care must then be exercised
VoLUMETRIC FEEDERS. In one type of volu- to prevent packing of the chemical, for this
metric feeder, the chemical is discharged would directly affect the accuracy of propor-
from a hopper onto a circular table, which tioning. The normal accuracy is within 5 per
revolves at a slow and constant rate of speed. cent.
One or more multi-section feed knives extend
GRAVIMETRIC FEEDERS. Gravimetric feeders
on the table, deflecting an amount of chemi-
measure the chemical to be fed by weight,
cal dependent upon the position of the knives.
resulting in accuracy of proportioning within
The feeder can be calibrated for use with
any chemical of relatively uniform composi- 1 to 2 per cent. A typical gravimetric feeder
tion and density. Once calibrated, the num- is shown in Figure 21-10. The chemical is
ber of pounds of chemical that will be de- fed to a traveling belt, the amount of chemi-
livered per unit of time, for the various cal admitted being controlled by a gate,
settings of the feed knives on the rotating which in turn is regulated by the total weight
table, will be constant. of the belt assembly and chemical. A constant
Chemicals rernoved from the rotating table weight of chemicals is maintained on the
are dissolved or suspended in slurry form by belt and variation in rate of feed to the
introduction to a dissolving pot agitated by treater is obtained by varying the speed of
a vortex of dissolving water or by a small the belt.
motor-driven agitator. There are several variations of this prin-
Still another type of volumetric feeder em- ciple, but essentially gravimetric feeders are
ploys an oscillating plate below the chemical scales, balanced in such a manner as to as-
hopper. The length of stroke of the plate and sure the delivery of the desired weight of
depth of chemical on the plate are both ad- chemicals to the system. The chemicals dis-
justable in order to obtain varying rates of charged by a gravimetric feeder are placed in
discharge. solution or suspension form, in the same
Accuracy of any volumetric type feeder is manner as for a volumetric feeder.
dependent to a large extent on the consis- Gravimetric feeders are considerably more

expensive than volumetric types, hence are addition which the analysis represents.
applied only to large systems where the ad- Conductivity, which may be related to the
ditional accuracy is required or when chemi- dissolved solids content of a solution, is also
cals are used whose variable density prevents a widely used measurement. This determina-
use of volumetric feeders. tion may be applied for continuous monitor-
ing of the total dissolved solids concentration
of boiler water or for automatic control of
Many installations are now equipped with continuous blowdown. Specific conductance
some type of device which automatically can serve also as the basis for monitoring
controls the operation of feeding equipment. condensate, and to actuate an alarm or to
CoNTINuous MoNITORING. The first step in divert the condensate stream when impuri-
providing complete and automatic control of ties are present.
the treatment program is the installation of Recirculating Cooling Tower Systems. The
continuous indicating and/or recording ana- class of analyzers noted in the case of boiler
lyzers. These analyzers are being applied not systems can be applied for the monitoring
only in monitoring plant influents and ef- and control of treatment programs for cool-
fluents, but also in the monitoring of external ing water systems. The differences are the
and internal water treating systems. Invalu- range of measurement and the parameters
able assistance can be gained by the judicious most critical to cooling systems (e.g. chro-
selection and installation of a continuous mate level, pH and conductivity).
analyzer to monitor critical variables in a Measuring instruments for pH and conduc-
water system. tivity utilize electrodes which are specific for
Analyzers commonly use a sampling tech- the parameters desired. In the area of chro-
nique whereby a discrete sample of the water mate, an electrode has been perfected which
is obtained, reagents are added in proportion does not require the addition of any reagent
to the sample and sufficient time is allowed to provide the measurement. This analyzer
for reactions. The sample-reagent solution is determines the chromate level on the basis of
then discharged into a measuring cell where the natural color imparted to the water by
a photocell determines the concentration of a chromate treatment.
specific parameter. The measurement is in- It appears that more investigation has
dicated and/or recorded, and serves as the taken place in the area of providing a com-
basis for manual adjustments or for remote- pletely automatic system for the control of
automatic regulation. cooling tower treating programs than in any
Continuous monitoring provides an extra other treatment area. Progress is evidenced
measure of safety by providing around-the- by the development of a completely inte-
clock surveillance of the system. Closer regu- grated system of critical parameter measure-
lation of treatment parameters is gained, ment and control, as shown by Figure 21-11.
thereby ensuring proper system protection CoNTROL MoDE. Most treatment control sys-
with minimum expenditure for treatment tems function properly using the simple on-
chemicals. off control mode. The method is suitable due
Boiler Systems. Continuous monitoring and to the large capacity of the water system
control devices are available for the measure- when compared to the relatively small quan-
ment of constituents, such as silica (control tity of chemical addition or blowdown re-
of demineralizer regeneration) , hardness quired by the treatment program. In fact,
(softener regeneration), phosphate (internal the sampling system used to provide a repre-
treatment control) and sulfite (oxygen scav- sentative sample of the water being controlled
enger control) . These parameters are con- is often the most distorting part of the whole
tinuously monitored, and serve as the basis analysis and control system.
for control of the specific unit or treatment Not all treatment systems can be controlled

with the on-off mode. There are occasions,

such as feed of sulfite to a deaerator and
treatment to a once-through cooling system,
where proportional control is a requisite.
However, on-off control can be applied quite
successfully on most systems.
SAMPLING. The sample to the analyzer must
be representative of the water in the system
under control. Care and precaution must be
taken in the selection of the sampling point
and in "sample preparation prior to introduc-
tion into the analyzer or to the electrode.
This sample preparation may involve filtering
of the sample, pressure and temperature re-
duction and, in some cases, flow control.
The manufacturer of the analyzer often
provides the sampling system. Regardless of
the supplier, every recommendation by the
manufacturer of the control system regarding
sampling should be followed to ensure ac-
curate and reliable measurement and control.
R. P. Lowe, "Chemical Feed Systems," Proceed-
ings, Engineers Soc. of Western Penna., pp. 155-
170 (1949)
J. H. Richards, "Chemical Feed Systems in
Modern Water Treatment," Power Generation,
Vol. 52, pp. 62, 63, 122-125 (Aug. 1948) and pp. Figure 21-11 • Modular Automation System
61-63, 118, 120, 124, 126 (Sept. 1948)
R. T. Sheen, "Liquid Chemical Feed Systems,"
Presented at 8th Ann. Water Conf., Engineers
Soc. of Western Penna., (1947)


Courtesy B-1-F Industries, a Division of The New York Air Brake Company
Figure 21-10 • Belt Gravimetric Feeder

Mechanism of Scale Formation .... ........... 152

Effect of Boiler Scale. . . . . . . . . . . . . . . . . . . . . . 153
Mud, Silt, etc. . . . . . . . . . . . . . . . . . . . . . . . . . . . 154

Boiler Scale Control
Calcium and Magnesium Salts .. ............. 154
Silica. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
Iron and Copper Oxides. . . . . . . . . . . . . . . . . . . . 156
Oil . .................................. 156
Identification of Scale Deposits . .............. 157
Internal Chemical Treatment. . . . . . . . . . . . . . . . 158
Phosphate Control. . . . . . . . . . . . . . . . . . . . . . 158
Carbonate Control. . . . . . . . . . . . . . . . . . . . . . 160
Organic Agents. . . . . . . . . . . . . . . . . . . . . . . . 160
Chelating Agents . ...................... 163
Chemical Testing. . . . . . . . . . . . . . . . . . . . . . . . . 164
Feeding of Internal Treatment Chemicals. . . . . . . 164
Feeding Methods for Clzelants. . . . . . . . . . . . . . . 165

HE formation of scale and sludge de- 22-1 illustrates the more common external
T posits on boiler heating surfaces is the
most serious water problem encountered in
treatment methods that are employed and
the range of residual hardness that usually
steam generation. The object of the majority results.
of the external treatment processes is to re-
move from the boiler feedwater those ob- MECHANISM OF SCALE FORMATION
jectionable substances which will contribute
to scale or deposit formations in the boiler. The primary cause of scale formation is the
Subsidence, coagulation and filtration may fact that the solubilities of the scale-forming
be employed to remove suspended solids from salts decrease with increase in temperature.
the boiler feedwater and so to prevent the Consequently, the higher the temperature
incorporation of mud, silt, etc., in boiler de- (and pressure) of boiler operation, the more
posits. Lime-soda softening, zeolite softening insoluble the incrusting salts become. No
and hot phosphate softening are used to re- method of external chemical treatment oper-
move from the boiler feedwater the incrust- ates at a temperature as high as that of the
ing salts of calcium and magnesium. Iron re- boiler water. Consequently, when the feed-
moval and silica removal processes have as water is elevated to boiler water temperature
their object the prevention of boiler scale and concentrations, the solubility of the
difficulties through the removal of iron and scale-forming salts is exceeded and they crys-
silica, both of which may .cause heat insulat- tallize from solution as scale on the boiler
ing deposits on the boiler surfaces. heating surfaces.
However, no system of external softening, The actual mechanism of scale deposition
regardless of a high degree of efficiency, is in can be viewed as taking place in two distinct
itself adequate protection against boiler scale operations. First, the precipitation from solu-
without the use of supplementary internal tion of a salt such as calcium carbonate
chemical treatment of the boiler water. In (CaC03) and the sedimentation of this al-
all cases, there will remain some small ready formed precipitate on the boiler heat-
amount of incrusting solids present in the ing surfaces where it bakes in place as scale.
boiler feedwater. In addition, it is also pos- Scale formation due to this mechanism un-
sible at time of softener regeneration, for doubtedly does take place particularly in fire-
appreciable hardness to enter the boiler feed- tube boilers and where the boiler feedwater
water system. Incomplete rinsing of a zeolite is high in hardness content. However, it can
bed also will introduce to the boiler feed- be stated that in general only a minor por-
water considerable quantities of scale-forming tion of scale formation develops in this man-
calcium and magnesium chlorides. Figure ner.
The chief mechanism of scale formation is
due to crystallization of scale-forming salts
from a solution locally supersaturated. The
thin viscous film of boiler water immediately
adjacent to the heating surface tends to be-
come more concentrated than the main body
of boiler water. Consequently, the solubility
of the scale-forming salts is first exceeded in
this thin film on the heating surface rather
than in the main body of the boiler water.
Crystallization of scale results directly on the
heating surface even when the solubility of
Figure 22-1 • Common External Processes the scale-forming salts has not been exceeded
qf Boiler Feedwater Preparation in the boiler water proper.

EFFECT OF BOILER SCALE graphically illustrated by Figure 22-2. For

simplification purposes no temperature drops
Boiler scale creates a problem in boiler oper- through gas or water films have been shown.
ation because the different types of scale Section A shows a cross-section of the tube
formed all possess a low degree of heat con- metal with a completely scale-free heating
ductivity. The presence of scale is therefore surface. There is a temperature drop across
equivalent to spreading a thin film of insula- the tube metal 'from the outside metal tem-
tion across the path of heat travel from the perature T 2 to the temperature of the boiler
furnace gases to the boiler water. The pres- water. Section B illustrates this same tube
ence of this heat insulating material will re- after the development of a heat insulating
tard heat transfer and cause a loss in boiler scale layer. In addition to the temperature
efficiency. Stack gas temperatures may in- drop from T 2 to T 1 , there would be an addi-
crease as the boiler absorbs less heat from the tional temperature drop through the scale
furnace gases. layer from T 1 to T 0 • This condition would,
There have been figures published show- of course, result in a lower boiler water tem-
ing large losses in boiler efficiency resulting perature T 0 • However, boiler water temper-
from even thin films of scale on the boiler ature is fixed by the operating pressure and
surfaces. Some of the early estimates of effi- operating conditions require that the same
ciency loss were erroneously high. The me- boiler water temperature T 1 be maintained
chanical structure of the scale, its porosity, and as before the development of the layer of
thickness as well as the design and method scale.
of operation of the boiler are all factors in-
fluencing the amount of heat lost. However,
it has been shown that the rate of heat trans-
mission may be reduced as much as 10-12%
by the presence of scale. While an equivalent TEMP~Itltrtllll

loss in heat utilization does not result, it is

possible for a scale of approximately Ys" T. BOIL~R
thickness to cause an over-all loss in boiler
efficiency in the neighborhood of 2.0-3.0%.
During the course of a year an increase of
2.0-3.0% in fuel consumption will amount to Sl!CTION A SEC'YION B ~I!CTIONC

an appreciable monetary loss. Figure 22-2 • Effect of Scale on

Even more important than the effect of Boiler Heating Surfaces
scale in causing heat loss, however, is the
effect of scale in causing overheating of the Section C illustrates the actual condition
boiler metal and consequent tube failures. that develops. Starting at the base boiler
Costly repairs and boiler outage are the re- water temperature of T 1 , the temperature
sult of such a condition. In modern boilers increase through the scale layer is repre-
with high rates of heat transfer, the presence sented by the line from T 1 to T 3 • The further
of even extremely thin deposits of scale will temperature increase through the tube wall
cause a serious elevation in the temperature is represented by the line from T 3 to T 4, the
of tube metal. The scale coating retards the outside metal temperature. It can be noted
flow of heat from the furnace gases into the that the outside metal temperature T 4 is now
boiler water. This heat resistance results in a considerably higher than the temperature T2
rapid rise in metal temperature to the point which was the outside metal temperature
at which failure results. prior to the formation of scale on the tube
The action which takes place in the blister- surfaces.
ing of a tube because of scale formation is If continued scale deposition takes place,

increasing the thickness of the heat insula- duction in calcium and magnesium content
ting deposits, further increases will take of the boiler feedwater. Internal chemical
place in the tube metal temperature until the treatment is used to preve~t deposit and
safe maximum temperature of the tube met- scale formation from the residual hardness
al is exceeded. Usually this maximum tem- concentrations remaining in the feedwater,
perature is taken as 900 F. and also to maintain clean boiler heating
With higher boiler pressures and attendant surfaces in those cases where no external
higher boiler water temperatures, there is softening is employed and where all chem-
less temperature difference between the boiler ical treatment for scale prevention must be
water temperature and the failing temper- applied directly to the boiler itself.
ature of the metal. Consequently, tube fail- Perhaps one of the most common sources
ures can result at higher boiler water pres- of scale is the breakdown of calcium bi-
sures with scale of considerably less thickness carbonate to form calcium carbonate under
than would cause failures at lower pressures. the influence of heat. This action is illus-
This condition explains the need for the ex- trated by the following equation:
tremely close control of boiler water condi- Ca(HCOlh + b.. = CaCOl +
tions in high pressure boiler operation. . calcium + heat = calcium +
bicarbonate carbonate
MUD, SILT, ETC. H20 + C02
water + carbon
The use of turbid river waters as makeup to dioxide
boiler systems is undesirable from several The precipitation from solution of calcium
standpoints, among which is the formation of carbonate to form boiler scale readily takes
boiler scale and sludge deposits. Settling of place where the boiler feedwater contains
mud at points of low flow may restrict boiler any appreciable quantity of calcium bicar-
water recirculation and cause water starva- bonate. The action illustrated above may
tion of water wall tubes. Baking of mud also be the cause for the formation of feed-
in place on tubes and sheets will result in line and economizer deposits and deposits on
their overheating and failure. Mud and silt open heater trays.
may be incorporated in scale deposits, add- Calcium sulfate is a more soluble salt than
ing to their volume and heat insulating calcium carbonate and is generally found
effect. only in boilers that have not received any in-
The removal of such suspended solids is ternal chemical treatment or have received
necessary prior to the use of the water for only an inadequate quantity of internal
boiler feedwater purposes. This removal is treatment. The presence of calcium sulfate
accomplished by the processes of subsidence, in a boiler scale can usually be taken as defi-
coagulation and filtration. No internal chem- nite evidence that internal chemical treat-
ical treatment of the boiler water is designed ment has been used only to an inadequate
to overcome difficulties due to suspended extent, possibly due to a flaw in the system
solids in the makeup water. The removal of chemical control. All methods of internal
of these impurities by external treatment chemical treatment rely on precipitating the
processes is required. calcium salts in some form other than cal-
cium sulfate.
CALCIUM AND MAGNESIUM SALTS Figure 22-3 illustrates the solubility of cal-
cium sulfate together with similar data for
The salts of calcium and magnesium are the calcium carbonate. The curves illustrated
most common source of difficulty with boiler show the solubility of the pure materials in
scale. The use of external treatment proc- distilled water solution. The presence of
esses such as lime-soda, zeolite and hot phos- other salts in the solution will affect the
phate softening have as their object the re- solubilities shown and consequently it is

40 salts commonly employed for this purpose.

Magnesium salts present in the boiler feed-
32 1\ water are usually more easily prevented from
\ forming scale than are the calcium salts.
;i ~ Normally, the magnesium salts found in-
.. -

16 ........... CALCIUM SULPHATE

corporated in boiler scale are magnesium hy-
droxide, magnesium silicate or magnesium
"'.... N phosphate. The latter salt, magnesium phos-
..."'< 8
.! r- 1--- phate, tends to form a very sticky deposit,
not very hard in nature, but requiring tur-
0 bining for its removal.
360 400 440 480 520 560 600
Figure 22-3 • Calczum Solubilit)', for Calcium
Su(fate and Calcium Carbonate As mentioned previously, calcium silicate de-
posits may prove quite troublesome in boiler
necessary to employ certain factors of safety operation. However, another source of con-
in interpreting these data. For example, the cern is the tendency for silica to form scales
usual boiler water will contain a variety of of a complex nature, which may or may not
salts in solution, tending to alter the theo- include calcium and magnesium in their
retical solubilities illustrated. Specifically, in structure. Particularly in combination with
the case of calcium sulfate, an increase in the aluminum, silica tends to form complex
quantity of sodium sulfate present will lower scales which have been identified by X-ray
the calcium sulfate solubility. diffraction methods. Formation of such com-
Because of the high solubility of calcium plex deposits usually occurs in boilers oper-
sulfate as shown in Figure 22-3 it is rela- ating at the higher pressures and proper
tively simple to introduce chemical treat- treatment of the boiler water requires the
ment to the boiler water and, with proper maintenance of both silica and aluminum
control, to avoid deposition of calcium sul- concentrations at a minimum. Under such
fate scale. operating conditions, the prevention of com-
The use of sodium carbonate (soda ash) plex silica scales is primarily a problem for
was one of the first materials employed to external treatment. The maintenance of high
prevent calcium sulfate scale formation. As phosphate and alkalinity concentrations in
can be seen from Figure 22-3 calcium car- the boiler water is an aid in control of the
bonate is considerably more insoluble than problem, but only limited silica concentra-
calcium sulfate and accordingly will be pref- tions can be controlled. Blowdown can be
erentially precipitated. However, calcium used to reduce silica concentrations to toler-
carbonate in itself can form a hard adherent able levels.
deposit and is frequently found as a constit- The concern for silica is not always be-
uent of boiler scale. cause silica concentrations are too high. In
Calcium phosphate and calcium silicate some cases, silica can be too low. This situ-
are frequently deposited as boiler scale. Cal- ation occurs primarily with waters that con-
cium silicate may be formed either from the tain high magnesium concentrations and low
combination of the calcium and silicate ions silica values. In these special cases, magne-
naturally present in the boiler feedwater or sium will tend to precipitate as magnesium
as a result of the misguided use of sodium phosphate which produces a sticky sludge.
silicate as internal chemical treatment. Cal- Precipitation in this form is aggravated when
cium phosphate, on the other hand, is in- the feedwater enters the boiler in a region of
variably the result of treatment of the boiler low alkalinity and poor circulation. Under
water by the use of the various phosphate these conditions, increasing the silica content

of the feedwater to avoid magnesium phos- line corrosion has eliminated this problem in
phate precipitation will prove helpful. How- several instances.
ever, this practice must be applied within Copper and copper bearing. metals consti-
the normal silica tolerances established for tute a large proportion of the metallic sur-
the boiler operating pressure. Other factors faces exposed in a modern power station.
that must be considered are increased boiler Solution of copper from surface condensers,
water alkalinity, the use of special organic pump impellers, stage heaters and evapo-
sludge conditioning agents and better mixing rators takes place as a result of corrosive
of the feedwater within the boiler. action. Dissolved oxygen, ammonia and car-
bon dioxide exert the major influences in the
IRON AND COPPER OXIDES corrosion of copper. Ammonia forms a sol-
Deposits of iron oxide, metallic copper and uble complex cupric ammonium ion which
copper oxide are frequently found in boilers is broken down on heating and can result in
operating with very pure feedwater. In such deposition of copper oxide in the boiler.
plants, the makeup water requirements may Where conditions are favorable to the depo-
be only 1-2% with evaporators or demineral- sition of metallic copper on the boiler metal,
izers supplying the makeup water needs. De- it is possible for localized pitting to take place
posits in drums and tubes may consist essen- beneath the deposit.
tially of iron oxide, copper oxide and metallic While internal corrective measures such as
copper. The usual sludge forming calcium the use of organic sludge conditioning agents
and magnesium salts may amount to less than may be applied to minimize the problem of
5% of the boiler deposit. iron oxide and copper oxide deposits, the
basic solution to the problem lies in cor-
The usual source of such deposits lies in
recting the corrosive condition.
corrosion external to the boiler. Corrosion of
iron and steel can result in the solution of In some cases, the source of iron oxide
iron by the condensate or feedwater, with its boiler deposits is not the result of corrosion
subsequent precipitation under the higher external to the boiler. Corrosive attack on the
temperature and alkalinity conditions of the boiler metal by high caustic concentrations
boiler water. The usual causes of such cor- or dissolved oxygen may be the cause of this
rosive action are dissolved oxygen and carbon condition.
dioxide. The prevention of the iron oxide
deposit in the boilers requires elimination of
the corrosive effect of these gases. The source The presence of oil in boiler feedwater is
of dissolved oxygen should be determined particularly undesirable. One of the most
and removed through the use of mechanical obvious difficulties resulting from oil con-
andjor chemical deaeration. If evaporators tamination is carryover of solids with the
are employed, carbon dioxide should be re- steam. However, oil may also contribute
duced to a minimum by suitable pretreatment greatly to the formation of heat insulating
or by changing the point to which evaporator scale deposits and is frequently the cause of
distillate is discharged. The use of ammonia, blistered tubes. Oil may be carbonized to
neutralizing amines and filming amines aid form a hard asphaltic scale or may be ab-
in the control of this problem. sorbed by porous scale already present. Ad-
Iron dissolved in the condensate and thus herence of such deposits to the heat transfer
introduced into the boiler can cause addi- surface may cause local overheating and
tional difficulties besides the deposition of failure of the metal.
iron oxide. Combination of iron with silica Oil should always be externally removed
present in the boiler water can result in the from the boiler feedwater. There is no
formation of iron silicate boiler scales. The method of internal chemical treatment spe-
use of filming amines to control condensate cifically adaptable to overcoming the hazards

presented by oil contamination of the boiler Fig. 22-4

water. While certain methods of internal Crystalline Scale Constituents, Identified
treatment may slightly aggravate or slightly by X-ray Diffraction
minimize the problem caused by oil, there is
no method of internal treatment which can
be relied upon to cope with any appreciable Acmite NazO.Fez03.4SiOz
oil contamination. Since the extent of the Analcite NazO.AizOa.4Si0z.2HzO
Anhydrite CaS04
difficulties caused by oil are more or less un-
Aragonite CaC03
predictable, external removal of oil from the Brucite Mg(OH) 2
boiler feedwater is the only safe policy. Calcite CaC03
Concrinite 4NazO.Ca0.4Aiz03.2COz.9SiOz.3H20
Hematite Fez03
Hydroxyapatite Calo(OH)z(P04)6
Magnetite Fe304
In addition to the microscope and gravi- Nose lite 4Naz0.3Aiz03.6Si0z.S04
metric chemical analyses, a useful tool in Pectolite Naz0.4Ca0.6Si0z.H20
identifying boiler scale deposits is the X-ray Quartz Si0 2
diffraction method. Any material possessing Serpentine 3Mg0.2Si0z.2Hz0
a crystalline structure will yield a distinctive Thenardite NazS04
Wollastonite CaSi03
X-ray pattern which will always be the same Xonotlite 5Ca0.5Si0z.Hz0
and which can be compared with the pattern
of standard materials on file. It is thus pos- ture whereas only the hypothetical structure
sible to identify complex chemical com- could be obtained from the usual chemical
pounds and their individual crystalline struc- analysis.

Figure 22-5 • Analysis of Boiler Deposits by X-ray Fluorescence


Figure 22-4 illustrates some of the com- whether or not the precipitate will tend to
mon and uncommon constituents of boiler tightly adhere to the boiler heating surfaces
scale which have been identified by X-ray -the chemical characteristics of the precipi-
methods. Some of the salts could readily be tate are relatively unimportant. Thus, from
determined by the conventional chemical a chemical standpoint sodium silicate will
analysis. Others, such as the complex precipitate the calcium and magnesium salts
alumino-silicates, could hardly have been just as well as a phosphate or carbonate.
identified except with the use of X-ray However, it is the physical characteristics of
diffraction methods. The use of X-ray fluo- the precipitate that are of importance, for a
resence also is developing. It is possible that precipitate of calcium silicate will tightly ad-
X-ray fluoresence eventually will prove more here to the heating surfaces.
useful in the identification of boiler deposits
than X-ray diffraction. Today it is no longer necessary to depend
on industrial application for experimentation.
This work is now performed in laboratory
experimental boilers and carefully studied in
For the internal chemical treatment of boiler each stage of development before application
water many strange and varied materials to the industrial field. One type of laboratory
have been employed ranging from the earlier experimental boiler used in this work em-
use of potato peelings, kerosene, molasses, etc., ploys ring circulation with electrical resistance
as "cure-ails" to the present use of specific heating, using specially wound coils. Elec-
chemicals applied to specific problems. It tronic water level control maintains water
can be accepted as axiomatic that no one level at +0.25 inches. The entire operation
particular material is universally applicable is automatic with automatic feedwater and
and that a "cure-all" is a physical impos- treatment pump control and automatic con-
sibility. Present day internal chemical treat- tinuous blowdown based upon boiler water
ment uses commercial chemicals with known conductivity. Electrical input is measured for
chemical properties to accomplish certain correlation with steam production. Upon
specific objectives. Many of the chemicals conclusion of a test run, the scaling surface
commonly employed are available to all on is removed from the boiler, the scale dis-
the open market; others are proprietary in solved in inhibited acid and analyses then
nature or their use may be subject to patent conducted on the acid solution for the vari-
licensing. Continued research in the field of ous scale constituents. Opportunity is pro-
boiler water conditioning is constantly de- vided for study of crystal structure and ob-
veloping new materials of value and finding servation of scale adherence prior to re-
new uses for well known products. moving scale from the heating surface. Boiler
The reason for the earlier use of "boiler water samples secured during operation are
compounds" with their usually disastrous re- analyzed for correlation with the scale analy-
sults is that the present day highly technical sis data.
science of internal boiler water treatment As previously stated, scale is caused pri-
developed slowly and to a large extent by marily by calcium and magnesium salts,
"trial and error" methods. Too often, the silica and oil. Each is a separate problem
"trial and error" method was applied to and must be dealt with individually.
industrial boiler plants without preliminary PHosPHATE CoNTROL. For the prevention of
laboratory and pilot plant experimentation. calcium and magnesium salts from baking on
It is well to realize that the prevention of boiler evaporative surfaces, the internal treat-
boiler scale cannot be predicted by any basic ment consists of precipitating the calcium
chemical principle. Where a precipitating and magnesium salts as sludge, and main-
type of treatment is employed, it is the taining this sludge in a fluid form to be re-
physical characteristics of the precipitate moved by the boiler blowdown. Calcium is
formed in the boiler water that determines considered to present a more difficult prob-

Figure 22-6 • Research Experimental Boilers

lem than magnesium. This is because the phate to be used depends upon economics
magnesium is readily precipitated by the alka- and upon the alkalinity of the boiler water.
linity of the boiler water to form a sludge of In some instances it is desirable to blend
magnesium hydroxide. two phosphates to obtain the desired alka-
The most common chemicals employed for linity. In others, it may be necessary to add
the precipitation of the soluble calcium salts alkaline materials to the boiler water such
are the sodium phosphates-trisodium phos- as caustic soda or soda ash.
phate, disodium phosphate, sodium meta- Once the calcium and magnesium salts
phosphate and monosodium phosphate. Once have been precipitated as a sludge, it is
any of these phosphates enter the boiler, necessary to maintain this sludge in a fluid
their action is exactly the same. They are all form. Generally, this is accomplished by the
converted into trisodium phosphate and as use of one or more forms of organic material.
such react with the calcium to form tri- After a fluid sludge of calcium and magne-
calcium phosphate, a rather flocculent pre- sium precipitates has been formed, it is
cipitate. In order for this reaction to take necessary to remove this sludge from the
place, sufficient alkalinity must be present in boilers by blowdown. The amount of blow-
the boiler water. At a pH value of approxi- down required depends upon the amount of
mately 9.5 or less the calcium is not pre- sludge formed and the nature of the sludge.
cipitated properly. The choice of the phos- Fluid sludges require less blowdown.

CARBONATE CoNTROL. Another type of in- adherent of the inorganic precipitates. How-
ternal treatment for the prevention of scale ever, even these precipitates will form scale
is commonly known as carbonate-organic and sludge on boiler heating surfaces, de-
treatment. This type treatment has definite pendent on many factors such as suspended
uses and limitations. Its principal application solids concentration, boiler pressure, heat
is on the so-called self-purging waters- transfer rate and chemical balance of the
waters in which the alkalinity is approxi- boiler water.
mately equal to or greater than the hard- Organic agents are required to prevent
ness. This type of treatment employs rela- scale and adherent sludge deposits resulting
tively high concentrations of active organic from these inorganic precipitates. It is proper
materials in the boiler water together with internal conditioning to use inorganic agents
the use of sodium carbonate (soda ash) to to precipitate the incrusting solids in the
maintain proper alkalinity concentrations. In most desirable inorganic form. Then the in-
cases of low feedwater hardness, this type organic treatment may be supplemented with
of treatment is generally not as applicable as organic agents to complete the formation of
~ phosphate type of treatment. However, in
a fluid sludge and to prevent the inorganic
a feedwater of high hardness, that is over 60 precipitates from adhering to the boiler heat-
to 70 ppm, and with high alkalinity con- ing surfaces. The organic agents commonly
centrations, carbonate-organic treatment can used in internal treatment of boiler water are
produce better results than phosphate treat- tannins, lignins, starches, and seaweed deriva-
ment, primarily through the elimination of tives.
the heavy sludge condition which usually ac-
companies phosphate treatment of high- Tannins, which are extracted from wood
hardness feedwaters. In some instances a and bark, are organic substances of high
combination of both types of treatment has molecular weight and complex structure. It is
produced better results than either one singly. characteristic of these compounds that they
are benzene derivatives and also are optically
ORGANIC AGENTS. The use of organic agents active. Tannins possess a glucoside structure
for the prevention of scale and sludge de- in that sugar is produced on fermentation or
posits is a necessary part of the internal hydrolysis. While synthetic compounds of
boiler water conditioning for most industrial tannin structure have been prepared, the
plants. Where external feedwater condi- tannins usually employed in boiler feedwater
tioning systems result in practically com- conditioning are of natural origin. Tannins
plete removal of incrusting solids and where may be conveniently grouped into catechol
very close chemical control is exercised, there and pyrogallol types and mixtures of the two.
may be no need for organic treatment. Central Among the catechol tannins are cutch, que-
station operation is an example of where or- bracho, hemlock, larch, and gambier. Pyro-
ganics are required only for special problems. gallo tannins include gallnuts, sumac, myro-
However, in a great majority of industrial balans, chestnut, valonia and dividivi. Mixed
plants, particularly where feedwater hard- tannins include wattle bark, English oak and
ness is high, organics are a definite necessity chestnut oak bark. These different tannins
if optimum conditions are to be maintained. are some of the most common, but many
As mentioned previously, a precipitating others also exist. Commercial products in
type of internal boiler water treatment is de- some cases contain mixtures of different tan-
signed to precipitate the calcium salts as tri- nins.
calcium phosphate and the magnesium salts Lignin is also a complex organic substance
as magnesium hydroxide. Under some cir- and is the chief non-cellulose component of
cumstances it has been found desirable to wood. The small portion of the total lignin
precipitate a portion of the calcium as cal- present which is in a free state and which
cium carbonate and some of the magnesium can be isolated with the use of solvents is
as magnesium silicate. These are the least referred to as "native lignin." Native lignin

contains four methoxyl groups, four hydroxyl some extent a coating action on precipitated
groups and one carbonyl group for each unit inorganic solids, decreasing their tendency
of molecular weight of 840. Commercial both to cohere and adhere to boiler surfaces.
methods of lignin extraction cause some Of prime importance, however, is the effect
chemical modification in structure. One of of organic agents in distorting crystal forma-
the most useful forms in which lignin is com- tion and thus inhibiting scale deposits directly
mercially available is as the lignosulfonate, at the point of formation. Regardless of the
produced by the sulfite digestion process. theory chosen to explain the action of or-
Starch is a high polymer carbohydrate. ganics, the end result with proper organic
The structure is presumed to consist of glu- concentrations in the boiler water is the
cose, joined by glucoside linkages. There is formation of a fluid boiler water sludge,
considerable range in the size of the mole- readily removed by blowdown. With proper
cules even in one variety of starch and the organic treatment of a boiler water there is
chemistry of starch is quite complex. In the also considerably less sludge of any sort re-
natural state, starch is insoluble in water, maining in the boiler upon opening.
but on boiling the individual cell walls burst Figure 22-7 shows the low and high heat
and a portion is dissolved, forming a milky transfer scaling surfaces (left-28,000 Btujsq
solution of starch paste. Commercially, starch· ftjhr and right-180,000 Btufsq ftjhr) from
is prepared from corn, wheat, potatoes and an experimental boiler run without the use
rice. The U. S. Navy standard boiler treat- of an organic sludge conditioning agent. The
ment formulation specifies the use of corn- deposit on the low heat transfer scaler was
starch for sludge conditioning and anti-foam 1.20 gramsjsq ft and on the high heat trans-
action. fer scaler 5.80 grams/sq ft.
Of the seaweed derivatives, the most prom- In contrast, Figure 22-8 shows the effect
inent are dispersions of sodium mannuronate, of an organic sludge conditioner under other-
Irish Moss, sodium alginate and agar-agar. wise identical test conditions. The low heat
'There are over 1100 kinds of seaweed from transfer scaler developed a deposit of 0.03
which a variety of water soluble gums can grams/sq ft and the high heat transfer
be extracted. Laminaria digitata, laminaria scaler a deposit of 0.12 gramsjsq ft.
longicruris and laminaria saccharina are Figure 22-9 illustrates data typical of the
varieties of seaweed widely used for extrac- results secured in experimental boiler studies
tion purposes. The molecular weight of sea- by superimposing various types of organic
weed derivatives employed in boiler feed- agents on phosphate control: While different
water conditioning varies with the degree of feedwaters were employed in the three series
polymerization and has been measured up to of tests covered by this figure, in each case
250,000. Sodium mannuronate dispersions the control run employed conventional phos-
will precipitate with polyvalent positive ions phate internal boiler water treatment. By re-
such as calcium and magnesium, forming a peating the test with all conditions held
flocculent precipitate which will adsorb sus- constant with the exception of the additional
pended inorganic precipitates. organic treatment of the boiler water, data
There are probably several different mech- were secured on the scale reducing properties
anisms by which organic substances function of the organic agents under the test condi-
to hinder boiler scale form~tion. With the tions employed.
sodium mannuronate polymers, a coagulant Figure 22-10 illustrates similar studies on
effect is secured with adsorption of inorganic the reduction of calcium carbonate type scale
precipitates in the organic floc. With tannins where no phosphate treatment was employed
and lignins it is believed that a major part of and the carbonate and hydroxide ions were
their action is in dispersion of precipitates. the only inorganic precipitating agents used.
All of the organic agents probably exert to Imposing organic agents under such condi-

Figure 22-7 • Scalers without Organic Treatment Figure 22-8 • Scalers with Organic Treatment

Fig. 22-9
Phosphate Scale Control Studies
Organic Cone. Type Suspended Solids Heat Transfer Operating Scale Formed %Scale
in Boiler Water Organic in Boiler Water Rate Pressure grams/ft 2 Protection
ppm ppm Btu/ft 2/hr psi

0 400 57,000 260 1.89

190 lignin 420 57,000 260 1.06 44.0%
180 Lignin 380 57,000 260 1.66 12.2%
180 Lignin 400 57,000 260 0.20 89.5%
0 250 24,000 320 0.684
150 Tannin 300 24,000 320 0.526 23.1%
150 Tannin 250 24,000 320 0.453 33.7%
150 Tannin 300 24,000 320 0.092 86.6%
0 900 57,000 650 3.50
180 Seaweed Derivative 800 57,000 650 1.98 43.5%
190 Seaweed Derivative 750 57,000 650 1.19 66.1%
180 Seaweed Derivative 800 57,000 650 0.55 84.2%

Fig. 22-10
Carbonate Scale Control Studies
Organic Cone. Type Suspended Solids Scale Formed
in Boiler Water in Boiler Water %Scale
ppm Organic ppm grams/ft2 Protection

0 400 4.10
200 Tannin 380 2.95 28.0%
200 Tannin 400 2.13 48.0%
185 Tannin 380 1.39 66.1%

Operating Pressure: 300 psi, Heal Transfer Rate: 57,000 Btu/f12/hr.


tions, which are similar to many types of Proper application of qualified chelating
railroad operation, can result in a marked agents involves optimum levels of suitable
decrease in boiler scale and sludge. While it supplementary dispersing agents and the
can be demonstrated by such experimental maintenance of favorable alkalinity as well
work that a major reduction of scale and as pH conditions. Where employed in a com-
sludge deposits is obtained by superimposing bination program, phosphate must be satis-
the proper type of organic agent on inorganic factorily regulated. With these factors in
control, the information secured on the balance, existing precipitated matter is re-
grams of scale per square foot of heating r110ved in a controlled manner and with a
surface is not the sole criterion for judging controlled degree of rapidity while prevent-
the applicability of a specific organic for in- ing new accumulations. Even certain complex
ternal boiler water treatment. Some organic scales, such as analcite and acmite, can be
agents have shown a fair degree of scale re- prevented or gradually removed with the ap-
duction, but have resulted in the develop- plication of a properly balanced chelant
ment of a thin, tightly adherent scale, high in program.
organic, and very difficult to remove from The applicability of the chelant type of
the heating surface. It is necessary to corre- treatments is generally in the area of low
late with the scale formation, information on feedwater hardnesses. That is, the treatment
the crystal structure of the deposit, its tend- is economical for most systems where the
encies for adherence and a complete chemi- feedwater hardness is 1.0 ppm or less. A full
cal analysis of the scale itself for proper chelant program is seldom economical when
interpretation of the tests. hardness in the feedwater exceeds 2.0 ppm.
CHELATING AGENTS Recent developments indicate that, under
certain conditions of application, chelating
Chelating or complexing agents have been agents provide dispersive powers of a nature
applied for a number of years to remove de- which are additive to the properties of nor-
posits of calcium and magnesium salts from mal organic dispersants. These characteristics
waterside surfaces of boilers. Developments can advantageously supplement standard
over the past few years have shown that these phosphate type treatments, particularly in the
agents can also prevent certain types of scale cases of systems where the feedwater hard-
and sludge accumulations in industrial boilers. ness is at a higher level. This property ex-
The completely new approach to the rather tends the useful range or the applicability of
common problem of preventing objectionable properly selected chelants.
deposits employs the tetrasodium salt of Chelating agents are useful for not only
ethylenediamine tetraacetic acid and the tri- problems of complex scale and sludge on
sodium salt of nitrilo triacetic acid as the waterside surfaces of industrial boilers, but
basic and most common chelating agents. A also for the control of iron and copper de-
number of complexing and sequestering posits in high purity feedwater systems. How-
agents are used in conjunction with these ever, chelating agents are not a cure-all for
basic agents in order to provide a more com- all types of boiler water problems. As an
plete and satisfactorily controlled scale and example, carryover problems are not cor-
sludge prevention program. Application is rected to any measurable degree in the aver-
in a continuous fashion on a long-term basis. age plant when a conversion to chelant
By their chemical nature, chelating agents treatment is made. In practically all instances,
form complex soluble salts with calcium, any carryover condition existing under a
magnesium, iron and copper as well as other phosphate treatment continues unabated with
metallic ions in exchange for sodium. There- a chelant approach. When mechanical or
fore, in the practical application, precipitates chemical characteristics quite apart from the
are avoided by the formation of soluble salts. small amount of suspended solids present

cause carryover, the use of chelants does not stances where contamination has been severe,
solve the problem. Field studies, based upon superheater problems due to carbonized oil
the sodium tracer technique, have shown also deposits have continued. It is well to note
that the application of chelating agents does that oil contamination is an undesirable situa-
not contribute to a carryover condition. tion with respect to boiler feedwater charac-
Use of a proper chelant program on a teristics. The best possible solution to an oil
continuous basis provides certain other ad- problem is to eliminate the source of con-
vantages which usually justify a higher cost. tamination or to remove the oil completely
Because of the reduction in suspended solids from the feedwater before introduction to the
and the elimination of precipitates, higher boiler.
total solids in the boiler water and hence CHEMICAL TESTING
lower blowdown rates are possible in many
instances. A further advantage of chelating The use of internal chemical treatment, re-
treatments is the fact that temporary upsets, gardless of the excellence of the materials
which may cause a scaling condition, can be employed, is valueless unless it is closely con-
mitigated with respect to aftereffects. That is, trolled by means of daily chemical tests of
the maintenance of proper treatment balances the boiler water by the plant personnel. De-
following the upset period causes the removal pendent upon the nature of the treatment
of scale. employed and the individual plant condi-
Longer and safer operation between turn- tions, the control of chemical treatment and
arounds is possible with the advance of che- blowdown may be based on tests for hard-
lant treatments. Some states allow longer ness, alkalinity, chloride, sulfite, phosphate,
operating periods between internal inspec- chelant, silica, nitrate, specific conductance
tions when waterside conditions are satis- and total and dissolved solids. Not all of these
factorily controlled, and such extension may tests are necessary in any one plant, but de-
be an incentive to apply a suitable chelant pendent upon the specific problems encoun-
program. Also, since less work is required for tered, several of these tests may be required
cleaning waterside surfaces with an effective for proper control. In addition, special tests
chelant program, less time is needed for a are occasionally necessary.
turnaround. Labor requirements are reduced. FEEDING OF INTERNAL TREATMENT
Of course, close control must be maintained CHEMICALS
in order to fully gain the advantages of a
properly designed chelant program. In order A satisfactory method of feeding chemical
to aid control, the development of accurate correctives is second only in importance to
and adequate testing methods is a continuing the consistent control of chemical balances.
effort in this new field of boiler water treat- However, some feedwater problems dictate
ment. the necessity of highly efficient feeding meth-
As indicated previously, chelants are not ods before adequate chemical control can be
cure-all treatments. The application of a established and maintained.
chelating program does not correct corrosion Obviously for proper results and to insure
problems due to oxygen and/or carbon di- maximum efficiency from any chemical treat-
oxide. Failure due to steam blanketing is not ment, continuous methods should generally
eliminated, and the potential for embrittle- be employed. Many chemicals, such as so-
ment must still be considered. As discussed in dium sulfite and the majority of the organics,
other chapters of this publication, treatments may be fed continuously to the suction of the
of a specific nature are required. boiler feed pump without fear of inviting
In some instances, the reduction of sus- feed line deposits, regulator troubles, etc.
pended solids has greatly reduced the effect On the other hand, phosphate treatment
of oil contamination. However, in isolated in- and in certain instances sodium carbonate

and sodium hydroxide, when applied in a

manner similar to the above, cause a prema-
ture reaction to take place in the boiler feed
line. Precipitates such as tricalcium phosphate
and calcium carbonate which result, readily
interlock on water distribution and regula-
tor surfaces to form adherent scale. These
premature reactions may frequently be ac-
celerated by high feedwater temperatures or
by chemical characteristics such as high feed- Figure 22-11 • Section of Boiler Feedline Containing
water hardness, excessive feedwater alkalin- Phosphate Deposit, Resulting from Continuous Feeding of
ity and high pH values. Phosphate to Cold Process Lime-Soda Sqftened Feedwater
Serious difficulty can be anticipated from
the consistent formation of feed line deposits. ment purposes is by pumping the phosphate
Usually, erratic operation of boiler feedwater solution direct to a boiler drum, utilizing for
regulators is the first symptom, followed by this purpose a small chemical pump. This
thermal loss in economizers or the necessity method of chemical feed is also suitable for
for increased feedwater pressure. Eventually the introduction of most other treatment
a point may be reached where insufficient materials. Sodium sulfite is usually pumped
feedwater can reach the boilers regardless continuously to the suction side of the boiler
of feed pump pressure. Sudden forced shut feed pump or to the storage section of the
down and boiler outage results, with conse- feedwater heater. Organic sludge condition-
quent possible equipment damage or certain ing agents may be fed continuously to the
loss in production schedules. boiler feedwater or direct to a boiler drum.
Chemicals added for alkalinity adjustment,
In the case of phosphates, the tendency
such as caustic soda and soda ash, may in
toward premature precipitation is not con-
some cases be fed continuously to the boiler
fined to the orthophosphates such as triso-
feedwater and in other cases, dependent on
dium, disodium and monosodium phosphate.
feedwater characteristics, must be fed directly
The problem of feed line deposition is also
to the boiler drum to avoid feedline deposits.
prevalent with the use of molecularly dehy-
drated phosphates.
In their original or unreverted states, the
molecularly dehydrated phosphates will not The basic chelating agents commonly used
precipitate as readily with calcium as the in boiler water treatment have an affinity for
ortho materials. This property is frequently boiler system metals under certain conditions.
utilized, in conjunction with certain surface Copper bearing materials, in particular, are
active organics, to prevent calcium carbonate subject to attack. Therefore, contact with
deposition. However, it should be remem- copper bearing alloys in feedwater stage
bered that such formulations (normally used heaters, control valves and check valves must
in relatively small quantities) are designed to be avoided. Copper cooling coils for sampling
stabilize the natural deposition tendencies of should be changed to a suitable stainless
a water. They are not expected to cope with steel, such as 304 or 316. Monel and similar
the problem of feed line deposition resulting nickel alloys are not satisfactory for this
from the continuous feed of relatively high service.
concentrations of chemical precipitants, such Suitably diluted chelant solution is usually
as sodium carbonate, sodium hydroxide and injected into the boiler feedwater through a
all phosphates. properly designed stainless steel quill. This
The ideal method of adding phosphate to quill is best located on the pressure side of
a boiler water for internal chemical treat- the feedwater pump, and preferably at a

point where the temperature is no greater REFERENCES

than the normal feedwater temperature. R. C. Corey, "What About Copper in Boilers,"
Because of their great activity, chelating Combustion, Vol. 16, pp. 43-46. (June 1945)
agents should be fed to a boiler system only J. A. Holmes, "The Development of Organics for
Water Treatment," Proceedings, Midwest Power
in a continuous fashion. This method pro- Conf., Vol. XIII, pp. 238-245 (1951)
vides continuous dilution. In addition, direct J. J. Maguire, "Organic Agents in Boiler Scale and
feed to boiler drums should be avoided, even Sludge Control," Proceedings, Midwest Power
though a suitably designed stainless steel Conf., Vol. XIII, pp. 246-256 (1951)
distribution line may be installed within the J. Van Brunt, "Tube Failures in Water-Tube Boil-
ers," Combustion, Vol. 17, pp. 32-37 (July 1945)
drum. The breaking of the line by chance
could lead to serious and undetected cor- J. R. Metcalf, "Chelant Treatment Programs for
Boiler Water," The Betz Indicator, Betz Laborato-
rosion problems. ries, Inc., Trevose, Pa. (June, 1967)

Boiler Corrosion Control

Control of pH . .......................... 168

Control of Dissolved Oxygen . .......... , . . . . 168
Protection of Idle Boilers. . . . . . . . . . . . . . . . . . . 171
Dry Storage. . . . . . . . . . . . . . . . . . . . . . . . . . . 171
Wet Method. . . . . . . . . . . . . . . . . . . . . . . . . . 172
Superheater Storage . ...... ; . . . . . . . . . . . . . 172

major consideration in the design of a

A feedwater conditioning system is pro-
tection of boilers and accessory equipment

Dissolved oxygen can be int~oduced into the

against corrosion, both during operation and system not only in the makeup water but also
when the equipment is out of service. Except due to air infiltration of the condensate sys-
in special cases, corrosion in boilers, feed tem. When dissolved oxygen is present in the
lines and economizers is due either to the feedwater, an attack of the feed line, closed
influence of low pH or dissolved oxygen. heaters and economizer can be expected with
the severity of the problem dependent on the
concentration of dissolved oxygen and the
CONTROL OF pH temperature involved. One of the most seri-
ous aspects of oxygen corrosion is that it
The solubility of iron in water increases as generally occurs as pitting so that the attack
the pH decreases below the neutral point. is con-centrated in a small area of the total
The attack on steel by water with an acid metal surface. With this type of corrosion,
pH is of a general nature over the entire failures can occur even though only a rela-
surface with little tendency toward localized tively small portion of the metal has been
corrosion. The obvious method of preventing lost.
such corrosion is neutralization of the acid The influence of temperature on the cor-
with an alkali. Soda ash and caustic soda are rosivity of dissolved oxygen is particularly
commonly employed for this purpose in the important in such equipment as closed heat-
treatment of boiler feedwater. Except in ers and economizers where the water temper-
special cases of corrosion in high pressure ature is increased very rapidly. Under such
boilers, experience has shown that it is de- conditions an additional driving force for
sirable to maintain a minimum boiler water ' reaction is present and for th1s.
the oxidation
pH of approximately 10.5. This pH is suf- reason, even very small quantities of dis--
ficiently high to stifle acidic attack. In the solved oxygen in feedwater can cause severe
proper treatment of boiler water for control corrosion in such equipment.
of scale and sludge deposits, the desired
When oxygen is present in the feedwater
boiler water alkalinity and pH for prevention
entering the boiler, a portion will be flashed
of acidic corrosion is automatically estab-
and will leave the boiler with the steam. The
lished. Therefore, except under unusual op-
erating conditions, pH is not a factor in remainder of the dissolved oxygen can attack
boiler water control. the boiler metal. While the point of attack
will vary with the boiler design and feed-
Trade waste contamination of the makeup water distribution, oxygen pitting is usually
water, sugar contamination of the condensate concentrated adjacent to the water level in
or acidic contamination from acid pickling the feedwater drum.
or plating baths are some of the most com-
The first and most important step in elimi-
mon sources of acid contamination of boiler
feedwater. Acid contamination can also enter nating the corrosive influence of dissol~ed
the system due to improper operation or oxygen is mechanical deaeration of the bmler
control of hydrogen zeolite softeners. feedwater. Efficient deaeration will reduce
the dissolved oxygen content of the boiler
If the makeup water or condensate is sub- feedwater to a very low value. It is advisable
ject to periodic acidic contamination, record- to follow mechanical deaeration by chemical
ing or indicating instruments are desirable deaeration in order to remove the last traces
to maintain a check on the purity of the of dissolved oxygen. Where mechanical de-
makeup or condensate. Such devices can be aeration is not employed, chemical deaera-
used either to operate automatic dump valves tion must be used for the removal of the en-
or to warn the operators to discharge the tire oxygen content of the feedwater.
contaminated water to waste.
Sodium sulfite is the chemical agent most

of the boiler feed pump. While intermittent

application is generally not recommended, it
has been found in some low pressure systems
that adequate protection is provided as long
as the additions of sodium sulfite are made
with sufficient frequency to continuously
maintain the proper residual concentration in
the boiler water.
Testing of the boiler water for sulfite resi-
dual and recording the quantity of sulfite re-
quired serves also as a quick check on heater
deaeration efficiency in those plants where
the oxygen content of the feedwater is not
Figure 23-1 • Solubility of Dissolved Oxygen determined regularly. Any decrease in boiler
water sulfite residual and consequent need
for increased feed of sodium sulfite, is an in-
commonly employed for chemical deaeration dication that heater operation should be
due to its low cost, ease of handling and its checked to ascertain and correct the reason
lack of scale forming properties. The oxygen for increased oxygen content of the boiler
scavenging characteristic of sodium sulfite is feedwater.
illustrated by the following reaction:
The speed of the sulfite-oxygen reaction is
2NazSOl + Oz. = 2NazS04 affected by a number of factors, the most
sodium sulfite + oxygen = sodium svlfate important being temperature. The reaction
The removal of 1.0 ppm dissolved oxygen time decreases with increased temperature.
theoretically requires 7.88 ppm of chemically In general, the reaction speed doubles for
pure sodium sulfite. However, use of a every 10 C increase in temperature. At tem-
technical grade of sodium sulfite combined peratures of 212 F and above the reaction is
with handling and blowdown losses as en- quite rapid. It has also been found that the
countered in actual plant operation usually presence of an excess or overfeed of sodium
requires the feed of approximately 10 lbs of sulfite will increase the reaction rate. Several
sodium sulfite for each pound of oxygen. investigators have shown that the reaction
Requirements will also depend on the con- proceeds most rapidly at pH values in the
centration of excess sulfite maintained in the vicinity of 9.0-10.0.
boiler water.
To assure complete oxygen removal, it is
necessary to maintain a residual concentra-
tion of sulfite in the boiler water. The resid-
ual required depends on a number of fac- ~
f3 10 0

tors such as the method of feed and the point ::; I/

~ 90
of application, the dissolved oxygen con- ...i50 •0
centration and the variation in the dissolved [....-
oxygen concentration of the feedwater. 1/ :;:::;
ill "'
Continuous feed of sodium sulfite is gener- 40
~ I
ally required for complete oxygen removal. 3
1/ v
In the majority of plants, the most suitable ~ 2
point of application is the storage compart- 4 • • 10
ment of the deaerating or open heater. In
other plants, sufficient reaction time will be Figure 23-2 • Effect of Di.fferent Waters on Time of
allowed with application to the suction side Oxygen-Sulfite Reaction

It has also been shown that some natural Catalyzed sodium sulfite is used in low
waters possess an inhibitory effect on the temperature systems for oxygen removal and
oxygen-sulfite reaction. Figure 23-2 shows the also finds application in boiler systems where
reaction rates at room temperature between the feedwater temperature is low, where me-
oxygen and sodium sulfite for three different chanical deaeration is not complete or where
types of waters. The reaction is fairly rapid it is essential to obtain rapid reaction for
with sea water, complete oxygen removal prevention of pitting in feed lines, closed
being secured in two minutes. The reaction heaters and economizers.
is much slower in distilled water, only 75%
of the oxygen being removed after 10 minutes
contact time. With a surface supply possessing I TI
inhibitory properties, after 10 minutes con- 10.0 I I I I
i 9. or-
t- t-
tact only approximately 3% of the oxygen SODIUM SULFITE -
f 8.0
has been removed. o1 t- f---+..-
~ 7.
Research directed toward increasing the ~ 6.
I -~
~ !1.0
-speed of the oxygen-sulfite reaction has de-
termined that certain materials act as cata-
lysts in speeding this reaction to completion.
0 I
2.0 i
The most suitable catalysts are the heavy I. 0
metal cations of two or more valences. Iron, ~ 20
2 3 4 !I 6 7 8 II 10
copper, cobalt, nickel and manganese are TIYE IN MINUTES

among the more effective catalytic aids to Figure 23-3 • Comparison of Sodium Sulfite vs Catalyzed
the oxygen-sulfite reaction. Combinations of Sodium Sulfite on Time of Oxygen Removal
several of these heavy metal cations have
proved most effective in providing a con- As boiler operating pressures have in-
tinuously active influence on the speed of creased, two disadvantages in the use of so-
reaction. dium sulfite as a chemical deaerant have
As a result of research on catalytic aids for become evident. The product of the sodium
oxygen removal, a catalyzed sodium sulfite sulfite-oxygen reaction is sodium sulfate
has been developed. Through the incopora- (Na2S04 ) which increases the dissolved
tion of suitable catalysts and sodium sulfite in solids concentration of the boiler water.
one formulation, a material is available While the increase in dissolved solids in
which will consistently provide practically low or medium pressure boilers is generally
instantaneous oxygen removal, even when the of little consequence, it can be significant in
water possesses natural inhibitory properties. high pressure boilers. Also, at high pressure,
the sulfite in the boiler tends to decompose
Figure 23-3 illustrates test results using
to form acidic gases, sulfur dioxide ( so2)
ordinary commercial sodium sulfite for oxy-
and hydrogen sulfide (H2S) which can con-
gen removal at room temperature with re-
tribute to corrosion in the return system.
sults obtained with catalyzed sodium sulfite.
As can be seen, the original oxygen content Hydrazine is a reducing agent which does
of 9.8 ppm has been reduced by sodium sul- not possess these disadvantages for high pres-
fite after 10 minutes contact to only 6.6 ppm. sure operation. Hydrazine removes dissolved
Little reduction has been made in the cor- oxygen in accordance with the following
rosive nature of the water. With treatment reaction:
by catalyzed sodium sulfite, however, com- N2H4 + 02 = 2H20 + N2
plete oxygen removal has been accomplished hydrazine + oxygen = water + nitrogen
in only 20 seconds contact time. Complete Since the products of this reaction are water
elimination of the corrosivity of the water and nitrogen, no solids are added to the
due to dissolved oxygen has been obtained boiler water. The decomposition products of
in this short contact time. hydrazine are ammonia and nitrogen. The

ammonia is alkaline and therefore will not in the boiler water to remove the normal
attack steel. However, if present in sufficient concentrations of oxygen without danger of
quantity, it can attack copper bearing alloys excessive ammonia liberation.
when oxygen is present. With proper appli- When feedwater is employed in desuper-
cation, the concentration of ammonia in the heating, it is desirable to feed sodium sul-
steam can be controlled so that the danger fite after the point from which the feedwater
of attack of copper bearing alloys will be for desuperheating is taken so as to avoid
minimized. At the same time, the ammonia adding solids to the water. When hydrazine
will neutralize carbon dioxide so that return is used as the chemical deaerant, this precau-
tion is not necessary since the hydrazine will
not add solids to the desuperheated steam.


Unless proper storage procedures are fol-

lowed, severe corrosion may occur in idle
boilers. The method of protecting idle boilers
depends primarily on the length of down-
time. Dry storage is preferred when the
boiler will be out of service for a month or
more while wet storage may be suitable for a
Figure 23-4 • Oxygen Pitting in a shorter period of downtime.
Two Inch Generating Tube
DRY STORAGE. The boiler should be drained,
line corrosion due to carbon dioxide will be thoroughly cleaned and dried completely by
reduced. means of hot air or a small wood fire. Close
Hydrazine is a toxic liquid and therefore attention should be given to complete elimi-
must be handled with the usual care observed nation of moisture from nondrainable super-
in the handling of all toxic materials. Pure heater tubes. A suitable moisture absorbing
hydrazine has a low flash point so a 35% material in a water tight container should be
solution, which has no flash point, is usually placed in the boiler drums or on top of the
employed. Theoretically, 1.0 ppm hydrazine flues in a fire tube boiler. The most widely
is required to react with 1.0 ppm of dissolved used moisture absorbants are quick lime and
oxygen. However, in practical application, silica gel. Silica gel is more efficient in ab-
hydrazine is usually required at approxi- sorbing moisture and can be regenerated by
mately 1.5-2.0 parts per part of oxygen. To heating so that it can be used over and over
avoid excessive ammonia concentrations in again. It is not a caustic substance like lime
the steam, the residual hydrazine concentra- so it can be handled more easily and safely
tion in the boiler water is usually maintained and is therefore generally preferred.
at 0.1 ppm or less. Requirements for quick lime approximate
In some instances, the feed of both sodium 20-30 lbs per 100 boiler horsepower or 7-10
sulfite and hydrazine is advisable in high lbs per 1000 lbs per hour boiler rating. Silica
pressure installations. In these cases, a very gel is recommended at the rate of 4 lbs per
low sulfite excess is carried in the boiler water 100 cubic feet of internal volume.
so that the potential for corrosion from After placing the quick lime or silica gel
acidic gases is minimized but a reserve of de- in the boiler, all openings should be tightly
aerant is available to react with an oxygen closed. The unit should be checked every
surge. In conjunction with the feed of so- two or three months, or as experience dic-
dium sulfite, the hydrazine is fed in a quan- tates, for renewal of the lime or regeneration
tity sufficient to maintain a very low residual of the silica gel.

WET METHOD. The boiler should be cleaned Nitrogen or other inert gases may also be
and inspected and then filled to the normal used for storage purposes. A slight positive
water level. If deaerated water is not avail- pressure of the gas is maintained after the
able, dissolved gases should be expelled by boiler has been filled to the operating level
boiling the water for a short time with boiler with deaerated feedwater. This method has
vented to atmosphere. The boiler water been employed in some high pressure utility
alkalinity should be adjusted with caustic stations or similar installations where short
soda to a minimum of 400 ppm. Sufficient periods of standby service are frequently re-
sodium sulfite should also be added to pro- quired.
duce a minimum sulfite residual of 100 ppm. SuPERHEATER STORAGE. In most modern
The quantities of caustic soda and sodium boilers it is not possible to separate the super-
sulfite required will approximate 3.0 and 1.5 heater section from the rest of the boiler.
lbs respectively per 1000 gals of water in the Accordingly, it is necessary to follow the
boiler. After the boiler is cooled and before same storage procedure for the superheater
a vacuum is created, the unit should be filled section as for the other portions of the boiler.
completely and all connections closed. Wet storage of drainable superheaters is rela-
Tests should be conducted on a weekly tively simple while wet storage of nondrain-
basis and additions of the treatment chemi- able superheaters is more complicated. In
cals should be made as necessary to maintain dry storage, care must be taken to remove
the recommended minimum concentrations. all of the moisture from the nondrainable
When treatment additions are required, the superheaters by reheating the superheaters
boiler water should be circulated by means sufficiently to evaporate all the water. This
of an external pump or by lowering the water may be accomplished by means of a small
to the operating level and steaming the boiler fire in the boiler furnace. In some cases it
for a short time. The boiler should then be may be possible to dry the nondrainable
completely flooded as outlined previously. superheaters with hot air diverted from the
The temperature of the boiler should be air heaters of one of the operating boilers.
maintained as low as possible since the cor- Depending on the actual design, there may
rosion rate increases at higher temperatures. be a choice as to whether the dry air is
When the boiler is returned to service, a directed over the external surfaces or used
high rate of blowdown should be maintained internally.
initially so that alkalinity and sulfite will be Since a residue will be left in nondrainable
reduced to the normal operating levels superheater tubes after boiling out, if the
rapidily. superheater has been flooded with water
In some small installations or where weekly containing boiler water salts, it is desirable
testing is not practical, chromate salts can be to employ a method of wet storage which
employed to protect idle boilers against cor- does not involve the use of solid chemicals.
rosion. The concentration maintained should Volatile chemicals or inert gas can be used
be 2000-2500 ppm as sodium chromate. The in the superheater section. The volatile chem-
boiler should be filled completely and closed icals recommended are hydrazine and am-
tightly. To assure good mixing, circulation monia or neutralizing amine. If high purity
of the water with a pump is recommended. water is not available to fill the entire boiler,
Boilers stored in this manner should be blown the superheater tubes can be filled with
down heavily to dissipate the chromate color, condensate or demineralized water from the
before being returned to service. The use of outlet end. The recommended treatment
chromate is not recommended if steam is to concentrations are approximately 100 ppm
be used in cooking, sterilization or in the of hydrazine and sufficient ammonia or
processing of food products since chromate neutralizing amine to elevate the pH to ap-
salts are toxic. proximately 9.0-10.0.

Figure 23-5 • Bottom of U-Bend in Superheater Tube Shol!Jing Pitting and Perforation


H. D. Ongman, "A Study of Reducing Solutions "Steam: Its Generation and Use", pp. 21-18, 19,
at Steam Boiler Temperatures", Combustion, Vol. The Babcock & Wilcox Co., New York, N.Y.
24, pp. 40-44 (Feb. 1953) (1955)
R. L. Reed, "Corrosion: Its Effect in Boiler Sys- "Suggested Rules for Care of Power Boilers",
tems", Combustion, Vol. 19, pp. 28-33 (May 1948) ASME Boiler and Pressure Vessel Code Section
and pp. 43-49 (June 1948) VII, Am. Soc. of Mech. Engineers, New York,
C. J. Schafer, "Protection of Idle Turbines and N.Y. (1954)
Superheaters," The Betz Indicator, Betz Labora-
tories, Inc., Phila., Pa. (March 1962)


Effects of Carryover . ...................... 175

Causes of Carryover. . . . . . . . . . . . . . . . . . . . . . . 175
Mechanical . .......................... 175
Chemical. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175
Prevention of Carryover. . . . . . . . . . . . . . . . . . . . 177
Mechanical. . . . . . . . . . . . . . . . . . . . . . . . . . . 177
Chemical. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178
Anti-Foam Agents . ..................... 180
Carryover Studies. . . . . . . . . . . . . . . . . . . . . . . 181

HE terms carryover, foaming and priming

T are closely associated and each term is
employed in describing conditions which
Carryover may result from mechanical and
cause the entrainment of boiler water solids chemical causes or from a combination of
with the steam. Because priming, foaming the two. Among the mechanical causes of
and carryover are sometimes loosely used carryover are boiler design, high water level,
interchangeably, it is well to define their method of firing and load characteristics.
exact meaning. Among the chemical causes are high total
Foaming is the condition resulting from solids of the boiler water, high suspended
the formation of bubbles on the surface of solids, high alkalinities, oil and improper
the boiler water. The foam produced may type of external or internal chemical treat-
entirely fill the steam space of the boiler or ment.
it may be of relatively minor depth. In either MECHANICAL. Boiler design will influence the
case, this foaming condition causes appre- solids content of the steam. Certain types of
ciable entrainment of boiler water with the boilers are known for their ability to con-
steam. sistently produce clean steam while certain
other types are traditionally recognized as
Priming is a more violent and spasmodic
subject to carryover.
action, resulting in the throwing of "slugs" of
boiler water over with the steam. This action Turbulence of the boiler water is a definite
is similar to the "bumping" experienced contributing factor to the mechanical carry-
when water is boiled in an open beaker. over of boiler water solids. In older boiler
designs the steam carrying or riser tubes dis-
Carryover is the term applied to the con- charged below the water level resulting in
tinuous entrainment of a relatively small severe turbulence. To minimize this condi-
quantity of boiler water solids with the steam. tion modern designs have the steam gener-
The term carryover is also often used to cover ating tubes discharge above water level or
the general conditions more exactly divided through steam separating devices. These de-
into foaming, priming and carryover. vices have become quite complex and are
generally very efficient. However, some pres-
EFFECTS OF CARRYOVER sure drop is experienced across the steam
purifier and thus any leak in the equipment
When boiler water solids are carried over in will result in impure steam.
the steam, deposits will occur in non-return High water level is another frequently en-
and stop valves, superheaters, steam piping countered cause of mechanical carryover
and engines or turbines. Such deposits can while boiler circulation is yet another factor.
promote failure of the superheater tubes and For example, at low load one burner may
loss in engine or turbine efficiency. Depo- cut out and the resultant disturbance of the
sition on governor valves can permit over- boiler circulation may influence the tendency
speeding and wrecking of the machines. to carryover.
Foaming or priming of the boiler water Obviously, operation of a boiler at loads in
can result in a false water level reading while excess of its designed rating will provide
the thermal shock resulting from slugs of more cause for carryover than will operation
water entering the steam distribution system within the designed rating. Sudden load
as a result of priming can seriously endanger changes will cause more difficulty than steady
all steam driven equipment. It should also be loads.
noted that when moisture is carried over CHEMICAL. The American Boiler Manufac-
with the steam a loss in thermal efficiency is turer's Association stipulates in its standard
experienced, as such carryover corresponds to guarantees on steam purity the following
additional blowdown through the steam line. boiler water concentrations.

Oil present in boiler water is highly un-

Total Suspended
Operating Total Solids, Alkalinity, Solids, desirable from the standpoint of carryover.
Pressure, psig ppm ppm ppm
Saponification of compounding agents by the
0 - 300 3500 700 300 alkalinity of the boiler water will produce a
301- 450 3000 600 250 soap, thereby causing a foaming effect.
451- 600 2500 500 150
601- 750 2000 400 100 Internal boiler water treatment with phos-
751- 900 1500 300 60 phate control is standard for most types of
901-1000 1250 250 40 feedwater. However, there is a tendency for
1001-1500 1000 200 20
1501-2000 750 150 10 the precipitated, finely divided calcium phos-
2001 and higher 500 100 5 phate to stabilize boiler water foam, thus
contributing to carryover of boiler water
The effect of dissolved solids, suspended solids with the steam. This carryover tend-
solids and alkalinity on carryover has been a ency must be controlled by increased blow-
subject of considerable research. In some down to lower suspended solids concentra-
plants alkalinity is the determining factor tions or by the use of organic agents to con-
while in other plants it is the total dissolved trol this increased foam load.
solids that limit the boiler water concentra- The increasing amounts of various syn-
tions that can be maintained without carry- thetic detergents and wetting agents, both
over. In general, it is agreed that part for for industrial and domestic use, may well
part suspended solids have a much more lead to increased foaming tendencies for
pronounced effect in causing carryover than many surface supplies to which wastes are
dissolved solids. discharged. Unlike soaps, which react with
The constituent having the most pro- the hardness present in water to form in-
nounced effect upon carryover-suspended soluble precipitates, these synthetic agents do
solids-is far less frequently shown on a boiler not react with hardness and their foaming
water analysis than alkalinity, dissolved solids, qualities are not thus eliminated, except by
sulfate, etc. Alkalinity is determined by titra- dilution. Contamination of surface waters by
tion with a standard acid; dissolved solids these agents in various industrial and do-
can be determined by conductivity methods mestic waste waters has caused difficulty
or by gravimetric analysis, but suspended with boiler foaming.
solids can only be determined by a tedious In many areas where the water supply is
gravimetric procedure. If sufficient data are critical, sewage plant effluents have been
available, however, it is possible to calculate used by industry for both boiler and cooling
the suspended solids content of the boiler water supply. Usually, the high concentra-
with reasonable accuracy. tion of organic wastes present in the sewage
plant effluent will lead to a foaming boiler
In certain sections of the country where
water condition unless controlled by anti-
the raw makeup water contains a high dis-
solved solids content it would be impossible foam agents.
to operate the boilers within the limits speci- The conventional mineral analysis of a
fied in the standard steam guarantee of the water does not indicate whether foaming
American Boiler Manufacturer's Association. tendencies are to be expected from organic
Boiler manufacturers are cognizant of this contamination. Even a determination of the
condition and will, in cooperation with the organic content of the water does not pro-
purchaser, supply boilers with extra large vide this information since many surface sup-
steam drums and with special baffling and plies from heavily wooded areas contain a
steam purifying devices in order that the relatively high organic concentration of
unusually high solids concentration of the harmless or beneficial organics of the tannin
boiler water can be maintained without type.
detrimental carryover. While all of the factors mentioned above

possess a definite bearing on the foaming drum internal purifying devices has become
characteristics of a boiler water, because of increasingly more complicated.
the inter-relation of these various influences, Primary separation of steam and boiler
it is not possible to generalize to the extent water is secured by means of abrupt changes
of specific maximum limits with respect to in direction of flow, utilizing the difference
alkalinity, suspended solids, dissolved solids, in density of water and steam as the means
etc. Minute traces of organic surface active of separation. Impingement of the steam and
contamination may materially alter the water mixture against a baffle for gravity
boiling characteristics of a concentrated boiler separation may be employed or centrifugal
water and yet the concentration of the sur- force may be used in a separating device
face active contaminants present may be so which causes the water to be thrown from a
small as to defy detection by chemical whirling mixture. The majority of the boiler
analysis. water impurities is removed in this primary
Operating conditions also vary to such an separation which also serves the purpose of
extent and are affected by so many factors removing steam from the recirculated boiler
that, except in a general fashion, it is not water returning to the tubes. Any appreciable
possible to predict the maximum water solids amount of steam in the downcomer circuits
that can be tolerated without carryover will reduce the head available for circulation
under certain operating conditions. Specific and will reduce the boiler circulation rate as
limits for any one set of operating conditions a result.
can usually be set only after a detailed indi- Primary separation can be secured by rela-
vidual study of operation at that particular tively simple devices, such as "end around"
plant. baffles over the discharge of the riser tubes
or curtain baffles protecting a simple dry
pipe. More complicated arrangements are
used to improve the efficiency of primary
The means that can be taken for the pre- separation. In many industrial boilers, pri-
vention of carryover can be divided into mary separation only is incorporated in the
mechanical and chemical methods. Of course, steam drum internals. For high pressure tur-
mechanical correction is required for the bine operation, however, it is advisable to
minimization of carryover resulting from follow primary separation with a secondary
mechanical causes, such as change in boiler separation step so as to reduce the solids con-
design, installation of baffles, etc. In addi- tent of the steam to the lowest value possible.
tion various mechanical devices are em- Secondary separation is also termed steam
ployed for the separation from the steam of scrubbing and steam drying. In this process,
the droplets of boiler water carried over. small amounts of moisture must be separated
Chemical methods for the treatment of the from large amounts of steam. This separation
boiler water are ineffective where the cause is accomplished by providing a large amount
of the carryover is mechanical. However, of contact surface and frequent reversals of
where carryover has a chemical origin, such steam flow so that the mist of boiler water
as high calcium phosphate concentrations in may collect and be drained from the sepa-
the boiler water, proper chemical treatment rating unit. Steam velocity is kept low to
methods can be highly effective in elimi- avoid reentrainment of the separated boiler
nating this condition. water and to insure maximum contact for
MECHANICAL. As pressures approach critical deposition of the boiler water.
the difference in steam and water density be- Steam washing is also sometimes employed
come progressively smaller and thus with to lower the solids content of the steam.
the trend to higher steam pressures and Steam washing employs the incoming feed-
higher steam generation, the design of steam water to mix with and dilute boiler water

solids with feedwater of lower solids content, Figure 24-2 illustrates another design of
thereby securing a lowering not of the mois- steam drum internals in which double rows
ture content of the steam, but of its solids of primary separators separate the entering
content. Various steam washing arrange- steam-water mixture. The second stage of
ments include bubble cap washers and feed- purification consists of so-called steam con-
water sprays through which the steam must densers which are U shaped finned tubes
pass. A condensing type steam washer is also through which feed water flows. Because the
used in which feedwater does not directly tubes and fins are of lower temperature than
contact the steam, but in which steam is con- the saturated steam a small portion of the
densed on finned tubes cooled by incoming steam condenses and maintains wet surfaces.
feedwater. In this design, the steam is washed Steam is thus effectively scrubbed and
in its own condensate. washed in pure condensate. The final stage
Figure 24-1 shows an installation of cen- of steam purification is secured in the dryer
trifugal primary separators in a boiler drum. cartons.
Steam and water from the risers enter the In still another design of steam drum inter-
separator tangentially. The water moves nals for large natural circulation boilers, the
downward in a helical path on the inside steam entering the drum (along with 20-40%
wall of the cylinder. Centrifugal force on the water by volume) is collected in a compart-
mixture boiling around the cylinder separates ment formed by internal baffles. The first
the steam from the water. The steam then stage of steam purification is in the centri-
flows through the hollow space in the center fugal separator which spins the steam-water
and passes out of the separator into cor- mixture, throwing water to the outside and
rugated plates at the top of the cylinder, steam to the inside. A skim-off lip above the
where moisture removed from the steam is skimmer collects the layer of water from the
arrested and returned to the cylinder. Sec- outer wall and returns it to the drum through
ondary separation is secured in the steam the annulus surrounding the inner chamber.
scrubber section. The steam then proceeds vertically upward
to the next stage.
Secondary separation is secured in two
Scrubber Element TO REDUCE SOliDS CARRY-OVER Scrubber Element
Corrugated Plates } TO EXTREMELY LOW VALUES Plates opposed banks of closely spaced thin cor-
rugated sheets which direct the steam in a
tortuous path at low velocity. Separated
moisture drains from the corrugated plates.
Additional separation takes place in the drum
space below the dryers and in the drying
Figure 24-3 is a line drawing of another
steam purifying device. In this separator the
centrifugal principle is again employed for
primary steam-water separation. The water
is removed through the bottom of the sepa-
rator through a slot discharging below mini-
mum water level while secondary steam puri-
fication is accomplished in a unit chevron
dryer as shown in Figure 24-4.
CROSS SECTIONAL ELEVATION CHEMICAL. As discussed earlier the principal
chemical causes of boiler water carryover are
Courtesy Babcock and Wilcox Co. high total solids, high alkalinity levels and oil
Figure 24-1 • Cyclone Separators contamination. By definition the total solids




co co
::::J ::::J
I- I-
(.!) (.!)
~ ~
<{ <{
::c ::c
u u

0 0
~ ~
I- I-
<{ <{
:;;: :;;:
~ :i
I- I-
(/) (/)



Courtesy Riley Stoker Corp.

Figure 24-2 • Steam Purifier for Large Capacity Boilers

of a boiler water is the sum of the dissolved mally be accomplished either by adjustment
and suspended solids. Of these, the sus- of boiler water controls, such as lowering con-
pended solids are considered to be the more centrations through additional blowdown, or
objectionable as their effect on foam stabiliza- by change or alteration of the external feed-
tion is more pronounced than that of the dis- water treatment facilities. For example filtra-
solved solids. If solids are determined to be tion following a hot process lime and soda
the cause of carryover, correction will nor- ash softener may be inadequate permitting a

facilities may be required such as the instal-

lation of a hydrogen zeolite unit to operate
in parallel with a sodium unit thereby re-
ducing the total alkalinity of the treated
The prevention of carryover difficulties
due to the presence of oil in the boiler water
is a problem that must be handled entirely
external to the boiler itself. No method of
internal treatment can be relied upon to over-
come a carryover problem caused by oil
contamination and the only safe policy that
can be followed is the removal of the oil
from the boiler feedwater.
ANTI-FOAM AGENTS. Frequently the cause of
a carryover problem cannot be economically
corrected through adjustment in boiler water
balances or installation of additional external
Courtesy Foster Wheeler Corp. treatment facilities. In many such instances
Figure 24-3 • Horizontal Steam Separator the use of effective anti-foam agents has
provided the desired result.
During recent years highly effective or-
ganic anti-foam agents have been developed.
In general the polymerized esters, alcohols
and amides are the most effective agents yet
developed. Various theories of foam inhibi-
tion by these compounds have been proposed
and one of the most logical is to the effect
that adsorption of the anti-foam agent on
the metal surfaces results in hydrophobic
conditions at the boiling surface. Fewer but
larger steam bubbles are generated under the
same heat transfer conditions and these
bubbles readily coalesce in place on the
heating surface, even prior to leaving that
surface and rising through the boiler water.
For the same amount of total steam produc-
Courtesy Foster Wheeler Corp. tion, tiny bubble formation is avoided and
Figure 24-4 • Unit Chevron Dryer non-foaming boiler water is the result of this
higher amount of suspended solids in the With the use of these anti-foams it has
boiler feedwater than can be tolerated. been possible to develop boiler water con-
Alkalinity as a contributing factor to carry- centrations in laboratory experimental boilers
over can also be corrected by adjustment or in excess of 150,000 ppm before carryover
changes in internal and/or external treat- was experienced. While no practical use can
ment facilities. If the alkalinity is a result of be made of such anti-foam power it clearly
the internal treatment program the use of demonstrates the effectiveness of the modern
less alkaline materials will correct the condi- anti-foam. Figure 24-5 graphically illustrates
tion. Should the cause be external, additional the results of one such laboratory experiment.


but also for techniques to measure such im-
purities in fractional parts per million. This
has been. achieved through the development
0 of highly sensitive conductivity equipment
and the sodium tracer technique which are

discussed in another chapter.
Carryover can be a serious problem in the
modern steam plant and frequently the
source of the carryover can only be estab-
'../ ./- lished by exhaustive studies employing sensi-
2000 4000 6000 8000 10000 12000 14000 16000
tive sampling and testing techniques. Science
has provided the water treatment engineer
with these tools and through their proper use
Figure 24-5 • Test Results Indicating Effect of
Anti-Foam Agent
he has been enabled to assist the plant
operator in the generation of maximum
While the prime purpose of anti-foam purity steam consistent with minimum blow-
application is the generation of high purity down requirements and clean boiler internals.
steam additional economic benefit is fre-
quently experienced through reduced blow- REFERENCES
down requirements. This has been proved on
numerous occasions by actual plant studies 0. de Lorenzi, "Combustion Engineering", pp. 20-
where anti-foam was applied initially to 1, 2, 3, Combustion Engineering-Superheaters,
Inc., New York, N.Y. (1949)
overcome a chemical cause of carryover and, M. H. Kuhner, "How Steam Is Released in
subsequent to the anti-foam feed, boiler water Water-Tube Boiler Drums", Power Plant Engi-
concentrations could be carried safely at neering; Vol. 48, pp. 92-94 (Sept. 1944)
much higher values without detrimentally R. A. Mumford, "A Theory of Carryover", Com-
bustion, Vol. 18, pp. 39-41 (Feb. 1947)
affecting steam purity. P. B. Place, "Investigation of Carryover Problems
CARRYOVER STuDIEs. The design of modern and Identification of Types", Proceedings, Engi-
neers Soc. of Western Penna., pp. 43-62 (1947)
steam driven equipment is such that toler- J, H. Richards, "Solving an Unusual Case of
ance for impurities in the steam is practically Carryover", lvfodern Power and Engineering, Vol.
non-existent. Thus there is an ever increasing 51, pp. 101, 102, 190-193 (Mar. 1957)
demand, not only for more effective steam C. D. Shields, "Boilers: Types, Characteristics,
and Functions", pp. 227-244, F. W. Dodge Corp.,
purifying equipment and anti-foam agents, New York, N. Y. (1961)

Silica Turbine Blade Deposits

Occurrence of Silica Deposits. . . . . . . . . . . . . . . . 183

Cause cif Silica Carryover. . . . . . . . . . . . . . . . . . . 184
Removal of Deposits.... . . . . . . . . . . . . . . . . . . . 184
Prevention ~~Silica Deposits. . . . . . . . . . . . . . . . 186

URBINE blade deposits develop from a culties may have numerous causes such as
T number of causes, but the general effect
is the same in all cases. That is, an adherent
condensate contamination or incorrect feed-
water treatment, they are not considered fur-
deposit develops in the steam passages and ther in this discussion. In recent years, a third
distorts the original shape of turbine nozzles and more troublesome type of carryover has
and blades. The deposits are frequently some- been observed, and this is the specialized
what rough or uneven at the surface, causing case of silica carryover.
increased resistance to the flow of steam. Dis- Non-selective carryover of boiler water
tortion of the steam passages will alter solids occurs most frequently in units operat-
steam velocities and pressure drops, and the ing at lower pressures, and various methods
overall effect will be one of causing less effi- of correction are available depending on the
cient recovery of the energy which would be cause. The deposits from such carryover have
available from the steam. Under sufficiently a large proportion of soluble salts, and wash-
severe conditions, the deposits may develop ing of turbines is usually sufficient to restore
unevenly and unbalance the turbine rotor. full efficiency. On the other hand, when
Corrosion or carryover are the principal selective carryover of silica takes place, the
causes of turbine blade deposits. Corrosion turbine blade deposits frequently contain up
may be caused by incomplete feedwater treat- to 95% silica, with the result that simple
ment, or by incorrect lay-up procedure when water washing is no longer adequate and
the turbine is not in service. The principal specialized cleaning must be used.
cause of corrosion is oxygen in the presence
of moisture, and less frequently, high carbon
dioxide content of the steam together with OCCURRENCE OF SILICA DEPOSITS
moisture. Deposits may also originate with Turbine blade deposits in which silica is a
normal carryover, gross carryover, or selective predominant constituent and which generally
carryover of boiler water solids with the tend to be insoluble are very rarely encoun-
steam. tered below 400 psi. As a result, such de-
With the deposition of solids on turbine posits were not recognized as an individual
blading, rapid loss in efficiency will occur. problem until comparatively recent years.
Records for stage pressures in the turbines Since that time, considerable investigation
will indicate progressively higher pressures. has been given to the problem and a number
Stage pressures will increase beyond the of groups are now cooperating to find a com-
manufacturer's recommendations, and fur- plete and effective solution to the problem.
ther operation will not be considered advis- The problem becomes increasingly severe
able. Most important will be the increase in with increase in operating pressure. At pres-
pounds of steam per kilowatt hour of power sures of 900 psi and above, the turbine blade
generated. deposits which contain a high percentage of
The term "normal carryover" refers to the insoluble silica are the rule instead of the
carryover which occurs in any boiler operat- exception.
ing under the best conditions. It is simply a With the development of fast steaming
matter of boiler design and solids content of high pressure boilers, advancement was also
the feedwater and the boiler water. With made with improved ability of the units to
modern, efficient steam generators, such generate steam of high purity. Limitation of
carryover is less than 1 ppm of total solids. total solids in the steam to a maximum value
Industrial plants, at times, encounter a of 1.0 ppm is now a standard specification
condition which may be described as gross and most large high pressure units consis-
carryover, but which is caused by priming or tently operate under such conditions. When
foaming of the boiler water, or by leakage of this value is compared to the solids content
steam purifying equipment. Since such diffi- of the boiler water, it is seen that the process

of purification is carried out to a high degree. dioxide in the steam on deposits of sodium
On the other hand, even a very small solids silicate resulting from normal carryover.
content of the steam will amount to carry- Straub has experimentally investigated the
over of several thousand pounds of boiler possibility of solubility of boiler water salts
water solids per year from a high capacity in steam, as well as vaporization of boiler
unit. Of this quantity, only a very small per- water· salts. The results of his investigations
centage need remain in the turbine steam indicate that the common boiler water salts
passages in order to form objectionable de- are soluble to an insignificant degree at pres-
posits. The problem is further complicated by sures below 2400 psi. Sodium chloride ap-
the fact that there is considerable evidence pears to be the most soluble common ma-
to prove that silica is carried selectively from terial, and appreciable amounts may be
boiler water at higher pressures in such a dissolved in steam when the steam is super-
manner that normal steam purification equip- heated and can come in contact with solid
ment cannot prevent contamination of the salt. This condition is not common in normal
steam. boiler water system operation. Since solubility
is based on temperature and pressure of the
CAUSE OF SILICA CARRYOVER steam, deposition of the salt would be ex-
pected by reduction in temperature and pres-
A number of theories have been proposed to sure, such as occurs in a turbine.
explain silica carryover and silica -deposits in
Evidence was also obtained to support the
turbine steam passages. Some of the better
theory that silica will actually leave the
known explanations are reviewed.
boiler as a vapor. The quantity in the steam
Normal carryover is due to mechanical will vary with boiler pressure and with con-
entrainment of minute droplets of boiler centration of silica in the boiler water. It was
water in the steam. Such droplets of water will also noted that pH value of the boiler water
evaporate to dryness in the superheater sec- has considerable influence on the quantity
tion of the boiler, and the solids will con- of silica in the steam. A decrease in silica
tinue to be carried along with steam in a carryover was noted with an increase in pH
manner similar to that by which dust par- value.
ticles are carried in air. If such carryover
Although silica is found on turbine blades
resulted in turbine blade deposits, it is to be
as crystalline and amorphous silica ( Si 0 2) it
expected that the deposits would have a
has been proposed that vaporization from the
composition similar to that of the boiler
boiler takes place as silicic acid (H2Si0s).
water. It is apparent that other factors enter
This would account for the variation in silica
the picture, since deposits containing up to
carryover with pH of the boiler water. To
95% silica are common.
some degree, this theory is supported by other
Experimental evidence has been developed experimental results. Other investigators have
to support the theory that silica is selectively reported a definite relationship between tur-
carried from the boiler as a vapor, or in solu- bine stage temperatures and the form in
tion in the steam. Other theories propose that which silica deposits. These results throw
the difficulty stems from mechanical entrain- some doubt on the theory that silica carry-
ment, but assume various methods by which over is in the form of silicic acid.
the deposits that result are principally silica.
The proposed theories are based on melting REMOVAL OF DEPOSITS
points of various salts as they pass through
the superheater tubes, adhesive properties of Two procedures are in widespread use for
sodium hydroxide particles resulting from in- removal of insoluble silica turbine blade de-
complete drying while passing through the posits. These are caustic washing of the tur-
superheater tubes, and the action of carbon bine and mechanical cleaning. Where the

facilities are available, and where operating Turbine manufacturers and central power
conditions permit, it is desirable to attempt station operators have developed procedures
caustic washing, since mechanical cleaning for turbine washing which experience has
generally takes a longer time and proves to shown can be safely performed without
be more costly. damage to the turbine. The procedures are
It is_ desirable to precede caustic washing available from a number of sources, and the
by plain washing, in order to determine the operation must- be carried out as recom-
extent to which the deposit consists of soluble mended in order to avoid the possibility of
salts. Shutting down and starting up a tur- damage to the turbine from mechanical and
bine after permitting the machine to cool thermal stresses.
accomplishes a certain amount of washing Caustic washing follows the same general
due to the condensation which takes place. procedure as water washing. Frequently, the
Where temperature of the steam varies with washing is started by using plain water, and
load, a certain amount of washing can also a solution of sodium hydroxide is gradually
be accomplished by operation at very light substituted. A 10 to 20% solution of caustic
load. Where washing is desired under con- soda is generally used, although under some
trolled conditions, it is preferable to use water circumstances, a stronger or weaker solution
injection in order to desuperheat the steam. may be desirable.

Figure 25-1 • Blading Deposit Analyzed to be 90-95% Silica


With simple washing, the progress of clean- 40 1"-
ing can be followed by measurement of
specific conductance of the condensate. How-
ever, when caustic washing is being used, it "'

is preferable to follow the progress of the "'
cleaning by silica determinations on the con- <l
() 10
densate. Progress of the cleaning must be fol- :::; 9
u; a
lowed closely, in order to determine that all 0:: 7 ......
UJ 6
deposits have been removed to the greatest ~ 5
possible extent.
Some plants are not in a position to resort
to caustic washing, since large quantities of
condensate are wasted during the operation. .........
In addition, if the deposits are sufficiently I'
heavy, it may be desirable to use mechanical I
500 600 700 800 900 1000 1100 1200 1300 1400 1500 1600 1700
cleaning in order to insure thorough removal BOILER DRUM PRESSURE, psig ,,
of all deposits.
Mechanical cleaning can be conducted by Figure 25-2 • Approximate Boiler Water Silica to
Limit Silica in Steam to 0.02 ppm
hand operations such as filing and scraping,
but is generally conducted by sand blasting.
Both fine sand and fly ash are widely used Steam purification systems which wash the
as the abrasive agents. During such an opera- steam with incoming boiler feedwater which
tion, care must be taken to prevent damage is low in silica, will slightly reduce the silica
to the metal surfaces. content of the steam. However, in most cases,
It is apparent that insoluble turbine blade the time of contact is too short for significant
deposits may be removed after they have de- reduction.
veloped, but it is also apparent that the The problem is essentially one of providing
procedure is costly, and in many cases highly feedwater with the lowest possible silica con-
inconvenient. Forced outage for cleaning may tent, and of preventing silica contamination
require overloading of other equipment, or from condenser leakage. After silica has en-
considerable expense for purchase of power tered the boiler water, the usual corrective
from an outside source. In the case of in- action will be blowdown within the ability
dustrial installations, plant shutdown may of the system to replace the blowdown water
even be necessary. The basic problem is with suitable makeup water.
therefore not one of removing deposits, but Internal treatment for removal of silica is
of preventing them in the first place. normally not practiced, but an interesting
application has been reported. Magnesium
oxide is widely used in hot process softeners
The experimental work of Straub comes for silica reduction with good results. Several
closest to fitting the facts of actual operating years ago, magnesium oxide was used intern-
experience, and on this basis, may be applied ally in a 1200 psi boiler. With feed of magne-
to actual operating conditions by making sium oxide, it was necessary to discontinue
suitable modifications, depending upon the feed of sodium phosphate, but particular care
circumstances of the individual problem. In was taken to practically eliminate calcium
units operating at 900 psi and above, it is contamination in the makeup water.
probable that the most significant factor of Condenser leakage is most readily detected
corrective action is maintenance of silica by providing instruments to continuously re-
content of the boiler water at least as low as cord specific conductance of the condensate.
10 ppm, but preferably less than 5 ppm. Such equipment must be maintained in good

working order and periodically checked. The circumstances, evaporated makeup IS too
chemical characteristics of the condenser costly and one or more of the other systems
cooling water must be kept in mind in inter- of treatment will provide satisfactory results
preting conductivity results on the conden- with favorable economy.
sate. For exarnple, the silica content of sea Evaporators may be obtained with per-
water is extremely low in comparison to the formance guaranteed so that the distillate
total solids content. At the other extreme, does not contain more than 0.5 ppm total
the silica content of some ground waters will solids with fresh water being used as evapora-
be quite high in comparison to the total tor feedwater. If sea water is used as feed-
solids. In addition to depending upon conduc- water the guarantee may specify that the
tivity measurements, it is also advisable to distillate will not contain more than ;,-4 grain
make periodic analyses for the silica content per gallon of total salts. Carryover from an
of the condensate. Determination of silica in evaporator corresponds to carryover from any
very low concentrations can be subject to low pressure boiler, and consists of mechani-
considerable error, unless the determination cal entrainment of the shell water in the
is conducted by an experienced analyst using evaporator vapor. The carryover from low
the proper laboratory facilities. pressure boilers is subject to many mechani-
The other principal source of silica con- cal factors, but evaporator design and opera-
tamination will originate with the makeup tion usually tend to eliminate or reduce the
water. The degree to which such contamina- factors which promote carryover. The most
tion can develop will depend upon the ex- important factor affecting vapor purity is con-
ternal system of treating the makeup water. centration of the shell water. Limiting concen-
The makeup water will usually be obtained tration may be specified by the evaporator
from a source such as condensate from a sys- manufacturer, and if it is not, then competent
tem operating at lower pressure, from evapo- advice should be secured.
rators, or from chemically treated water. Feedwater to the evaporator should be
Since it is necessary to maintain control deaerated and depending upon characteristics
over other factors aside .from silica content, of the feedwater, chemical treatment may
makeup water not from evaporators or con- also be advisable to eliminate scale formation
densate is usually treated to have essentially and sludge accumulation within the evapora-
"zero" hardness. This requirement calls for tor shell.
final treatment such as zeolite softening, de- The use of anti-foams to reduce carryover
mineralization, or hot phosphate softening. in high pressure boilers is well known. Similar
However, these processes may be preceded anti-foams may be used in evaporators to
by coagulation and filtration, hot or cold improve steam purity. The need for anti-
process softening using lime and other agents, foams is increasing due to expanding use of
and acid treatment. Choice of the exact treat- synthetic detergents which are present in in-
ment system is based upon chemical character- creasing quantities m many surface water
istics of available sources of makeup water supplies.
considered in conjunction with economy of The technology of demineralization has
appropriate systems. advanced to the extent that demineralization
Where a high pressure system operates is generally the most common method of
under conditions such that very little makeup preparing makeup water for high pressure
water is required, evaporation or deminerali- utility boilers. Present day demineralization
zation of the makeup water is generally the systems will produce water with very low
the most satisfactory system. Many industrial solids content including silica. In plants
installations use high pressure steam for proc- where makeup requirements are more than
ess requirements, and appreciable quantities about 10%, demineralization may not be the
of makeup water are required. Under these most economical method. Organic contamin-

Fig. 25-3
Occurrence of Silica in Natural Waters
Buffalo Elko Laurel Haverhill Savannah Warren Sea
N.Y. Nevada Miss. Mass. Ga. Ohio Phil. Is. Water

Hardness as CaC03, ppm .................. 118 180 10 30 107 160 84 6250

Calcium as CaC03, ppm .................. 85 135 7 25 79 105 41 1000
Magnesium as CaC03, ppm ................ 33 45 3 5 37 55 43 5250
"M" Alkalinity as CaC0 3, ppm....... ······ 90 436 66 17 109 24 130 115
Sulfate as S04, ppm ...................... 20 119 13 11 9 137 16 2650
Chloride as Cl, ppm 0 •••••••••••••• 0 •••••• 19 45 7 6 6 9 5 19000
Silica as Si02, ppm -··· 1.6 80 50 0.8 53 0.2 90 0.02-4.0
Source ....... ············· .............. Lake Well Well Lake Well River Well Ocean

ation of the available water source may make Figure 25-3 illustrates the wide variation
th.e use of ion exchange resins questionable in silica content of natural waters which may
unless thorough pretreatment is provided. be encountered. Some of the waters are of
The other systems of zeolite softening, hot course rarely encountered, but serve to illus-
and cold process softening are well known trate the fact that where water treatment is
and widely used. Each system has certain concerned, the problem calls for individual-
inherent limitations, and these are the guid- ized attention.
ing factors in selecting a suitable system. As The means by which silica is transported
an example, the silica removal which can from the boiler water to the turbine steam
be obtained with each system can be sum- passages, and the means by which deposition
marized in general as follows: takes place are problems which are still be-
a. Silica reduction to approximately 1 ppm ing widely investigated, and on which con-
with hot process softening. tinuing progress can be expected. A generally
b. Forty to seventy percent silica reduction applicable, clear-cut solution is not available
with cold process softening. for the problem and each plant presents an
individual problem. In addition to the prob-
c. No silica reduction with zeolite soften- lem of preventing insoluble silica turbine
ing, and the possibility of silica increase with blade deposits, a plant may be faced with
a siliceous type exchange medium. the problems of preventing corrosion, scale
With any of the systems of treatment out- or embrittlement. These complicating factors
lined, design and operation of the equipment also influence selection of the proper system
as well as choice of treatment chemicals de- of treatment.
termines whether optimum results are ob-
Some plants are fortunate in having more
than one source of makeup water available. E. E. Coulter, E. A. Pirsh and E. J. Wagner, Jr.,
"Selective Silica Carry-Over in Steam", Trans-
When this situation arises, a careful study actions, Am. Soc. of Mech. Engineers, Vol. 78,
should be given to each source of makeup pp. 869-873 (1956)
water, in order to avoid future difficulties F. G. Straub, "Steam Turbine Blade Deposits",
and provide best economy. Eng. Exp. Station Bul. 364, University of Illinois,
Urbana, Illinois ( 1946)

Measurelllent of Stealll Purity

History. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 190
Total Solids Determination. . . . . . . . . . . . . . . . . 190
Routine Mineral Analyses. . . . . . . . . . . . . . . . . . 190
Specific Conductance. . . . . . . . . . . . . . . . . . . . . . . 190
Tracer Techniques. . . . . . . . . . . . . . . . . . . . . . . 191
Steam Sampling. . . . . . . . . . . . . . . . . . . . . . . . . 193

ACCURATE measurement of steam purity is TOTAL SOLIDS DETERMINATION

1""\..of considerable significance in the oper-
ation of a steam plant. It is particularly so The collection of a sample of condensed
for the modern higher pressure plants where steam and the evaporation of 'the sample to
tolerances for deposition can be very critical. dryness and weighing of the solids residue is
With advancement in technique it is now another method employed in the determina-
possible to determine steam purity in parts tion of steam purity. This method is cumber-
per billion and it seems safe to state that some and time consuming, not lending itself
present knowledge permits reliable measure- to routine control. Satisfactory accuracy can
ment of steam contamination to a degree be obtained only with careful handling and
which should satisfy the demand of critical technique. The presence of 1 ppm of solids
high pressure operation. Thus it is possible in the steam will yield a weighable residue
to accurately check on boiler guarantee or of only 0.001 gram from a 1 liter sample.
effectiveness of steam purifying equipment. Naturally, therefore, a larger sample must
be taken in order to avoid the error in weigh-
ing and the sample must be carefully pro-
tected from dust, etc. The large size sample
must be evaporated down to dryness, under
The importance of steam purity measurement carefully controlled conditions. ASTM Ten-
in boiler operation was investigated as early tative Methods of Test for Suspended and
as 1865 and for some fifty years steam quality Dissolved Solids in Industrial Water (ASTM
determinations were accomplished mainly by Designation: D-1888-6lT) is intended for
steam calorimetry. Of the many types of application to water containing 5 ppm or less
calorimeters developed for this purpose the of total solids, and for which samples of
most familiar was the throttling calorimeter more than 4 liters must be evaporated. The
in which a sample of saturated steam is ex- American Boiler Manufacturer's Association
panded through an orifice to atmospheric stipulates that the total solids in the steam
pressure. The total heat in the steam sample for referee analyses shall be determined by
is greater than the total heat of the saturated testing no less than 10 samples of condensed
steam at atmospheric pressure, and the excess steam and taking the average of all results.
heat in the steam sample will superheat the
expanded steam or evaporate the moisture ROUTINE MINERAL ANALYSES
originally present in the steam sample.
The use of the conventional mineral analyses
From measurements of the pressure and covering such determinations as hardness,
temperature of the inlet and outlet steam, it alkalinity, chloride, pH, sulfate, etc., is of
is possible to calculate the moisture present
little significance in the analysis of condensed
in the original steam sample. The accuracy steam samples except of course where the
of the throttling calorimeter is usually taken
contamination of steam is excessive. The
as -+- 0.2% moisture which can mean an ap-
limit of accuracy of various test methods
preciable error in the solids content of the
such as hardness, chloride, etc. is such that
steam. For example, at a boiler water con-
they cannot be applied where the condensed
centration of 3500 ppm, an error of 0.2%
steam sample is reasonably pure.
moisture would correspond to an error in the
calculated total solids of the steam of 7.0
The wide use of superheated steam along The specific conductance of a condensed
with the need for more precise measurement steam sample, which is the measure of the
of steam purity and our broadening knowl- flow of an electric current through the
edge of steam properties have practically sample, is probably the most universally used
eliminated calorimetric measurement. method of determining steam purity. Pure

water has a very low conductance, but as and apply correction to the conductance
dissolved solids in a steam sample increase reading and the other is to degas the sample
so does conductance. This method of de- before measuring conductance. Accurate cor-
termining purity is rapid and reasonably rection curves are available to establish the
accurate. effect on conductance of carbon dioxide or
One of the disadvantages to the use of ammonia in a sample. While constant im-
specific conductance and which materially provements have been made in effectiveness
affects accuracy is the fact that some dis- of degassing equipment, traces of these gases
solved gases have a marked effect on the con- can remain. A further advancement in the
ductance of water. Such gases as carbon di- removal of dissolved gases has been the use
oxide, ammonia and hydrogen sulfide ionize of hydrogen ion exchange which can effec-
in water solution and thus interfere with the tively reduce ammonia to a negligible value.
use of conductance as a measure of dis- However, should the sample being tested
solved solids present. This interference can contain solids which are primarily in the
be appreciable when determining purity on form of sodium sulfate and sodium chloride,
a high quality steam sample. For example, in hydrogen ion exchange will produce acids
a sample containing less than 1 ppm dissolved which will result in 1an increase in conduc-
solids, specific conductance may be in the tance. There is no doubt that proper use of
range of 1.0 to 2.0 micromhos. The presence conductance and proper sample preparation
of any ammonia or carbon dioxide in this have greatly increased the reliability of this
sample will materially increase the conduc- method of steam purity measurement. Never-
theless the results of steam purity studies
using specific conductance and recently de-
veloped tracer techniques do indicate that
when measurement for solids in the order of
1. 0 ppm or less is desired, the accuracy of
conductance becomes questionable.


While the steam purity guarantee for the

modern high pressure steam generators states
that the steam will contain not more than
1.0 ppm of total solids for given boiler water
conditions, it is now generally accepted that
steam purity of this order is not satisfactory
for high pressure operation if deposition in
superheaters or turbines is to be avoided.
Tracer techniques involve testing for a com-
Figure 26-1 • Specific Conductance Flow Cell and
ponent of the solids present in the boiler
Recorder with One of the Earliest Dega.r.rer.r
water, which component can reasonably be
tance reading. Ammonia will impart con- expected to be present in the steam at the
ductance of 8.0 to 9.0 micromhos per ppm same ratio to the total solids as is the case for
of this gas, while carbon dioxide will increase the boiler water. For tracer technique studies
conductance on an average of 5.0 micromhos published so far radioactive phosphorus
per ppm of carbon dioxide in the lower (p 32 ), which was added to the boiler water
range of concentration. and the sodium ion which is normally present
Two procedures may be employed to off- as the ion in greatest concentration in the
set the effect of dissolved gases. One is to boiler water, have been employed. While the
measure the content of the gas in the sample use of radio-active phosphorus resulted in

Figure 26-2 • Sodium Recording Equipment Set Up for Plant Study

extreme sensitivity, it presents a problem of densed steam with a precision of 0.0004 ppm.
handling and disposal of the boiler water and An average ratio of 2.5 to 3.0 ppm of solids
would be too costly to be practical for actual per ppm of sodium has been established for
plant steam purity studies. For this reason boiler waters so the sensitivity of this tech-
the work on the use ·of tracer techniques as nique is in the range of 0.001 ppm of solids.
a means of determining steam purity has Initially, the use of the sodium tracer tech-
concentrated on the use of the sodium ion. nique for steam purity evaluation possessed
Employing a flame spectrophotometer it is the disadvantage of the steam sample having
possible to detect sodium content of con- to be tested in the laboratory. Extreme

Figure 26-4 • Strip Chart Showing Effectiveness

Figure 26-3 • Close-up of Burner Housing of Air Filter

care had to be taken in the sampling pro- and it was generally assumed that with the
cedure to avoid contamination by sodium use of either an ASTM or an ASME steam
which is always present in the atmosphere to sampling nozzle, correctly installed, good
some extent. This procedure was time con- steam sampling would be experienced. As
suming and did necessitate frequent resam- methods of measuring steam purity become
pling to verify results. Steady progress, how- increasingly sensitive, it has become obvious
ever, has been made in the technique of that our analytical techniques are far ahead
measuring sodium by flame spectrophotom- of our sampling techniques. A great deal of
eter. It is now possible to operate the experimental work on steam sampling
equipment in the boiler plant with contin- methods has been conducted recently in an
uous measurement and recording of the so- attempt to improve knowledge of this sub-
dium content of a constant flowing condensed ject. This work has been conducted under
steam sample. The problem of contamination the supervision of the Steam Contamination
from sodium in the air was eliminated Subcommittee of the Joint Research Com-
through the use of a highly effective air filter mittee on Boiler Feedwater Studies of the
which provides a sodium free air to the ASME.
burner of the spectrophotometer at a slight
positive pressure. Experience gained with this REFERENCES
equipment in many actual steam studies has
shown that steam purity evaluations to the R. V. Cobb and E. E. Coulter, "The Prevention
level of 0.001 ppm of solids can readily be of Errors in Steam Purity Measurement Caused
by Deposition of Impurities in Sampling Lines",
obtained. Proceedings, Am. Soc. for Testing Materials, pp.
It is of interest to note that of the many 1386-1395 ( 1961)
W. A. Crandall and W. Nacovsky, "The Develop-
plants which operate equipment for the con- ment and Operation of An Ultrasensitive Record-
tinuous monitoring of steam purity by con- ing Flame Photometer", Proceedings, Am. Power
ductivity, some steam generating plants op- Conf., Vol. XX, pp. 726-738 (1958)
perating at extremely high pressures have J. J. Maguire, "Sodium Test Measures Steam
Purity Accurately", Power Engineering, Vol. 62,
found the use of conductivity to be of in- pp. 81-82 (Sept. 1958)
sufficient accuracy for their needs. As a re- S. 0. Meyer, "Measurement of High Purity Steam
sult there are now several installations where by Continuous Sodium Recording", The Betz In-
dicator, Betz Laboratories, Inc., Phila., Pa. (Oct.
continuous monitoring of steam purity is ac- 1959)
complished by constant measurement and re- T. A. Miskimen, "Results of Steam Sampling Noz-
cording of the sodium content of the steam. zle Tests on Evaporator Vapor", Paper Number
59-A-301, Am. Soc. of Mech. Engineers, New
York, N.Y. (1959)
STEAM SAMPLING H. Phillips, "Closing Our Gaps in Our Knowledge
of Steam Sampling", Paper Number 59-A-287,
For our present knowledge of steam purity Am. Soc. of Mech. Engineers, New York, N.Y.
determination to be of practical value it is ( 1959)
essential that the sample being tested is truly E. A. Pirsh and F. G. Raynor, "Instrumentation
for the Determination of Steam Purity", Pro-
representative of the steam being generated. ceedings, Engineers Soc. of Western Penna., pp.
Detailed instructions for the installation and 79-88 ( 1956)
construction of steam sampling nozzles are J. H. Potter, "Steam Calorimetry", Combustion,
Vol. 29, pp. 51-55 (July 1957)
available in the ASME Power Test Code
J. K. Rice, "Steam Quality Measurements by
and the ASTM Standards. For many years Flame Photometer", Proceedings, Engineers Soc.
these standards have been widely accepted of Western Penna., pp. 89-101 (1956)

Condensate Return
Line Corrosion
Causes of Return Line Corrosion. . . . . . . . . . . . . . 195
Sources of Dissolved Oxygen. . . . . . . . . . . . . . . . . 196
Sources of Carbon Dioxide. . . . . . . . . . . . . . . . . . 197
Corrosive Effects of Carbon Dioxide. . . . . . . . . . . 197
Reduction of Carbon Dioxide Content of Steam. . . . 199
Internal Methods of Correction for Carbon Dioxide. 199
Ammonia . . . . . . . . . . . . . . . . . . . . . . . . . . . . 200
Neutralizing Amines. . . . . . . . . . . . . . . . . . . . 200
Filming Amines . . . . . . . . . . . . . . . . . . . . . . . 202
Heat Transfer. . . . . . . . . . . . . . . . . . . . . . . . . 203

problem of corrosion in return con-

densate systems can be brought under
control in most industrial systems by one of
several proven methods. This problem, once
responsible for expensive maintenance costs,
need no longer be of concern. In new plants,
precautionary measures should be taken dur-
ing design stages, rather than waiting until Figure 27-2 • Typical Grooving Attack by
expensive shutdowns show the need for Carbon Dioxide


Failures most frequently occur at threaded

joints because the metal thicknesses have been
reduced at these points. Figure 27-1 is an ex-
ample of such attack. Horizontal lines are
usually more severely attacked than vertical
lines due to incomplete draining of the con-
densate in systems that operate intermittently.
The corrosive action manifests itself in the
form of grooving or channeling for the most
part. Figure 27-2 is a cutaway section of pipe
illustrating the typical grooving attack caused
by carbon dioxide. Figure 27-3 shows another
pipe sample typical of carbon dioxide attack.
Figure 27-4 shows a condensate line in
which the corrosive attack is primarily in the
form of pitting. Such pitting attack is typical Figure 27-3 • Example of Pipe Wall Grooving ~y
of oxygen. Carbon Dioxide
The principal causes for steam and con-
densate corrosion are the dissolved gases car-
bon dioxide and oxygen. These gases must
be in solution before they become aggressive.
There are other influencing factors such as
rate of condensation, time of contact, tem-

Figure 27-1 • Carbon Dioxide A/lack at

Threaded Joints Figure 27-4. • Pz'tting Attack Typical of Oxygen

perature, metal heterogenity, galvanic action (5)

and others but only because they supplement 2Fe(HC03h + lf2 Dz = FezOl +
the basic factors. ferrous bicarbonate + oxygen = ferric oxide
The corrosion reaction for dissolved oxygen + 4COz + 2Hz0
carbon dioxide + water
may be illustrated as follows:
(1) A combination of reactions ( 4) and ( 5)
4Fe + 6H 0 +
2 30 2 = 4Fe(0Hh can be written to show ·the production of
iron + water + oxygen = ferric hydroxide Fe 3 0 4 , the magnetic oxide of iron (magnetite)
as illustrated by equation 6.
The ferric hydroxide formed may later re-
vert to iron oxide rust as follows : (6)
3Fe(HC03h + lf2 Oz = Fe304 +
ferrous bicarbonate + oxygen = magnetite
2Fe (OHh = FezOl + 3Hz0 + 6COz + 3Hz0
ferric hydroxide = ferric oxide + water carbon dioxide + water
When carbon dioxide dissolves in water it
forms carbonic acid which is a weak acid and In some cases, the decomposition of ferrous
produces corrosion in keeping with the fol- bicarbonate has resulted in the production of
lowing equation: deposits of ferrous carbonate:

(3) (7)
Fe + 2HzC03 = Fe(HCOllz Fe(HCOlh FeC03 +
iron + carbonic acid = ferrous bicarbonate ferrous bicarbonate = ferrous carbonate +
+ Hz COz + HzO
+ hydrogen carbon dioxide + water

The above reaction proceeds rapidly at pH From the above reactions, it can be seen
values below 5.9. The reaction product, fer- that deposits of FeO, Fe20a, Fea04, FeCOa
rous bicarbonate, tends to elevate the pH and and mixtures of these can be found present
as the pH increases above 5.9, the reaction in steam and condensate systems and are the
rate becomes reduced. The hydrogen formed type deposits usually responsible for plugging.
in the reaction will also have a retarding
effect particularly when the pH exceeds 5.9. SOURCES OF DISSOLVED OXYGEN
The ferrous bicarbonate formed in this
reaction is soluble and can be carried along Oxygen in steam and condensate systems may
in the condensate. However, at points of pres- originate directly from the boiler feedwater
sure drop or reduced carbon dioxide content or may enter at various points in the con-
in the vapor phase it will precipitate from densate system. Oxygen is present in most
solution in keeping with the following re- makeup waters and unless the makeup water
action: is brought to saturation temperature in a
properly vented open or deaerating heater,
appreciable amounts will remain in the feed
Fe(HCOlh FeO + water entering the boilers. When the feed
ferrous bicarbonate = ferrous oxide +
2COz + HzO water is subjected to boiler temperature the
carbon dioxide +water oxygen will be released and a portion of it
can be entrained with the steam and find its
When dissolved oxygen is also present, it way into the condensate system. Oxygen that
will influence the reaction rate by removing · enters in this fashion can be coped with
the hydrogen to form water. The ferrous bi- readily by use of mechanical or chemical
carbonate will precipitate as follows: deaeration of the feedwater.

In vacuum return systems or in gravity (8)

type return systems there is ample opportu- 2(HC0l)- + heat = (COli= +
nity for oxygen infiltration. While condensate bicarbonate + heat = carbonate +
that is returned under pressure will usually C02 + H20
be free of oxygen, infiltration can take place carbon dioxide +water
where such systems operate intermittently. (9)
(COli= + ~hO + heat = 2(0H)- +
It is not uncommon for cooling water to be carbonate + water + heat = hydroxide +
injected into the vacuum pumps to lower the C02
temperature of the condensate. This practice carbon dioxide
is also followed in certain gravity type return Reaction (8) proceeds to 100% completion
systems where the condensate is collected in whereas reaction (9) is only partially com-
a centrally located receiving tank for pump- plete with decomposition taking place to ap-
ing back to the boiler plant. The cooling proximately 80%. Based on these reactions
water is usually laden with oxygen and its it can be calculated that for each 1 ppm of
introduction into the condensate generally bicarbonate alkalinity (expressed as calcium
creates a corrosion problem that manifests carbonate) in the boiler feed water, 0. 79 ppm
itself in the form of pitting. Where cooling of carbon dioxide will be evolved. For each
water must be added, it should be controlled
1 ppm of carbonate alkalinity (expressed as
thermostatically so that the amount intro- calcium carbonate) 0.35 ppm of carbon di-
duced is held to an absolute minimum. To
oxide will be evolved.
avoid corrosion from the introduction of cool-
ing water, catalyzed sodium sulfite, which The other possible and usually minor
permits rapid reaction of the sulfite with sources of carbon dioxide are the free, gase-
oxygen in the cold, can be proportioned to ous carbon dioxide that is dissolved in most
the cooling water to react with the oxygen natural waters and the carbon dioxide which
and eliminate it from the water. forms from decomposition of soda ash that
may be used for regulation of boiler water
In general, the elimination of oxygen from alkalinity. Free carbon dioxide is almost com-
condensate systems can be accomplished pletely eliminated by efficient feedwater de-
through relatively simple means and for this aeration and consequently is rarely a factor
reason corrosion and pitting due to dissolved in steam and condensate corrosion. The use
oxygen is readily controlled. of soda ash is inadvisable where return line
corrosion is being experienced and should be
SOURCES OF CARBON DIOXIDE replaced by caustic soda.

Carbon dioxide is the usual cause of steam CORROSIVE EFFECTS OF CARBON DIOXIDE
and return line corrosion and the corrosion
is characterized by a general thinning of the The curves illustrated by Figure 27-5 show
pipe wall or grooving along the bottom of the that other conditions being fixed, corrosion is
pipe. proportional to the carbon dioxide concentra-
The chief source of carbon dioxide is the tion of condensates.
bicarbonate and carbonate alkalinity of the The rapid increase in corrosion rate with
makeup water to the boilers. The bicarbon- increase in carbon dioxide concentration and
ate and carbonate alkalinity when subjected with increase in flow rate is illustrated
to boiler temperature undergo thermal de- graphically.
composition and liberate carbon dioxide Even at low carbon dioxide concentrations,
which becomes entrained with the steam. The however, corrosion can take place at isolated
decomposition is illustrated by the following points in the condensate system. For example,
equations: the curve of Figure 27-5 shows that for any

1- 40
l.LJX 30
0 a::
I ~
. r: >- 20
<1:·· 0::
z a..
CJ) 10
0 0+-----~------~-----r------~-----r------r------i
0 100 200 300 400 500 600 700
Figure 27-5 • Effect of Condensate Flow and Carbon Dioxide Content on Rate of Corrosion

concentration of carbon dioxide, the rate of negligible amount of carbon dioxide dissolved
corrosion increases with the quantity of con- in it, due to the fact that the partial pressure
densate flowing, or in other words, corrosion of the carbon dioxide in the vapor phase is
is a function of the total pounds of carbon low. Assuming that the incoming steam con-
dioxide. Therefore, in a given system corro- tains 20 ppm of carbon dioxide, the initial
sion may only be a serious problem where the condensate discharged by the unit will have
quantity of condensate handled is high, such less than a fraction of 1 ppm. The remaining
as in the main return line. carbon dioxide will accumulate in the vapor
In order for carbon dioxide to be corrosive phase above the condensate in the unit. As
it must be in solution. The laws of Henry condensation continues the accumulated car-
and Dalton govern the solution of gases in bon dioxide will develop a partial pressure
liquids and state in simple form that the con- sufficiently high to permit re-solution of more
centration of a gas dissolved in a liquid is and more carbon dioxide, until equilibrium is
proportional to its partial pressure (or con- attained, under which condition the carbon
centration) in the contacting gaseous phase. dioxide content of discharged condensate will
The application of these laws serves to ex- equal that of the incoming steam.
plain severe corrosion that occurs in units It can be calculated that the quantity of
condensing steam that operate intermittently. carbon dioxide that must accumulate in the
For example, when a drier or unit heater is gaseous phase will be in the range of several
first supplied with steam and condensation hundred to more than a thousand ppm before
occurs, the resulting condensate will have a equilibrium conditions are attained. If such

units operate intermittently, on cooling and as in the anion exchange resin. Normally, de-
pressure on the unit decreases considerable mineralization is limited to those plants where
quantities of carbon dioxide will go into solu- high percentage of returned condensate is
tion. The pH of the condensate will be expected, due to the cost of this process.
lowered and accelerated corrosion will take In addition to hydrogen zeolite neutralized
place. It is apparent, therefore, that corrosion with caustic, raw water or blending with
in isolated units can take place even though sodium zeolite, other ion exchange processes
the concentration of carbon dioxide in the can be used. Softeners operating on the
steam is relatively low. Because of the ac- sodium cycle do not reduce the bicarbonate
cumulation of carbon dioxide that is possible, alkalinity of the raw w;ater. However, acid
venting of the units continuously during oper- treatment followed by aeration can be used
ation or shortly before shut-down would following the softening process. Dealkaliza-
prove of material benefit. It is also essential tion by chloride-anion exchange is another
that such units be provided with proper method, in which the makeup water is first
drainage. passed through a sodium unit and then
through a second anion exchange unit re-
generated with salt. In this latter process,
REDUCTION OF CARBON DIOXIDE CONTENT alkalinity is reduced without the use of acid,
OF STEAM but chemical efficiency is poor compared to
other systems. The use of chloride-anion ex-
change is generally limited to small plants
The various softening processes in which lime where objection to handling acid is the in-
is employed, such as lime softening, lime-soda, fluencing factor.
and lime-gypsum softening, secure a reduc-
tion in alkalinity by precipitation of calcium By selection of the proper external treat-
and magnesium bicarbonate by the lime. Hot ment processes, it is possible to produce a
lime-hot ion exchange is another example of minimum carbon dioxide content of the
this type of process. It should be emphasized steam in the range of approximately 5 ppm
that while these processes reduce the feed- when using 100% makeup water. In most
water alkalinity, they do not completely systems, except at points where stratification
eliminate it. With normal balances, utilizing may occur, corrosion should be kept to a
hot process lime and soda softening and satisfactorily low level. Also, at this low level
100% makeup, the carbon dioxide content of carbon dioxide, neutralizing amines can be
economically considered.
of the steam will be in the range of 17 - 21
The various processes using acid, such as INTERNAL METHODS OF CORRECTION
direct acid neutralization or hydrogen zeolite FOR CARBON DIOXIDE
softening, will reduce the alkalinity of the
treated water to a relatively low value. With Of historical interest only was the attempt to
hydrogen zeolite softening, chemical efficiency use inorganic chemicals to control carbon
is lower, but softening is accomplished simul- dioxide corrosion in condensate lines, as these
taneously. Where very low carbon dioxide methods proved to be ineffective. These
values are required, necessary alkalinity for methods involved the use of an alkali, such
good boiler operation can be obtained by as sodium hydroxide to neutralize carbon
caustic soda addition. dioxide or polyphosphates to form an iron-
Almost complete carbon dioxide removal phosphate film. The injection of these in-
is secured by demineralization, in which both organic materials to the steam was objec-
cations and anions are removed either by tionable for the same reasons that carryover
aeration, vacuum deaeration, or adsorption of boiler water solids is undesirable.

AMMONIA industrial plants, and consequent higher am-

monia requirements, the problem of copper
Ammonia has been employed for neutraliza- attack limits the application qf ammonia.
tion of carbon dioxide and elevation of pH.
Ammonia can be fed to the system as am- NEUTRALIZING AMINES
monium hydroxide (a solution of ammonia
gas in water) or an inorganic salt such as Neutralization of carbonic acid can safely be
ammonium sulfate can be used. At boiler accomplished by certain volatile amines such
temperatures, the ammonium sulfatq will be as cyclohexylamine, CsHuNH 2 and morpho-
decomposed with the liberation of ammonia line C 4 H 9 NO. These amines when fed to a
gas with the steam. boiler volatilize with the steam and combine
Ammonia has been successfully employed with the carbon dioxide in the condensate
for control of C0 2 corrosion and iron pick-up to neutralize its acidity. Unlike ammonia the
in central stations with low percentage make- amines in low concentrations are not corro-
up and low carbon dioxide concentration in sive to copper and zinc bearing metals. Con-
the steam. While a reduction in corrosion of trol of treatment is usually based on feeding
ferrous metals can be secured in this manner, sufficient amine to raise the pH of the con-
copper and zinc bearing metals can be seri- densate to 7.0. Satisfactory reduction of car•
ously corroded, particularly when oxygen is bon dioxide corrosion is obtained with the
also present and there is opportunity for use of the neutralizing amines in the absence
ammonia to concentrate. Because of the of oxygen.
higher C02 concentrations usually found in Figure 27-6 shows cyclohexylamine require-


25 50 75 100 125 150 175 200 225

Figure 27-6 • pH AdJustment with Cyclohexylamine




MORPHOLINE (40%), ppm

Figure 27-7 • pH AdJustment with Morpholine

ments for different concentrations of carbon in this section of the turbine to minimize cor-
dioxide in the condensate. Increased carbon rosion and iron pick-up. Morpholine is par-
dioxide concentrations require increased ticularly suited to this use, because it is less
quantities of cyclohexylamine. Cyclohexyl- volatile than ammonia or cyclohexylamine.
amine requirements are shown for the 40% The distribution ratio of ammonia and
solution since this is a common commercial the amines can be used as one means of
strength employed to minimize fire hazard selecting the proper material. Distribution
with the material. Approximately 3.0 ppm ratio is the ratio of the concentration in the
cyclohexylamine (40%) are necessary for steam compared to the concentration in the
each 1.0 ppm carbon dioxide to elevate the condensate. Ammonia has a distribution ratio
pH of the condensate to 7.0. of 10, cyclohexylamine about 3 and morpho-
Morpholine requirements for different line only 0.4. This favorable distribution ratio
levels of carbon dioxide are shown in Figure for morpholine makes it the more applicable
27-7. Morpholine is usually employed as a amine for central stations to protect the wet
40% solution to ;minimize fire hazard. Ap- steam sections in high pressure turbines, since
proximately 3.6 ppm morpholine ( 40%) are high pH is imparted to initial condensed
required for each 1.0 ppm carbon dioxide moisture at a minimum feed rate. Cyclohexyl-
to elevate pH to 7.0. An interesting applica- amine can be expected to remain with the
tion of the neutralizing amines has been in steam. At points of initial condensation,
the prevention of corrosion in the wet end higher rates are necessary to maintain desired
section of high pressure central station tur- pH levels. This property of cyclohexylamine,
bines. In these sections of the turbine, con- however, is of benefit in extended systems
densed moisture is traveling at high velocity utilizing reduced pressure wet steam, espe-
and tends to remove any protective oxide cially where excess condensate is frequently
films. It is desirable to develop a pH of 9.0 trapped off.

This difference in the action of these two

most commonly used neutralizing amines can
be utilized to select the proper material for
any given plant. Blended neutralizing amines
are available, possessing the desirable char-
acteristics of morpholine and cyclohexyl-
amine. Such a blend may be the most ap-
plicable material in many industrial plants
where protection is desired both at points of
initial condensation, as in turbines, and in
low pressure steam systems used for heating.
Neutralizing amines do not protect return
lines against oxygen attack. However, with
proper deaeration of feedwater and other
steps to prevent oxygen pick-up, this dis-
advantage can be considered minor. Since
the neutralizing amine requirement is a func-
tion of the carbon dioxide content of steam,
judicious selection of external treatment
methods may be necessary to achieve greatest


The filming amines function on a completely

different principle from the neutralizing Figure 27-8 • Water Repellancy of Octadecylamine
amines. The filming amines do not neutralize Film at Left. Untreated Specimen at Right.
carbon dioxide. Instead, they function by
forming on the metal surfaces contacted an
impervious non-wettable film that acts as a a filming amine. It can be seen how the
barrier between the metal and the conden- amine film causes the water to assume drop-
sate, protecting against both oxygen and car- let form, on both the large and the small
bon dioxide attack. When the filming amines drops. The untreated specimen at the right
are adsorbed on a metal surface, water will does not repel water and the water spreads
not wet that surface. The film formed and wets the metal surface.
through the use of these amines is of sub- The filming amines of value in the preven-
stantially monomolecular thickness and does tion of corrosion are the high molecular
not increase in thickness with continued treat- weight amines and amine salts having straight
ment. The amine will penetrate corrosion carbon chains containing 10-18 carbon atoms.
products present on a metal surface and Octadecylamine (C1sHa1NH2), hexadecyl-
deposit a protective film on the metal. While amine (C 16 HaaNH2) and dioctadecylamine
continuous treatment is advised, the adsorbed (C 36 H 74 NH) are examples of useful materi-
film on the metal surface is quite durable als of this type. In plant practice the most
and is not removed during short periods of successful results have been obtained with
discontinued treatment. the dispersed octadecylamine.
Figure 27-8 is a photograph of two test Octadecylamine and octadecylamine ace-
specimens on which a few drops of water tate are available in either dry, flaky form or
have been placed. The specimen on the left in the form of a dispersion which simplifies
has been exposed to condensate treated with the problem of solution and chemical feeding.

Condensate should be used in preparing the from the standpoint of improved heat trans-
chemical soultion which is then fed with a fer is the ability of the filming amines to
small chemical pump to the main steam line loosen and remove old corrosion films. Dur-
at a point where it will be well mixed with ing initial start-up, the amines are applied at
the steam. The material is dispersed readily a low rate to permit gradual removal of
with the steam and is carried through the corroswn products in the system.
·steam and condensate lines, forming a non-
wettable film on the metal surfaces contacted. REFERENCES
Unlike the neutralizing amines which must
be fed in direct proportion to the carbon di- A. A. Berk, "Treating Steam Chemically to Reduce
Return Line Corrosion", Industry and Power, Vol.
oxide concentration of the steam, filming 53, pp. 79-81 (1947)
amines are required at a rate sufficient to L. F. Collins and E. L. Henderson, "Corrosion in
establish and maintain the desired corrosion Steam Heating Systems", Heating, Piping and Air
Conditioning, Vols. 11 and 12, Sept. 1939-May
resistant film on the metal surfaces. 1940 incl.
E. Elliott and P. J. Gaughan, "Plant Stops Return
Line Corrosion-Saves $14,200 a Year", Power
HEAT TRANSFER Engineering, Vol. 55, pp. 104-109 (Aug. 1951)
H. L. Kahler and J. K. Brown, "New Polar Film
As previously stated, filming amines function Treatment for the Control of Return Line Corro-
as a barrier treatment forming a non-wettable sion", Proceedings, Engineers Soc. of Western
Penna., pp. 115-126 (1949)
film of substantially monomolecular thickness.
G. A. Mierendorf, "Protecting Heating Units and
This film does not impede heat transfer due Condensate Lines Against Corrosion by the Use of
to its very thin nature. Also, due to the non- Filming Amines", Proceedings, Am. Power Conf.,
wettable feature, dropwise condensation is Vol. XIV, pp. 453-458 (1952)
W. A. Tanzola and J. G. Weidman, "Film Forming
promoted actually producing an increase in Corrosion Inhibitors Also Aid Heat Transfer", The
heat transfer. Of even greater importance Paper Industry, Vol. 36, pp. 48-50 (April 1954)

Elllbrittlelllent of Boiler Metal

Causes of Embrittlement. . . . . . . . . . . . . . . . . . . 205

Embrittling Characteristics. . . . . . . . . . . . . . . . . 206
Treatment Methods . ...................... 207
Status of Problem . . . . . . . . . . . . . . . . . . . . . . . . 208

cracking or caustic
metal embrittlement generally occurs· in
boilers along riveted seams and at tube ends.
Cracking of the metal may cause dangerous
weakening of its structure and subsequent
boiler failure.


There are three factors necessary to produce

intercrystalline cracking of boiler metal.
These are:
1-Leakage of the boiler water must take
place so as to permit escape of steam and Figure 28-1 • Embrittlement Cracks in a 1 Inch Plate.
concentration of the boiler water at the point The Cracking in this Drum Was So Extensive that the
of leakage. Drum Had to be Replaced.
2-The boiler metal must be subjected to
high stress. This stress may be internal, re-
sulting from cold working of the metal, or the
stress may be external, due to expansion and
3-The concentrated boiler water must
possess embrittling characteristics and chemi-
cally attack the boiler metal.
Leakage of boiler water from riveted seams
or rolled-in tube ends can produce the re-
quired concentration of the boiler water. In-
vestigations have shown that intercrystalline
cracking is not produced at "normal" boiler
water concentrations, but that continued
evaporation may build up local concentra-
tions of sodium hydroxide to a range such as Figure 28-2 • Cracks in the Rolled End qf a Tube.
75,000 to 500,000 ppm which will cause in- These Cracks Were Thought at First to be the Result qf
tercrystalline cracking under stress and in the Excessive Rolling, but Further Investigation (and many
absence of an inhibitor. tube failures) Showed the Cause to be Embrittlement.
Even in boilers with welded drums, inter-
crystalline cracking may manifest itself at
rolled-in tube ends. In general, therefore, the
power plant engineer has little control over
the factors of leakage and stress.
Inasmuch as all three factors of leakage,
stress and embrittling character of the boiler
water are required for the production of in-
tercrystalline cracking, if it can be definitely
shown that one of these three factors is ab-
sent, it can reasonably be assumed that the
danger of embrittlement is non-existent in Courtesy Mutual .Boiler and Machinery Insurance Co.
that instance. The factor that can best be Figure 28-3 • Mud Drum of Boiler Which Exploded as
shown to be present or absent is the embrit- a Result of Caustic Metal Attack Weakening the Seam.

Courtesy Mutual Boiler and Machinery Insurance Co.

Figure 28-4 • Wreckage of the Power House Caused by the Explosion of the Boiler Shown in Figure 28-3.

tling character of the boiler water. If a boiler must be taken to render it non-embrittling.
water is naturally non-embrittling or has been
made non-embrittling by the addition of
embrittlement inhibitors, the power plant en-
gineer can neglect the other two factors of Naturally, it is desirable that chemical con-
leakage and stress as affecting the embrittle- trols of some nature be established so that
ment problem. However, should it be defi- chemical analysis of the boiler water will
nitely shown that the boiler water does pos- indicate the ability of that boiler water to
sess embrittling characteristics, the definite crack steel. However, the methods of study-
possibility exists for intercrystalline cracking ing embrittling characteristics today are more
should the additional factors of leakage and direct. They consist of placing a steel speci-
stress also be present. men in contact with a boiler water under ex-
Good engineering practice, therefore, dic- treme concentrations and simultaneously
tates that the boiler water be investigated for placing stress on the steel specimen in order
embrittling characteristics. If the boiler water to reproduce under test conditions the same
is found to be embrittling, naturally steps circumstances which will cause cracking in

shown that cracking of the specimen will re-

sult if the boiler water is embrittling.
WATER IN The Embrittlement Detector is most con-
veniently installed in the continuous blow-
down line of a boiler. If continuous blowdown
is not installed, boiler water may be taken off
at any convenient point and either returned
to the boiler or run to waste. The require-
ment is that water at substantially boiler
temperature circulate rapidly through the
The use of the Embrittlement Detector,
while requiring attention once daily, possesses
the advantage of operation over a thirty,
sixty or ninety day period which tends to
take into account fluctuations in water char-
acteristics during that time. In addition, the
unit is operated on the actual boiler water at
the plant. Occasionally, the specimen will
show definite cracking on its removal from
the block although in many instances the
WATER OUT cracking is revealed only on further bending
in the course of a subsequent examination of
the specimen bar in the laboratory of the
Figure 28-5 • Embrittlement Detector supervising organization.

the boiler itself.
This testing is conducted with the use of The use of sodium nitrate is practically stand-
the Embrittlement Detector as developed by ard for the inhibition of embrittlement. While
the United States Bureau of Mines. This nitrate is naturally present in many water
testing unit, covered by U.S. Patents 2,283,954 supplies, the inhibition of embrittlement re-
and 2,283,955, is illustrated in Figure 28-5. quires a definite ratio of nitrate to the caustic
It consists of a rectangular block 5 inches alkalinity present in the boiler water. The
long, 214, inches thick and 3;;2 inches wide formula for calculating this sodium nitrate/
with a % inch hole bored through it, as illus- sodium hydroxide ratio in the boiler water is:
trated, through which boiler water circulates. NaN0 3/NaOH ratio=
The test specimen is a steel bar 5 inches Nitrate as N0 3, ppm X 2.14
long, % inch wide and ;;2 inch thick which M.O. alkalinity as CaC03, ppm-
is bolted into a slot in the block by a clamp- Phosphate as P04, ppm
ing plate and four studs. The clamping of
The ratios recommended by the U. S.
the specimen into place creates a high stress
Bureau of Mines depend on the boiler oper-
in the metal surface in contact with the
ating pressure and are as follows:
block. Then, by adjustment of the clamping NaN0 3 /NaOH ratio
plate and the adjusting screw, water is up to 250 psi 0.20
allowed to leak out very slowly under the up to 400 psi 0.25
specimen with escape of steam leaving a con- up to 700 psi 0.40
centrated solution in contact with the stressed Figure 28-6 illustrates the 'before and after'
surface. Laboratory and plant tests have appearance of two test specimens exposed in

control of this method of preventing em-

brittlement, it is advisable to check on the
hydroxyl ion content of the boiler water.


A large portion of our present knowledge

of intercrystalline cracking and of the means
that can be taken to combat this condition is
due to the study of this subject at the U. S.
Bureau of Mines under the sponsorship of
the Joint Research Committee on Boiler Feed
Water Studies. Development of the Embrittle-
ment Detector by W. C. Schroeder and A. A.
Berk of the U. S. Bureau of Mines was a
great step forward in this work. Industry as a
whole is indebted to these research investi-
gators for the continuous logical develop-
ment of the basic principles involved and for
the continuous follow-up on methods of pre-
Figure 28-6 • Detector Specimen Bars venting embrittlement.
from a 300 psi Boiler
From the viewpoint of the power plant
engineer, therefore, the status of the em-
brittlement problem may be briefly sum-
a plant operating at 300 psi. The first bar marized. Embrittlement may strike unex-
cracked badly. Sodium nitrate was then em- pectedly, when we least anticipate it. It may
ployed for correction of the embrittling char- occasion costly repairs as well as serious inter-
acteristics of the boiler water. The second ruption in production. Each engineer should
bar, which exhibits no cracking, was obtained know whether his boiler water can cause
after the use of sodium nitrate. intercrystalline cracking of the boiler metal if
Coordinated pH-phosphate control pre- the required leakage and stress conditions are
vents the boiler water from developing em- also present. The embrittling characteristics
brittling characteristics by eliminating 'free'
sodium hydroxide from the boiler water.
Since it is desirable to avoid low pH in the IU
boiler water, the proper pH is provided
through the use of an agent such as trisodium
phosphate. Figure 28-7 illustrates the ap-
~ 10.5

~ -
proximate pH of solutions of trisodium phos- /
phate of various P0 4 content. To insure the ;::
absence of 'free' sodium hydroxide and the a. 10.0
absence of any alkalinity other than that of I
trisodium phosphate, it is necessary to main- 9.5
tain the pH of a boiler water below the curve.
This coordinated pH-phosphate method of 9.0 J
20 40 60 80 100 120 140 160 180 200
preventing embrittlement is primarily appli- PHOSPHATE AS PO
cable to plants employing evaporated or de-
ionized makeup water. Although the curve in Figure 28-7 • Approximate pH Values of Trisodium
Figure 28-7 can be used as a guide in the Phosphate Solutions

of the boiler water can readily be checked in REFERENCES

the plant itself through the use of the Em- A. A. Berk and W. C. Schroeder, "A Practical
brittlement Detector under the direction of a Way to Prevent Embrittlement Cracking", Trans-
supervising laboratory. If embrittling char- actions, Am. Soc. of Mech. Engineers, Vol. 65, pp.
701-711 (1943)
acteristics of the boiler water are found, there
A. A. Berk, "The Prevention of Embrittlement
are definite established methods available for Cracking", Transactions, Am. Soc. of Mech. En- .
overcoming such undesirable conditions. Ap- gineers, Vol. 73, pp. 859-864 (1951)
plication of specific chemicals or methods of J. D. Betz, "Preventing Boiler Metal Embrittle-
ment", Petroleum Refiner, Vol. 24, pp. 98-102
treatment can be tested by further embrittle- (July, 1945)
ment test runs until specimens no longer 0. H. Preis, "Embrittlement A Modern Menace",
crack in the Detector. It is then possible to Power Engineering, Vol. 64, pp. 68-69 (Apr. 1960)
establish controls over the embrittlement in- S. F. Whirl and T. E. Purcell, "Protection Against
Caustic Embrittlement by Coordinated Phosphate
hibitor and to so control the boiler water as pH Control", Proceedings, Engineers Soc. of West-
to maintain it in a nonembrittling condition. ern Penna., pp. 45-53 ( 1942)

Purpose for Blowdown. . . . . . . . . . . . . . . . . . . . 211

29 Manual Blowdown . .................... 212

Continuous Blowdown . .................. 213

Boiler Blowdown Control

Equipment Employed . ..................... 213
Manual Blowdown . .................... 213
Deconcentrators . ....................... 215
Surface Blowdown. . . . . . . . . . . . . . . . . . . . . . 215
Continuous Blowdown . .................. 215
Blowdown Control. . . . . . . . . . . . . . . . . . . . . . . . 216
Total Solids. . . . . . . . . . . . . . . . . . . . . . . . . . 216
Dissolved Solids. . . . . . . . . . . . . . . . . . . . . . . 217
Sulfate, Silica and Alkalinity . ............ 220
Chloride . ............................ 220
Specific Gravity. . . . . . . . . . . . . . . . . . . . . . . . 220
Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220

term blowdown means to eject or

drain a portion of the water from a
system. Boiler water blowdown is the removal
100 y
Since the total chlorides entering boiler must
from a boiler of some of the concentrated
equal total chlorides leaving boiler
water to be replaced with feedwater, thus
effecting a general lowering of concentra- xa = yb
tions in the boiler. 100
Multiplying both sides of equation by---
Boiler blowdown is usually expressed m equals xb
percentage and is defined as follows:
100 a 100 y
Quantity blowdown water
- - - - - - - - - ( 1 0 0 ) =%Slowdown b X
Quantity feedwater
100 y
Thus a 5% blowdown means that 5% of Since by definition--- = k, then
the water fed to the boiler is removed by
blowdown. The percent blowdown normally 100 a
k ----or
is determined by means of the chloride test. b
The formula and its algebraic derivation is Cl in feedwater
shown in Figure 29-1. X 100 = % blowdown
Cl in boiler water
PURPOSE FOR SLOWDOWN The American Boiler Manufacturer's Asso-
ciation in its standard guarantee of steam
Under normal operating conditions the only purity stipulates the following limits for
reasons for boiler blowdown are to lower the solids content of a boiler water:
suspended solids and dissolved solids content
of the boiler water. These solids are brought Operating Total Solids,
in by the feedwater and since the steam Pressure, psig ppm ppm
leaving the boiler is relatively pure, concen-
0- 300 3500 300
tration of solids will develop in the boiler 301- 450 3000 250
water. Excessive amounts of suspended solids 451- 600 2500 150
will cause deposition of sludge and may alter 601- 750 2000 100
751- 900 1500 60
the surface tension of the boiler water This 901-1000 1250 40
effect, in turn, may be severe enough to 1001-1500 1000 20
1501-2000 750 10
cause carryover in the absence of suitable 2001 and higher 500 5
anti-foam agents. Excessive amounts of dis-
solved solids have the same deleterious effects These recommended limits for solids are
as suspended solids and, in addition, when not entirely applicable to each individual
the solubility of the salts present in the boiler
case. Rather, they are limits of a general
water is exceeded, scale w:ll be deposited.
nature, providing a suitable basis from which
Fig. 29-1 to vary solids in accordance with the specific
Algebraic Proof of Slowdown Formula needs of a particular plant.
In the past few years, highly effective anti-
Let x = quantity feedwater foam agents have been developed and in
y = quantity blowdown water many instances, where they have been prop-
a = chloride concentration in feed- erly applied, excellent steam purity has been
water produced even though the total solids content
b = chloride concentration in boiler of the boiler water was maintained con-
water siderably in excess of the range stipulated by
k = percent blowdown the A.B.M.A. In fact, in many cases the
Therefore: by definition of percent blowdown effect of boiler water solids on steam purity

need no longer be considered in controlling lowering the dissolved solids concentration of

boiler water balances and therefore the sole the boiler water, it will also remove a portion
controlling factor in boiler blowdown is the of the sludge which is generally more con-
need to control sludge concentrations in the centrated in the lower part of the boiler.
boiler for the prevention of deposits on the When continuous blowdown is also in-
boiler heating surfaces. stalled, the primary purpose of manual blow-
The total solids content of the boiler water, down is to remove the suspended solids or
however, is only one factor controlling boiler sludge which may have accumulated in the
blowdown. In many cases, especially boilers area. In most cases, it is necessary to employ
operating above 600 psig, blowdown is based two or more short sludge blowdowns per day
upon limiting the silica content of the boiler in order to procure proper sludge removal.
water below certain values so as to limit the However, in cases where the boiler feed-
amount of silica vaporized with the steam. water is exceptionally pure, such as resulting
Vaporization of silica with the steam is not from an evaporator or with a high per-
usually a problem where boiler pressures are centage of returned condensate, blowdown
maintained below 400 psig. Unless silica can be employed less frequently as very little
values are properly restricted, a siliceous type sludge is formed in the boiler.
of scale may be deposited on the boiler When continuous blowdown is not in-
heating surfaces. Therefore, it is quite possible stalled, it is necessary to use the manual blow-
to find a plant operating with an apparently down to control concentrations in the boiler
high blowdown rate even though A.B.M.A. water. Whenever the dissolved solids or sus-
limits might permit higher solids and the use pended solids approach or exceed pre-
of an anti-foam might permit a very high determined limits it is necessary to use manual
total solids content before carryover occurred. blowdown to lower these concentrations. In
In addition to silica, other chemical factors practice, the valve of the bottom blowdown
which frequently limit or control the final is opened periodically in accordance with an
percentage of blowdown are alkalinity and operating schedule and/or chemical control
suspended solids. The two objectives gov- tests. From the standpoint of control, econ-
erning the boiler water limits which are omy and results, frequent short blows are
established are the production of pure steam preferred to infrequent lengthy blows. This
and the obtaining of clean boiler heat trans- is particularly true when the suspended solids
fer surfaces. content of the water is appreciable. Figure
In addition to the chemical balances main-
tained in the boiler water, there are many
mechanical factors which influence the total
solids concentrations which can be carried
safely. These include boiler design, rating,
water level and load characteristics and
whether water level and load variations are
relatively constant or subject to extremely
rapid variation.
MANUAL BLOWDOWN. There are two princi-
pal types of boiler water blowdown-inter-
mittent manual blowdown and continuous
blowdown. Manual blowdown, or sludge
blowdown, is necessary for the operation of a -----FREQUENT SHORT BLOWS
boiler regardless as to whether or not con-
tinuous blowdown is also installed. The blow-
down take-off is usually located at the lowest Figure 29-2 • Effect of Different Types of Blowdown on
part of the boiler so that in addition to Boiler Water Concentrations

29-2 illustrates a hypothetical case. With heat exchanger can be used, providing a
the use of frequent short blows, a more uni- definite savings in initial cost of equipment
form concentration of the boiler water is since a considerable portion of the heat re-
maintained and in general a smaller quantity coverable from the boiler blowdown is con-
of water is blown down because of the better tained in the amount of flash steam obtained.
control. Figure 29-3 illustrates typical calculations
Bottom blowdown, properly controlled, ac- for determining the savings involved in a
complishes the lowering of the dissolved heat recovery system utilizing a low pressure
solids and suspended solids, providing for flash tank and heat exchanger. Installation
satisfactory boiler operation. It is frequently of heat recovery equipment is of value only
undesirable, however, from an operating when the low pressure steam or heat re-
standpoint, to blow down a boiler manually covered from the blowdown water can be
with sufficient frequency to properly control utilized for increasing the feedwater tem-
concentrations. This condition is especially perature. If an excess supply of exhaust or
true if the boiler feedwater is high in solids low pressure steam is already available, natu-
content. rally, there would be no savings and instal-
lation of heat recovery equipment would not
CoNTINUous BLOWDOWN. Continuous blow-
be justified.
down as the name implies is the continuous
removal of concentrated water from the Other applications of continuous blow-
boiler. It offers many advantages not pro- down water and methods of recovering the
vided by the use of bottom blowdown alone. heat from this water are practiced. Under
Continuous blowdown may be viewed as an certain conditions, an advantage is obtained
extension of the practice of frequent short in the recirculation of boiler water blowdown
to hot process softeners. In other instances, a
blows. In actual plant practice, periodic ad-
small portion of the blowdown water is re-
justments are made in the setting of the con-
trol valve in order to increase or decrease the turned to the feedwater heater for the pur-
amount of water blown out of the boiler in pose of elevating the feedwater pH value
accordance with controlled test results, thus and thus decreasing the possibility of cor-
making it possible to maintain close control rosion in the feed lines.
of boiler water concentrations at all times.
Another important advantage of continu- EQUIPMENT EMPLOYED
ous blowdown is that heat recovery is possible.
With the use of an efficient heat exchange MANUAL BLOWDOWN. The equipment em-
unit the only heat loss is the terminal differ- ployed for manual blowdown of the boiler is
ence between the incoming cooling water considered a part of the boiler and is installed
and the blowdown water to the sewer. This along with the unit. This equipment generally
terminal difference usually amounts to ap- consists of a take-off line, a quick opening
proximately 20 F. valve and a shut-off valve. The take-off line
The use of the flash steam principle prior is always located in the lowest part of the
to a heat exchange unit further increases the boiler proper where it is anticipated the
possible savings which can be realized by greatest concentration of sludge will form
continuous blowdown. By flashing a portion and also to permit proper draining of the
of the boiler water into steam at a lower unit.
pressure, the resulting low pressure steam Water tube boilers of various types gener-
can be used quite readily for process heating ally have more than one blow-off connection,
or, as is the usual case, for feedwater heating one or more being installed so as to blow
by direct introduction to the feedwater down from the mud drum and additional
heater. With the use of a flash tank in con- connections may be made to various headers.
junction with a heat exchanger, a smaller Blowdown connections are installed on the

Figure 29-3-Typical Calculations on Savings Possible through

Use of Heat Recovery on Continuous Slowdown. Basis-1 Day
Evaporation = 5,000,000 lbs steam

Boiler pressure = 200 psig

Feedwater temperature = 200 F (live steam used)

Makeup water temperature = 60 F

Rate of blowdown = 5.0%

Boiler efficiency = 75%

Fuel·= Oil at 150,000 Btu/gal, 7.0¢/gal delivered and fired

Feedwater = Steam plus blowdown

5,000,000 lbs
Feedwater 5,263,000 lbs

Blowdown = 263,000 lbs

Employing a flash tank at 5 psig, the quantity of 5 psig steam available may be calculated
from the formula:
Hb- Hr
% flashed steam = X 100

heat of liquid at boiler pressure, Btu/lb

heat of liquid at flash pressure, Btu/lb
latent head of vaporization at flash pressure, Btu/lb
362- 196
% flashed steam = X 100 = 17.3

Flashed steam available at 5 psig = 263,000 lbs X 0.173 = 45,500 lbs

Total heat of flashed steam at 5 psig = 1,156 Btu/lb

Heat savings in flashed steam = 1,156 Btuflb - 28 Btu/lb X 45,500 lbs = 51,324,000 Btu

The drain water from the flash tank is passed through the heat exchanger and thence to
the sewer. It is assumed that the temperature of the water leaving the exchanger is 20 F
above the incoming makeup water or 80 F.
Heat of liquid at 80 F = 48 Btuflb
Heat of liquid at 5 psig =.196 Btu/lb
Heat recovery = 148 Btu/lb

Heat savings from heat exchanger = 263,000 lbs X 0.827 X 148 Btujlb 32,190,000 Btu

Total heat savings = 51,324,000 Btu plus 32,190,000 Btu = 83,514,000 Btu

At a boiler efficiency of 75% there is actually utilized 150,000 Btu/gal X 0.75 120,000
Btu/gal oil.

83,514,000 Btu
Fuel savings 696 gal
120,000 Btu/gal oil

Daily savings = 696 gal X 7¢/ gal oil = $48.72

headers for the chief purpose of blowing similar device in an attempt to make the
down these sections of the boiler in order to installation more effective. A further refine-
remove suspended solids which may accu- ment of this idea involves the application of
mulate at such points and cause a restriction a floating trough which is designed to re-
of circulation. Usually, the boiler manu- move water at a point close to the boiling
facturer will stipulate certain restrictions in surface regardless of any variations in water
the blowdown of these water wall headers level in the boiler. Skimming devices, in gen-
and, in all instances, the recommendations of eral, have been more commonly applied in
the manufacturer should be followed. marine practice and have not found wide
DECONCENTRATORS. A deconcentrator is a application in stationary plants.
cylindrical tank so connected to the boiler as CoNTINuous BLOWDOWN. Continuous blow-
to receive concentrated boiler water. By down equipment is usually considered aux-
means of special baffiing, a settling of the iliary boiler equipment although many manu-
suspended solids takes place and a clarified facturers now install the internal collection
water is returned to the boiler. Periodic blow- pipe or at least will make provision for such
down of the deconcentrator results in the an installation during the course of manu"
blowdown of water containing a higher per- facture. The exact location for the continuous
centage of suspended solids as compared to blowdown take-off line within the boiler de-
the normal boiler water. A deconcentrator in pends primarily upon the water circulation
no way affects the dissolved solids or alka- circuits which are involved. The take-off line
linity content of the boiler water. Theoreti- is installed in such a manner as to assure the
cally, deconcentrators should be of value in removal of the most concentrated water from
cases where excessive suspended solids or the boiler. It is essential also that possible
sludge develops. In actual plant practice, short-circuiting of boiler feedwater or chemi-
however, they can be recommended only on cal feed solution into the take-off line be
rare occasions since installation of softening avoided. The size of the lines and control
equipment can be more readily justified in valve depends upon the feedwater character-
most such instances. istics, quantity of blowdown required and the
SURFACE BLOWDOWN. Connections are in- boiler rating. Figures 29-4 and 29-5 illustrate
stalled occasionally with the take-off line the method of installation in typical boilers.
located slightly below the working water level In most boilers, the take-off line is located
for the purpose of skimming sediment and approximately two inches below low water
oil from the surface of the water and the level although in certain boiler designs, the
name, surface blowdown, is generally applied take-off line is located close to the bottom of
to such an installation. The end of the take- the drum at a point where the most con-
off line may be equipped with a funnel or centrated water is found.

2nd pau




CONTINUOUS BLOWDOWN:-Take-off pipe should be _ _ , extra heavy iron pipe lo-

cated under cyclone separators in 1st gas pass and at least 2" below low water level. End
of pipe should be capped and __ , hole drilled therein. On top side of pipe, _ _ equally
spaced __ , holes should be drilled. External to drum, install globe valve and reduce line
to _ _ " pipe. Install forged steel _ _ " _ _ - V port flow control valve similar to the Han-
cock Flocontrol at convenient location.
CHEMICAL FEED: -Distribution line should be _ _ , extra heavy iron pipe installed be-
neath cyclone separators in 3rd gas pass and at least 2" below low water level. Line
should remain open ended, discharging at point between 2nd and 3rd gas passes, as indi-
NOTE:-Use iron or steel valves and piping on all continuous blowdown and chemical feed
Figure 29-4 • Continuous Blowdown and Chemical Feed forB & W Type F and FH Boilers

BLOWDOWN CONTROL blowdown rate, it is essential that suitable

boiler water tests be conducted in order to
It is important that blowdown be such that obtain a constant check on concentrations
boiler water concentrations do not exceed developing in the boiler water. Regardless of
certain limits. Likewise, it is equally impor- the factor limiting the boiler water concen-
tant that excessive blowdown be avoided be- trations, the control of boiler blowdown is
cause of the heat units that are wasted as usually based on the determination of total
well as the additional chemical treatment re- solids content of the boiler water, determined
quired to react with the additional incoming gravimetrically, the dissolved solids as meas-
feedwater. Therefore, it is essential that a ured by specific conductance, the chloride,
system of control be established so as to main- sulfate, silica or alkalinity content of the
tain blowdown within determined limits. In boiler water or any combination of these
order to illustrate the savings which may be tests.
obtained by proper blowdown control, Figure ToTAL SoLIDS. Determination of the total
29-6 shows the amount of heat saved when solids content of the boiler water provides an
assuming a reduction in blowdown from 6o/o excellent control over blowdown. While usu-
to 2%. In order to maintain an economical ally applicable from a technical standpoint,






Globe Valve

Reducer _,.


CONTINUOUS BLOWDOWN: Take-off line should be __ inch pipe located at least 2

inches below low water level in top, rear drum as indicated. Cap end of take-off and drill
with __ inch hole. In top side of the take-off, __ , _ _ inch holes spaced equally should
be drilled within length A- A'. Immediately external to the drum, install globe valve and re-
duce line to __ inch. Insert __ inch, forged steel, _ _ - V port flow control valve at an
easily accessible location.
CHEMICAL FEED: Distribution line should be __ inch pipe discharging open ended
approximately 2 inches below low water level and external to trough if provided. With tee
arrangement as in illustration "A", the open ends should discharge at third points of drum.
NOTE: Use extra heavy iron or steel piping and fittings for all continuous blowdown and
chemical feed lines.
Figure 29-5 • Continuous Blowdown and Chemical Feed for Three-drum, Bent-tube Boiler

it is rarely employed as the analysis requires carbonates liberating carbon dioxide gas, etc.
much time and is too difficult for a routine DisSOLVED SoLIDS. Determination of the
control. There are objections to the use of specific conductance of a boiler water will
the total solids test for calculation of the provide an accurate measure of the dissolved
rate of blowdown. A comparison of the total solids content and usually can be applied to
solids content of the boiler water with the blowdown control with considerable advan-
total solids content of the boiler feedwater tage. However, calculation of the rate of
will not necessarily provide an accurate blowdown on the basis of the relative specific
measure of the concentration that has taken conductance of the feedwater and boiler
place within the boiler due to the boiler water is not technically correct. Not only will
water samples not showing a representative specific conductance be affected by loss of
suspended solids content, effect of internal carbon dioxide with the steam and the solids
treatment, breakdown of bicarbonates and introduced by internal chemical treatment,

Figure 29-6-Typical Calculations on Sav- but the specific conductance of a dilute solu-
ings Possible Through Reduction in Blow- tion such as feedwater and a concentrated
down. Basis-1 Day solution such as boiler water cannot be com-
pared directly. The specific conductance of a
Evaporation . . . . . = 2,000,000 lbs. steam sample is caused by ionization of the various
Boiler Pressure . . . = 200 psig gage salts present in the sample. In dilute solu-
Feedwater Temper- tions, ionization increases and greater specific
ature . . . . . . . . . . =215 F conductance results. In more concentrated
Makeup Water solutions, ionization is repressed and specific
Temperature . . . = 60 F conductance decreases per part of salt pres-
Present Blowdown ent. This effect is illustrated by the fact that
Rate .......... =6% the solids content of very dilute solutions
Reduced Blowdown such as condensate may be calculated by
Rate .......... =2% using a factor of 0.5 to 0.6 ppm of dissolved
solids per micromho of specific conductance.
Boiier Efficiency . . = 80% In more concentrated solutions, such as boiler
Fuel = Oil at 150,- water, the factor will vary between the limits
000 Btu/gal., of 0.55 and 0.9 ppm of dissolved solids per
7.0¢/ gal. delivered micromho of specific conductance. If gallic
and fired
Feedwater . . . . . . . =Steam plus blowdown
Feedwater at 6.0%
This chart is used to calculate the percent
Blowdown . . . . . = 2,128,000 lbs.
of boiler water discharged by a continuous
Feedwater at 2.0% blowdown system that can be flashed into
Blowdown =2,041,000 lbs. steam at a reduced pressure and recoverable
Saving m Feed- as low pressure steam for heating or process.
water . . . . . . . . . = 87,000 lbs. ExAMPLE: A boiler operates at a pressure of
Heat of Liquid at 450 psig. Continuous blowdown amounts to
Boiler Pressure, 10,000 pounds per hour. What percentage of
Btu/lb. .. .. .. .. =362 blowdown water can be recovered as flashed
Heat of Liquid at steam at 5 psig?
60 F, Btu/lb. . . . =28 SoLUTION: Locate 450 psig on lefthand axis.
Heat Saving, Follow horizontally toward the right to the
Btu/lb. . . . . . . . . = 334 intersection with 5 psig "flash" curve (point
Total Heat Saving A) . Drop vertically downward to the bottom
= 87,000 lbs. x axis and read 25.5%. (25.5% of 10,000
334 Btu/lb. . . . . = 29,058,000 Btu pounds per hour blowdown = 2550 pounds
per hour of flash steam at 5 psig pressure.)
At Boiler Efficiency
of 80% There is These curves have been prepared from the
Actually Utilized formula:
150,000 Btujgal. He- HF
Percent flashed steam - - - X 100 where
x 0.80 . . . . . . . . . = 120,000 Btu/gal. VF
HB = heat of liquid at boiler pressure m
29,058,000 Btu Btu/lb
Fuel Savings 242 gal.
120,000 Btu/gal. HF = heat of liquid at flash pressure m
Daily Savings = 242 gal. oil x $0.07/ gal. V F = latent heat of vaporization at flash
=$16.94 pressure in Btujlb


1\ 'I\ I I \ II if 'I I I
I\ \ \ II I\ II Ill/ 1/ 1/ II
\ 1 1\ 1 \ ~ I I II I I I
700 ~~~~~~++~~+1\wi\~I~X
I~ \ 1\ I 1\ J JII II
\ \ \ I\ I II! II \ I II I I I
' 1 I~- - ft ~fll-· -11- ·f'r-
POUNDS PER SQUARE IIKH 6.1/U6E 1\ I~ 1\ l '\ Ill I I 1300 g
g 6oo \ 1J 1 \ I/ I 1\ 1 11 ~
~ 1/ 11 ' 1\ '1]}1111 J
1~ '
\/1\ 1\ 1\ \
V\1\ 111
t\ Jlii!IV1 l' lii

soo' H-l--1--+-+-1-+--+-+-+-+-+-+IH-1

1/ ~
\lf.l/.f-ltr\Vh4-'-vlf-HI--t-+-t-t-H 1200 ':'
,1\1\ ~
IY 1' A 1/1\,
:g 1-· - -I- --- +· -1--- - - rt _, f-1 -, -t-. ~~if If I'V 1\ ~
~ J7 ~ '1\ 1/ I IJ\ :0

~ J.J..--+\I\r+l\¥1r\-\+\-l\-+,\A!,'Jlf-JA/IY-III/+1JI
40oH-l-t-++-l-+-+-+-i+J+-+-+I+- f-1 H-+-+1\-\t-t-t-t--H--t 1100 ~
~ 11 I\ lA ' \ OC . 1 \ ~· ~
~ II J ) lllJ\ ~ v II I I I\ z
8: I II 1\ 1\ V\ IV l~ I K I I\ £
~ V II \lXII '\r\ I/ \d 11 ~
g 300~4-~+-+~V~~~~~r-t-~'+-++A~~VAV~~~I~v~~'-r+\~-M'\~-t-~looo ~
I/ I II Vl"il~l\ ~~ \
J 1/ v ~ ~\ ~ 1\

\ 1\

\.1-j.U 800
0 5 10 15 20 30 3S

Figure 29-7
Flashed Steam Recoverable from Continuous Blowdown Systems
NOTE: For boiler pressures between 100 ond 800 psig, use "flash" pressure curves slanting from lower
lefthand to upper righthand corner and the bottom axis. For boiler pressures above 800 psig, use
"flash" pressure curves slanting from lower righthand to upper lefthand corner and top axis.

acid is employed to neutralize the phenol- content will cause an appreciable error in the
phthalein alkalinity, the specific conductance calculated rate of blowdown.
multiplied by the factor 0.9 will give a very SPECIFIC GRAVITY. The specific gravity of a
close approximation of the actual dissolved boiler water is proportional to the dissolved
solids content of the boiler water. solids. However, the accuracy in determining
SuLFATE, SILICA AND ALKALINITY. Under dissolved solids by means of a hydrometer
certain circumstances, the sulfate, silica and measuring the specific gravity of a boiler
alkalinity content of the boiler water can be water is so poor that it cannot be recom-
used for blowdown control where they are mended for proper blowdown control.
factors which, in that particular case, limit
the boiler water concentrations that may be CONCLUSION
safely tolerated. Usually, these determina-
tions are not suitable for calculation of the It is to the benefit of any plant to maintain
rate of blowdown since both sulfate and boiler water concentrations consistently at
silica, may be removed from solution as scale. the optimum value. This control is necessary
The use of sodium sulfite will increase the to assure the production of pure steam and
sulfate content of the boiler water and alka- maintain heat transfer surfaces free of scale
linity is usually affected by either variable and other objectionable deposits. The eco-
softener balances or by the internal treat- nomics of blowdown requires careful study in
ment employed. All these conditions prevent order to make certain that blowdown is es-
accurate determination of the relative feed- tablished upon a proper basis, taking into
water and boiler water concentrations which account all chemical and mechanical factors
is the basis for calculation of the rate of affecting the problem. Material savings in
blowdown. fuel and water consumption may be made
CHLORIDE. The chloride determination can either by the application of proper chemical
be used not only for blowdown control, but treatment methods or by the installation of
it is also suitable for the accurate calculation heat recovery equipment. The entire subject
of the rate of blowdown provided the chlo- of blowdown is of sufficient importance to
ride concentration of the feedwater is suffi- justify in every case a suitable study of all
cient to permit accurate measurement. Since factors pertaining to it.
chloride is not subject to precipitation in the
boiler water, the determination of the rela-
tive chloride concentrations of the feedwater 0. de Lorenzi, "Combustion Engineering", pp.
and boiler water provides the most accurate 21-17 & 18, Combustion Engineering-Superheat-
basis for calculation of the rate of blowdown. ers, Inc., New York, N.Y. (1949)
W. L. Nieland, "Control and Economics of Boiler
Therefore, the percent blowdown can be Blowdown", Combustion, Vol. 16, pp. 34-38 (Oct.
calculated substituting the determined chlo- 1944)
ride values in the formula as shown in Fig- W. L. Nieland and J. H. Richards, "Boiler Blow-
ure 29-1. down Control", Industry and Power, Vol. 57, pp.
64-68 (Mar. 1954)
The chloride test is unsuitable for this cal- R. T. Sheen, "Continuous Blowdown on Boilers",
culation only in those cases where the feed- Power Plant Engineering, Vol. 49, pp. 90-93 (Jan.
water chloride is quite low, as in the range
C. D. Shields, "Boilers: Types, Characteristics,
of 0.5-1.0 ppm. In this range a slight analyti- and Functions", pp. 227-244,. F. W. Dodge Corp.,
cal error in determining feedwater chloride New York, N.Y. (1961)

Chemical Cleaning of Boilers

Alkaline Boil-out. . . . . . . . . . . . . . . . . . . . . . . . 222

Inhibited Acid Cleaning. . . . . . . . . . . . . . . . . . . . 224
Hydrochloric Acid . ..................... 224
Phosphoric Acid. . . . . . . . . . . . . . . . . . . . . . . 226
Removal of Copper Deposits . . . . . . . . . . . . . . . . 226
0 ther Acids. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227
Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 228

0 NE of the most serious problems encoun-

tered in steam generation is the forma-
tion of scale and sludge deposits on boiler
sate from steam engines or pumps. It may
also originate from steam or water used in
certain processes. Even thin films of oil on
heating surfaces. The presence of oil on boiler boiler heating surfaces may seriously interfere
heating surfaces may also give rise to serious with normal heat transfer and the resultant
difficulties. In high pressure generating sta- increase in metal temperature may be suffi-
tions boiler deposits containing metallic cop- cient to cause blistering and loss of boiler
per and copper oxide have been experienced. tubes and plates. ·
This condition may result in boiler tube fail- For the removal of oil and grease from
ures. In addition to mill scale, new boilers boiler heating surfaces an alkaline boil-out is
also contain deposits of oil, grease and pro- used. With new boilers the alkaline boil-out
tective coatings. It is important that the removes the protective coatings applied dur-
proper steps be taken to remove deposits from ing erection together with oil, grease and any
the boiler heating surfaces. Removal of de- dirt that may have accumulated during erec-
posits may be effected by one or a combina- tion. The boil-out should be so formulated to
tion 'of chemical methods, depending on the provide the greatest detergent, emulsifying
particular nature of the deposit. and dispersive action on the greases and oils
normally found in new boilers. With boilers
ALKALINE BOIL·OUT that have been in operation, the boil-out
should exert the same effect on oil present as
The presence of oil in the boiler feedwater a result of feedwater contamination.
originates from the exhaust steam or conden- While it is not possible to obtain maximum

Figure 30-1 • Thin Scale Particles Cemented into a Dangerous Mass


emulsifying and dispersive action by the use rate in accordance with recommendations of
of inorganic chemicals alone, readily avail- the manufacturer. When steam is flowing
able inorganic chemicals may be used when freely from the vents, all vents should be
a specially prepared boil-out mixture cannot closed except those specified by the manu-
be obtained. One of the most satisfactory facturer on boilers containing superheaters,
combinations of inorganic chemicals includes where superheater protection is required.
3 pounds of caustic soda, 3 pounds of diso- The pressure 'is then raised to a maximum
dium phosphate and 1 pound of sodium ni- of 25 psi and maintained for approximately
trate for each 1000 pounds of water required 48 hours. For high pressure boilers it is usual
to fill the boiler to operating level. Because to boil-out at one-half the operating pressure
of the high alkalinity developed during the to obtain circulation. In all cases the recom-
boil-out, the sodium nitrate is included in mendations of the manufacturer should be
the combination as a precaution against caus- followed regarding boil-out pressure. The
tic metal embrittlement. boiler should be blown down one-half gauge
Prior to boiling-out a boiler, it is advisable glass approximately every eight hours through
to replace the normal high pressure gauge the bottom blow-off valve. Where there is
glasses with temporary glasses to prevent at- more than one blow-off connection, alternate
tack on the glass. After addition of the re- valves should be used. After each blowdown
quired quantity of boil-out material, all man- the boiler should be refilled to the top of the
holes should be closed and the boiler filled to glass.
the top of the gauge glass with the vents After completion of the boil-out the unit
open. The boiler should be fired at a low is cooled slowly, drained and flushed thoroughly

Figure 30-2 • This Superheater Tube Showing Carryover Deposits is an Application for Chemical Cleaning

with a high pressure hose. If inspection sists of black magnetic iron oxide. The pres-
should indicate that any scum or oil remains, ence in the boiler water of the black iron
the boil-out procedure should be repeated. oxide suspension may cause. a new boiler to
Some loose mill scale may be removed from run "black water" for several weeks under
a new boiler during the boil-out process. some conditions. These iron oxide deposits
However, removal of mill scale is not the also may accumulate to a point where they
primary objective of the boil-out since acid impede circulation and interfere with heat
cleaning is required for effective removal of transfer. They may also cause obstructions
mill scale. in steam separators and dryers, possibly af-
fecting the flow of steam in the drums. A
INHIBITED ACID CLEANING carryover condition may be aggravated,
causing iron oxide to deposit in superheaters,
The principal cause of scale formation on strainers and turbine blading. In the presence
bbiler heating surfaces is the fact that the of the alkaline boiler water, concentrated cor-
solubility of scale forming salts decreases with rosion can occur at breaks in the mill scale
an increase in temperature. When the feed layer and tube failures may result due to
water temperature is elevated to boiler op- these pits. Before acid cleaning, a new boiler
erating temperature, the solubility of scale should be given an alkaline boil-out so as to
forming salts is exceeded and crystallization · remove oil, grease, protective coatings and
takes place on the boiler heating surfaces. any dirt that may have accumulated on these
The use of inhibited acid for the removal of surfaces during erection.
scale from boiler and metal surfaces has been Prior to acid cleaning the usual protective
practiced for many years. Within the past measures should be taken such as replace-
few years, however, there has been a con- ment of any brass or bronze parts temporarily
siderable development in the technique and with steel or steel alloy equipment. Steps
a number of new inhibitors are now avail- should be taken to provide for proper vent-
able. Acid cleaning requires no great dis- ing of acid vapors and any valves connecting
mantling operation and labor cost is held to the boiler with the main steam header should
a minimum. Scale is removed from all sur- be closed.
faces reached by the acid solution and the HYDROCHLORIC Acm. Inhibited hydrochloric
cleaning, in most cases, is more complete acid is most widely used in acid cleaning be-
than could be obtained by mechanical means. cause of its relatively low cost and the num-
By acid cleaning the internal metal sur- ber of satisfactory inhibitors presently avail-
faces of a boiler are left metal clean and as able. Two general methods used in acid
such the true condition of the metal is re- cleaning of boilers are the "circulation"
vealed. Opportunity is provided for more method and the "fill and soak" method. In
complete inspection of questionable areas the first method, the cleaning solution is
within the boiler tubes, headers and drums. circulated continuously through the unit un-
In many cases pits and pitted areas may go til the cleaning action is completed, as shown
unnoticed until inspected after acid clean- by tests on the effluent. This method is suit-
ing. Cleaning down to the metal surface will able for units with positive liquid flow paths
help reveal any stress areas and cracks which and offers the advantage of regular checking
may be due to fatigue, embrittlement or during the cleaning operation. In the second
corrosion. method the cleaning solution is permitted to
Acid cleaning of new boilers is considered soak in the unit for a prescribed length of
desirable by some for the removal of mill time. This method is used where positive
scale. Mill scale is the term generally applied circulation of the cleaning solution cannot
to the oxide film produced when iron and be readily obtained.
steel surfaces are heated and cooled during In the "circulation" method the unit is
manufacture and fabrication. Mill scale con- filled to overflow with clean warm water at

the maximum temperature permitted for the metal surfaces from corrosion. This protec-
particular inhibitor used. This water is circu- tion can be accomplished by employing either
lated until a constant temperature is obtained the wet or dry lay-up procedures.
on the effluent which is returned from the In the "fill and soak" method inhibited
highest point of the unit to a tank equipped acid is added in the proper ratio with warm
with a clean water line, steam line and drain. water or condensate (at the maximum tem-
For circulation, a corrosion resistant pump perature permitted for the inhibitor) to give
should be employed. Inhibited acid is added a cleaning solution of the desired acid strength
in sufficient quantity to provide an acid solu- and the unit is filled to overflow. The vent is
tion of proper strength entering the unit and opened and soaking is carried out for a pre-
this feed is continued until tests on the efflu- determined length of time, depending upon
ent solvent indicate a maximum acid content the type and estimated quantity of deposit
has been reached. The concentration of acid and on the rate of reaction between the de-
used will generally vary from 3 to 6%, de- posit and solvent, as indicated by laboratory
pending upon the particular conditions. tests.
Should continued testing of the effluent sol- At frequent intervals samples can be drawn
vent reveal a 1% drop in acid content, suffi- from available points and tested for acid
cient inhibited acid should be added to main- strength. However, these samples will not be
tain the desired acid content of the solvent indicative of conditions throughout the unit
influent. Circulation of the solvent through although they will give an indication as to
the unit continues until the acid content of the amount of deposit removed. Vent gases
the effluent is constant, thus indicating the may contain hydrogen and therefore open
end of the reaction between the solvent and flame is not permitted in the venting area.
the boiler deposit. With new boilers the soaking period, using
Following acid cleaning, the boiler is this method, may be approximately six hours
drained and flushed with clean warm water although treatment may be discontinued after
until tests on the returned flushing water a shorter period, depending upon control
show it to free of acid and soluble iron salts. tests.
Flushing is followed by a neutralizing or At the end of the acid cleaning period the
metal conditioning boil-out. Soda ash is gen- unit is flushed with clean warm water as
erally used for this purpose although triso- described previously. The neutralizing solu-
dium phosphate, sodium tripolyphosphate, tion is added and the unit is permitted to
sodium chromate, caustic soda or other chem- stand for a short period of time. The level
icals may be used. The alkaline boil-out can then be lowered to normal and the unit
serves to neutralize any residual acid, re- fired as described under the "circulation"
leases from the metal surface the hydrogen method. Following this neutralizing boil-out
generated during the cleaning procedure, and the unit should be drained and flushed, as
provides a slightly protective film on the previously described. Also, if the boiler is to
metal surface. To aid in neutralization the stand idle for more than a few days, proper
unit can be fired at 50 psi or somewhat steps should be taken to protect the metal
higher after circulation of the neutralizing surfaces.
solution has been stopped and the water level In addition to the properly inhibited acid
dropped to normal. During this period the it may be desirable under some conditions to
unit should be vented on occasion to permit include certain inorganic chemicals which
escape of the liberated gases. will aid in the removal of one particular
At the conclusion of this boil-out the unit constituent of the scale. The use of the proper
is drained and flushed with clean warm wetting agent along with the inhibited acid
water. If the boiler is to stand idle for more serves to reduce the rate of corrosive attack
than a few days before being returned to on the metal during cleaning. The lower the
service, steps should be taken to protect the degree of wetting of the metal by the inhib-

ited acid, the greater is the effect of the wet- REMOVAL OF COPPER DEPOSITS
t~ng agent in reducing the amount of corro-
siOn. The increased use of copper .and copper alloy
Sulfuric acid and nitric acid may also be equipment in the feedwater cycle has given
used in chemical cleaning. Sulfuric acid is rise to a new problem of boiler deposits in
economical and may be easily inhibited. high pressure generating stations that have
However, it does form insoluble salts, such low makeup water requirements. Deposits
as calcium sulfate, during the cleaning proc- have been composed principally of iron
ess. Although nitric acid is relatively inex- oxide, metallic copper and copper oxide.
pensive, there are only a few suitable inhibi- While acid cleaning is the answer to the
tors available. Both sulfuric acid and nitric problem of removing iron oxides, difficulties
acid are dangerous to handle and therefore may arise due to acid cleaning when copper
extreme care must be taken when these acids or copper oxides are present in the sludge.
are used. Although metallic copper is relatively in-
soluble in hydrochloric acid, it tends to go
PHOSPHORIC Acm. Inhibited phosphoric acid into solution during acid cleaning due to the
has been used successfully for removal of presence of ferric ion in the acid solution.
mill scale from new boilers. This agent pos- Copper can then replate on the boiler sur-
sesses certain advantages over inhibited hy- faces and it may adhere to these surfaces.
drochloric acid when used for chemical
cleaning. Inhibited phosphoric acid can be When the boiler is returned to service, the
boiled in the unit by direct firing of the boiler copper may be loosened from the boiler sur-
with neglible attack on the boiler metal and faces and deposit in areas where circulation
it does not create a problem of noxious or is less rapid. Under these conditions normal
corrosive fumes. Natural circulation resulting circulation may be affected and over-heating
from this direct firing promotes good distribu- may occur, resulting in tube failures. While
tion of the cleaning solution. Boiler surfaces copper deposits may be removed mechani-
cleaned with phosphoric acid are resistant to cally, this is a long, expensive process and as
corrosion whereas metal surfaces freshly such would nullify any benefits gained in
cleaned with inhibited hydrochloric acid re- acid cleaning. As in any mechanical process
act with the atmosphere to form rust if not there are locations that cannot be cleaned
properly protected during draining, and es- mechanically and thus deposits not removed
pecially if left in a moist condition, such as would be potential sources of difficulty dur-
before neutralization. ing later operation.

Prior to cleaning with inhibited phosphoric Various chemicals have been used for re-
acid, a new boiler should be given an alka- moval of copper deposits from boilers. In
line bbil-out, as previously described. A 5% some cleaning operations, the solvents were
inhibited phosphoric acid solution boiling at ammonia base oxidizing agents and the ef-
atmospheric pressure has been successfully fectiveness of these solvents is based on the
used in the removal of mill scale from new reaction of the oxidizing agent with metallic
boilers. Following the cleaning, some authori- copper to form cupric oxide. This oxide is
ties recommend that an acid rinse be em- dissolved in ammoniacal solution to form the
ployed for removal from the boiler surfaces familar cupric-ammonium complex ion. This
any iron that may have deposited upon series of reactions is reversible and it is neces-
draining the acid cleaning solution. A 0.1% sary that certain fundamental conditions be
solution of phosphoric acid may be used as maintained. During the treatment period ex-
a rinse. The final step in the cleaning pro- cess oxidizing agent must be available in the
cedure consists of an alkaline treatment for solvent and the ammonia present must be
purpose of neutralizing the remaining acid maintained at a uniformly high level of con-
and for conditioning of the metal surfaces. centration. The cupric-ammonium ion con-

centration must be uniform throughout the the principal oxidizing agent. The solvent is
solvent. circulated continuously through the boiler at
The first solvents used for removal of cop- a temperature of 140 F to 160 F. While this
per from boilers were those containing chlo- type of solvent dissolves copper at a less
rate as the major oxidizing component. One rapid rate than solvents containing persul-
of this type of solvent, made up of sodium fate, it is faster. acting than the chlorate type
chlorate, ammonia and ammonium sulfate, solvents. During the cleaning action analyti-
must be used at temperatures in excess of cal control can be set up to measure rate of
175 F in order to rapidly oxidize and dis- oxidant consumption and quantity of dis-
solve copper. Another type solvent, contain- solved copper in the solvent.
ing sodium chlorate, slight amounts of potas- More recently, ammoniated citric acid has
sium bromate, ammonia and ammonium ni- been employed for removal of water side de-
trate, must be used at temperatures between posits containing high percentages of copper
140 F and 190 F. With this solvent the bro- oxide. Also, complexing agents specific for
mate reacts first and limited amounts of copper have been used in an acid medium
chlorate react depending on the temperature for single stage cleaning of metal surfaces
maintained. containing significant quantities of copper
With the use of these solvents the boiler oxide in a multi-component deposit.
is generally fired intermittently to obtain re-
quired temperature and to induce circulation. OTHER ACIDS
Certain disadvantages are connected with the
use of these solvents. Where over-heating In the past few years, sulfamic acid
occurs at localized areas of the boiler, in- (NH 2 S03H) has come into use for chemical
creased attack on the boiler metal by the cleaning of equipment. This material, blended
oxidizing agent will occur. In over-head with an inhibitor, is available in powder form
areas, the ammonia content may be depleted and becomes active as an acid cleaner when
causing copper already taken into solution to placed in solution. In the cleaning operation
replate on the boiler metal. In the cleaning the solution of inhibited sulfamic acid is
process, any ammonia driven off into the air handled in a manner similar to that described
space in the boiler drums decreases the con- previously. Inhibited sulfamic acid has an
centration of the solvent, rendering it less advantage over hydrochloric acid inasmuch
effective. as it is easier to handle, does not produce
Another type of solvent contains ammo- noxious fumes on dissolving and is less corro-
sive than inhibited hydrochloric acid, particu-
nium persulfate, ammonia and sodium hy-
droxide. This solvent is highly reactive with larly at higher concentrations and tempera-
tures. Sulfamic acid, while more expensive
metallic copper even at room temperatures.
than either hydrochloric or sulfuric, is ef-
The solvent is recirculated in the boiler for
fective in removing calcium carbonate scale
a short period, then the unit is drained and
as well as iron oxide and organic deposits.
flushed with clear water. While the solvent
However, its rate of reaction is slow against
may be used at room temperature there are
calcium sulfate. In chemical cleaning, sul-
some disadvantages inasmuch as it reacts
famic acid has been used primarily for small
rapidly with iron as well as copper. As such,
boilers and other plant equipment such as
accurate chemical control of both oxidant
condensers and compressors.
consumption and quantity of copper dissolved
is difficult. Citric acid (C6Hs0 7 ) may also be used in
Another type of solvent that has been used chemical cleaning of power plant equipment.
s~ccessfully contains potassium bromate, Recirculation of a 3% solution of inhibited
shght amounts of sodium chlorate, ammonia citric acid at elevated temperature has been
and ammonium carbonate. The bromate is successful in removal of deposits from boilers

and other power plant equipment. The ease conditioning practices.

of removal depends on the composition and Although the technique of acid cleaning
quantity of deposit. The addition of a wetting has progressed to a point where there is little
agent to the cleaning solution may hasten danger from corrosion during the cleaning
removal of deposits. Where a boiler deposit process, it is advisable that the procedure
is found to be particularly resistant, the addi- be performed under the supervision of ex-
tion of ammonia to the citric acid solution in perienced personnel. During acid cleaning
sufficient quantity to elevate the pH, tends to every safety precaution must be exercised.
speed removal. In recent years, citric acid has The ever present danger of hydrogen ex-
been used successfully in cleaning of stain- plosion exists because hydrogen is evolved in
less steel superheater tubes of utility boilers. the reaction of acid with iron. In pockets
Citric acid may be used for removal of where hydrogen mixes with air, any spark or
mill scale from new boilers. Prior to use of open flame will ignite the mixture. Other
the acid the boiler should be given an alka- gases may also be generated during the clean-
·Iine boil-out to remove oil, grease, protective ing process, such as hydrogen sulfide. There-
coatings and any dirt accumulated during fore, adequate venting is essential.
erection. Following draining and flushing of Personnel handling any acid as well as
the boiler a 3% inhibited citric acid solution most inhibitors, should be fitted with proper
at elevated temperature is added and circu- safety clothing, including protective goggles
lated throughout, until all mill scale is re- and rubber gloves. Protective masks are also
moved. Periodically the iron content of the necessary for some vapors produce a suffo-
cleaning solution is checked to determine ef- cating effect and irritate the mucous mem-,
ficiency of the cleaning action. branes.

S. Alfano, "Cleaning The Iron System Using Am-
Deposits of oil and grease as well as boiler moniated Citric Acid", Combustion, Vol. 33, pp.
scale, iron oxide and copper deposits may be 17-24 (11ay 1962)
removed from the boiler metal surfaces by P. H. Cardwell and L. H. Eilers, "Use of Wetting
Agents in Conjunction with Acid Inhibitors", Ind.
chemical cleaning. When employing any of and Eng. Chern., Vol. 40, pp. 1951-1956 (1948)
the chemical cleaning methods the proper "Chemical Cleaning of Internal Heating Sur-
technique must be followed and the neces- faces", Bulletin G-39A, The Babcock & Wilcox
Company, New York, N.Y. (1946)
sary precautions taken. With regard to boiler C. 11. Lqucks, "Chemistry Tackles Plant 11ainte-
scale, it should be recognized that acid clean- nance", Chemical Engineering, Vol. 69, pp. 103-
ing is not a substitute for proper feedwater 120 (11ar. 5, 1962)

Cooling Water Treatlllent:
Once-Through Scale Control

Scale Formation. . . . . . . . . . . . . . . . . . . . . . . . . 230

Prevention of Calcium Carbonate Scale . . . . . . . . 232
Iron Deposits . . . . . . . . . . . . . . . . . . . . . . . . . . . 235

HE term "cooling water" is applied "scale" applies to deposits which result from
T wherever water is circulated through
equipment to absorb and carry away heat.
crystallization or precipitation of salts from
solution. Wasting away of a, metal is the re-
This definition includes air conditioning sys- sult of corrosion. While a deposit results in
tems, engine jacket systems, refrigeration both cases, the mechanism of formation is
systems as well as the multitude of industrial different and different corrective methods are
heat exchange operations, such as found in required to prevent the deposit.
oil refineries, chemical plants, steel mills, etc. Calcium carbonate is usually the chief in-
The once-through system, as the name gredient of scale formed in once-through
implies, is one in which the water is passed cooling water systems. Calcium carbonate is
through the heat exchange equipment and more insoluble than calcium silicate and
the cooling water is then discharged to waste. calcium sulfate and this fact accounts for its
Usually, a once-through system is employed more frequent appearance in the form of
only where water at suitably low temperature scale. Figure 31-1 illustrates analyses of sev-
is. readily available in large volume and at eral typical scales formed in once-through
low cost. The usual source of once-through cooling water systems. It can be noted that
cooling water is from wells, rivers and lakes the combination of calcium and carbonate
where the cost involved is that of pumping accounts for the major portion of each de-
only. In a once-through system, no evapora- posit.
tion takes place and consequently the water The factors which affect scale formation
does not concentrate. Circulating water char- are temperature, rate of heat transfer, the
acteristics are the same as the makeup water. calcium, sulfate, magnesium, silica, alkalinity,
dissolved solids and pH of the water. In a
once-through system, the magnesium, silica
and sulfate concentrations are normally suf-
Deposits in lines, heat exchange equipment, ficiently low that these factors do not enter
etc., may originate from several causes. For into the problem.
example, the precipitation of calcium carbon- In 1936, the work of Prof. Langelier was
ate will form scale, but products of corrosion published dealing with the conditions at
also result in a deposit of iron oxide. In speak- which a given water is in equilibrium with
ing of deposits which form in cooling water calcium carbonate. The use of the equation
systems, it is Important to bear in mind the developed by Langelier made it possible to
mechanism causing the deposit, otherwise predict the tendency of natural or condi-
confusion may result. In general, the term tioned water to either deposit calcium car-

Figure 31-1-Analysis of Typical Scales from Once-Through Cooling Water System

% % % %
Calcium as CaO 47.2 34.4 43.5 52.3
Magnesium as MgO 2.1 1.7 0.6 4.1
Iron as RzOa 3.1 22.7 10.72 3.00
Aluminum as Al 2 0a 1.5 4.2 0.88 0.20
Carbonate as C02 34.2 22.9 34.0 35.2
Sulfate as SOa 0.9 0.1 1.4 0.3
Silica as Si02 4.9 9.8 2.1 1.1
Loss on Ignition 5.9 4.1 6.4 2.1
2.7 2.6 2.5 2.4 2.3 2.2 2.1 2.0 1.9 1.8 1.7 1.6 1.5 1.4 1.3 1.2 1.1 ~
pCa >-1
5ooo I I I I Ill I I I I I\ I I I I I I ~ I I I I I 1\ I I I I I ft I I I I l I I I Ill I I I Ill I I I ~ I I I l I I I I I I I I I I l'ol I 1\ I I II I I ~ I I I I A I I A 5000 >

:::: IIIII NII II Nlllllllllll M1111\1111\111 NIll ill IN IIIlii NII NII NIf.J Itllltll II:::: ~
z 1000 1000 "tl 0
0 > 0
::::; ::u zt;
.... -1
i 500 500 (II
~ 300
"tl ......
300 ::u z......
en 200
....!,... z0
100 100 z

50 50
30 30
20 20

10 10
4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0

TEMP.= 140F. pH= 7.80 SUM= pHs = 6. 76

Figure 31-2 • Langelier Saturation Index Chart <.>:>

bonate from solution or to dissolve calcium While a high hardness water, with a positive
carbonate with which the water is brought Saturation Index will definitely lead to cal-
in contact. The equation takes into considera- cium carbonate scale formation, a low hard-
tion readily obtained analytical values such ness water, with the same positive Saturation
as pH, calcium, total alkalinity, dissolved · Index, may not form any appreciable calcium
solids and temperature. carbonate scale.
In simplified form, and for waters in the In an attempt to secure a quantitative
pH range of 6.5-9.5, Langelier's equation for index, Ryznar has proposed the Stability
the pH at which the water is in equilibrium Index. This index is empirical, based on a
with calcium carbonate (pHs) may be study of actual operating results with waters
written as follows: of various saturation indexes.
pHs = (pK2 t - pKs 1 ) + pCa + pAik Stability Index = 2(pHs) - pH
The two latter terms of the above equa- With waters having a Stability Index of
tion are negative logarithms of the molal 6.0 or less, scaling increases and the tendency
··· a~d equivalent concentrations of calcium and to corrosion decreases. When the Stability
titratable base respectively. The terms pK 2 1 Index is above 7.0, a protective coating of
and pKs1 are respectively the negative loga- calcium carbonate scale may not be de-
rithms of the second dissociation constant for veloped. Corrosion would become an increas-
carbonic acid and the activity product of ingly greater problem as the Stability Index
calcium carbonate. The difference (pK 21 - increases above 7.5 or 8.0. In some cases, the
pKs 1 ) varies with ionic strength and tem- use of the Stability Index along with the
perature. The term pHs represents the pH Saturation Index will help in predicting more
at equilibrium assuming no change has taken accurately the scaling or corrosive tendencies
place in the water bringing it to equilibrium. of a water.
The Saturation Index is defined as the
algebraic difference between the actual meas- PREVENTION OF CALCIUM CARBONATE SCALE
ured pH of the water and the calculated
pHs at saturation with calcium carbonate. Because of the large volumes of water used
Saturation Index = pH - pHs in once-through systems, external softening
This index shows qualitatively the tend- of the raw water is not usually economically
ency for deposition or solution of calcium feasible as a means of scale prevention. How-
carbonate. A positive Saturation Index will ever, a softening plant may be installed to
indicate a tendency to deposit calcium car- supply softened water for a variety of plant
bonate. A negative Saturation Index will purposes which may also include some once-
indicate an undersaturation condition with through cooling uses. In such cases, the
respect to calcium carbonate and hence a treated water balances are controlled so as
tendency to dissolve any existing calcium car- to provide a Saturation Index that will pre-
bonate. A "zero" Saturation Index denotes vent calcium carbonate deposition at the tem-
that the water is exactly at equilibrium with peratures involved.
respect to calcium carbonate. However, once-through cooling water is
Langelier's Equation has been slightly generally not softened by external treatment
modified by other investigators. Various nom- processes even though such softening facil-
ographs have been prepared to simplify the ities may be installed for other plant uses.
calculation of the pHs. A nomograph suit- The chemical treatment cost in softening the
able for this purpose is shown in Figure 31-2. large volumes of once-through cooling water
It should be emphasized that the Saturation is usually prohibitive. The investment cost in
Index is only a measure of the directional the treatment plant for handling such large
tendency or driving force of a water. It is in volumes of water is also another factor dis-
no way intended to be a measure of capacity. couraging such softening.

The most common method employed to systems of this type, the feed water of definite
inhibit calcium carbonate scale formation in composition is pumped at a constant rate
once-through cooling water systems is the through a scaling chamber, and scale forma-
use of various anti-nucleating agents. Except tion develops on a cartridge heater. Temper-
under unusual conditions, these materials ature, rate of flow and electrical input are
provide very satisfactory inhibition of scale automatically controlled.
formation, at a low chemical treatment cost. Consistent results are obtained with the use
An anti-nucleating agent employed for this of the same type of heating surface. Figure
putpose is one which possesses the property 31-6 illustrates a system of this type.
of preventing crystal growth and thereby scale
Figure 31-3 is illustrative of the data se-
formation. In effect, these agents broaden
cured under such controlled conditions in the
or increase the solubility of the scale forming
prevention of calcium carbonate scale forma-
salts and permit an oversaturated condition
tion. While most of the starches tested ex-
to exist without precipitation from solution.
hibit relatively low power in preventing cal-
The materials most commonly used include
ciun'l carbonate scale under test conditions,
the polyphosphates, tannins, lignins, starches,
appreciable power is shown by some of the
polyacrylates and combinations of these ma-
tannins and lignins. The most effective re-
terials. By blending several of the various
sults have been secured with blends of sev-
agents it is possible to secure the additive
eral different agents. For example, some
benefits of each for application to that par-
agents exhibit very high scale prevention
ticular problem.
powers per ppm in the low treatment range
In addition, chelants can be used in certain of 1-5 ppm. Increasing amounts of treatment
types of once-through cooling water systems above this range does not result in any ap-
for the purpose of chelating metal ions, such preciable increase in scale prevention powers.
as calcium and magnesium. Chelants react Other materials show practically a straight
stoichiometrically, and the reaction products
are usually very stable. Typical agents of this
type are ethylenediamine tetraacetic acid, Fig. 3.1-3
citric acid and gluconic acid. Effect of Surface Active Treatments in
Crystal distortion has been advanced as Reducing Calcium Carbonate Scale
one of the main forces in reducing carbonate
scale by the use of anti-nucleating agents. Heat
Treatment Type Transfer Rate CaC03 Scale % Scale
Figure 31-4 illustrates crystals of pure cal- Cone., ppm Organic Btu/1! 2 /hr Grams/1!2 Reduction
cium carbonate at 450 magnifications. Figure 0 16,000 26.8
31-5 is a photomicrograph of crystals of cal- 50 Starch A 16,000 23.3 13
cium carbonate formed in the experimental 50 Starch B 16,000 26.8 0
50 Starch C 16,000 18.2 32
laboratory system shown in Figure 31-6. The
crystal distortion shown by Figures 31-7, 31-8 0 24,000 18.7
and 31-9 illustra,tes the effect of various treat- 50 Tannin A 24,000 16.4 12
50 Tannin B 24,000 6.6 65
ments. Laboratory correlation of crystal dis- 50 Tannin C 24,000 12.3 34
tortion with the amount of scale actually
formed on the heat transfer surface indicates 0 17,000 34.6
that those which do not produce crystal dis- 50 Lignin A 17,000 25.6 26
50 Lignin B 17,000 38.1 -lQ
tortion also do not reduce the amount of 50 Lignin C 17,000 10.7 69
scale formation.
0 17,000 37.7
Study of the inhibitory effects of various 2 Blend A 17,000 18.1 52
agents for the prevention of calcium carbo- 25 Blend B 17,000 5.6 85
nate deposition can be made with relatively 25 Blend C 17,000 2.7 93
50 Blend D 17,000 0.4 99
simple once-through laboratory systems. In

Figure31-4 • Pure CaC0 3 (Merck), 450 Magnifications Figure 31-7 • Distorted CaCOa Crystals with Tannin

Figure 31-5 • CaCOa Crystals (17,000 Btu/Ft2/hr), Figure 31-8 • Distorted CaC0 3 Crystals with
450 Mag. No Treatment Polyphosphate Treatment

Figure 31-6 • Once- Through Cooling Water Figure 31-9 • Distorted CaC0 3 Crystals with
Experimental Unit Polyphosphate- Tannin Treatment Blend

line relationship with scale properties in- ing iron bearing well waters. This precipita-
creasing in direct proportion to the concen- tion of iron can be prevented by the use of
tration of treatment employed. By combin- sequestrants or dispersants, which are very
ing two or more such agents into a blended specific in their activity on ferrous or ferric
treatment, it is possible to secure their addi- ions. Dispersants are adsorbed on the iron
tive benefits over a wide range of treatment particles while the iron is still in the ionic
concentrations. The results in Figure 31-3 for form. Residual charges on the adsorbed parti-
such a combination treatment indicate the cles are similar, and hence cause a stable,
high degree of scale preventive power that dispersed condition. Polyphosphates, lignin
can be incorporated into a single blended derivatives and organic phosphates are ex-
treatment. amples of treatments employed for this pur-
With the use of anti-nucleating agents, it pose.
is possible to secure complete prevention of Iron deposits also result from corrosion
calcium carbonate precipitation unless the products. In this case, the basic problem is
water is highly supersaturated with respect prevention of the corrosion, although poly-
to calcium carbonate at the temperature of phosphates may be employed to minimize the
use. In such cases, it is necessary to decrease "red water" problem, as discussed more fully
the positive Saturation Index, usually through in the next chapter.
the use of acid treatment. Complete elimina-
tion of the positive Saturation Index is not REFERENCES
required under such circumstances, but only H. L. Kahler, "Once-Through and Recirculating
enough adjustment in this index to bring the Cooling Water Studies," Proceedings, Engineers
water thus treated within the effective range Soc. of Western Penna., pp. 39-62 (1944)
W. F. Langelier, "The Analytical Control of Anti-
of surface active agents. Corrosion Water Treatment," Journal, Am. Water
Works Assoc., Vol. 28, pp. 1500-1521 (1936)
IRON DEPOSITS J. W. Ryznar, "A New Index for Determining
Amount of Calcium Carbonate Scale Formed by a
A special case of scale formation from fer- Water," Journal, Am. Water Works Assoc., Vol.
36, pp. 472-486 (1944)
rous bicarbonate may be encountered on J. M. Donohue, "Organic Topping Agents," South-
once-through cooling water systems employ- ern Engineering, Vol. 85, pp. 52-54 (Jan. 1967)

Causes of Corrosion . . .......... /' . . . . . . . . . . 237

Water Characteristics. . . . . . . . . . . . . . . . . . . 237
Effect of Temperature ................ ··: : 238

32 Effect of Flow Rate . ..... : . . . . . . . . . . . . . . 238

Dis~imilar Metals . .................... ; 238

Cooling Water Treatment:

Corrosion Control
Prevention of Corrosion . ................... 240
CaC0 3 Protective Scale. . .. . . . . . . . . . . . . . . . 240
Mechanical Deaeration. . . . . . . . . . . . . . . . . . 240
Chemical Deaeration. . . . . . . . . . . . . . . . . . . . 240
Corrosion Inhibitors . .................... 241
Polyphosphates. . . . . . . . . . . . . . . . . . . . . . 241
Silicates. . . . . . . . . . . . . . . . . . . . . . . . . . . 242
Dian odic-:(inc Dianodic . . . . . . . . . . . . . . 243

o~RosroN of ferrous metals in. distribu- 0>0

tiOn systems, once-through coolmg water

systems, plant piping, etc. is a serious prob- "" v
lem from several standpoints. Not only is the 0.20 /
line replacement costly in itself, but the labor !"
cost of removing and replacing buried lines,
or those imbedded in walls, is very high. Cor-
"" ,v <f)Y


rosion products also reduce the carrying ca- / v

pacity of lines, increase frictional resistance
and pumping costs. The iron oxide formed as
a result of corrosive action occupies a volume

~ r:.--

' 4

- 48 F


many times greater than the original metal

and eventually may result in complete block- Figure 32-1 • Effect of Oxygen Concentration on Corro-
age. "Red water," due to corrosion products, sion at Different Temperatures
gives an unsightly appearance to the water
and may interfere with processes sensitive to geri at the cathodic surface, forming water
iron. and depolarizing the surface, thus permitting
additional iron to dissolve. Figure 32-1 shows
CAUSES OF CORROSION the effect of oxygen concentration on the cor-
Corrosion is defined as the destruction of a rosion of low carbon steel specimens at dif-
metal by chemical or electrochemical reaction ferent temperatures. Philadelphia (Delaware
with its environment. In the corrosion of fer- ·supply) tap water was used in these tests,
rous metals in contact with moisture, rever- with a pH of 6. 7. Linear rate of flow past
sion of the metal to the oxide form takes the test specimens was 2.5 feet per second.
place. The driving force or the tendency to The chief s~urce of dissolved oxygen in
corrode is electrochemical. However, the rate water is contact with the atmosphere. Deep
at which corrosion proceeds is dependent on well waters are usually devoid of oxygen, while
the resistance to continued attack caused by most surface supplies are saturated with oxy-
the products of the corrosion reactions. gen. The solubility of .oxygen in water is de-
In water systems, the principal factors in- pendent on the temperature and atmospheric
fluencing the corrosion of ferrous metals are pressure. Increased pressure increases oxygen
the water characteristics, temperature, rate of solubility, while increase in temperature low-
water flow and contact with dissimilar met- ers the sol~bility of oxygen.
als. The chief variables controlling the cor- Generally, oxygen produces an easily iden-
rosive clta:racteristics of a water are its dis- tified corrosion in the form of small pits or
solved oxygen concentration, carbon dioxide depressions. As corrosive attack continues, the
content, pH and dissolved solids. Other fac- pit usually increases, not only in area but also
tors are involved) of course, such as the pos- in depth, and the nodule of corrosion prod-
sible presence of free mineral acid, hydrogen ucts becomes enlarged. The detrimental ef-
sulfide, sulfur dioxide, etc. In general, how- fects of corrosive attack are more severe when
ever, dissolved oxygen and carbon dioxide are this corrosion is localized in the form of pits
the chief problems, with pH directly affected than would be the case if general metal wast-
by the carbon dioxide concentration. In- age occurred at the same rate over a large
creased dissolved solids, particularly chloride area. The centralization of corrosion at the
and sulfate, enhance the corrosive effect of points of pitting permits deeper penetration
oxygen and carbon dioxide. of the metal and more rapid failure at these
WATER CHARACTERISTICS. Dissolved Oxygen. points.
Oxygen attacks ferrous metals, causing cor- pH. At pH values below approximately 4.3
rosion and pitting. Oxygen reacts with hydro- (the region of free mineral acidity), the chief

controlling factor in promoting corrosion is anodic to the colder areas.

pH. Acid waters, such as those contaminated EFFECT OF FLow RATE. Increase in the rate
with acid mine drainage, are quite corrosive. of flow tends to increase the rate of corro-
Neutralization of such acidic characteristics sion. Figure 32-2 illustrates data secured
is a necessity for practically every water use. through the exposure of Y4" tubular low car-
However, in the range of most natural wa- bon steel specimens to Philadelphia (Dela-
ters, between pH 6.0-8.0, the pH of the water ware supply) tap water at a temperature of
is not the controlling corrosive factor, in the 120 F. The corrosion rate increases rapidly
presence of oxygen. with increases in the rate of flow although at
Carbon dioxide concentration, pH and al- higher flow rates this influence is less marked.
kalinity of a water are directly related. For a Similar effects have been noted on tests with
given alkalinity, the higher the carbon diox- Pittsburgh, Pa. tap water and with sea water.
ide, the lower the pH. In the absence of oxy- The influence of flow rate on corrosion is
gen, as encountered with the use of deep well probably primarily the depolarization effect
wa.ter, pH is a controlling influence on the secured by bringing larger quantities of oxy-
corrosion rate. At lower pH values, corrosion gen into contact with the cathodic surface,
increases. Carbon dioxide is a factor in such together with the effect of greater amounts
corrosion since it influences pH. In the of ions (such as chloride and sulfate) on the
presence of oxygen, however, carbon dioxide anodic surface.
like pH is not the major influence on the rate
of corrosion. 0.35

Dissolved Solids. In general, increase in the

dissolved solids content of a water increases ~ 0.30
the corrosivity. Higher solids content increases a:
the conductivity of the water, thus favoring ... 0.25
increased corrosion. The chloride ion, in C/)
particular, accelerates corrosive action al- ~ 0.20
though at very high concentrations (brine) a z"
decrease in corrosivity is observed. Various ~ 0.15
explanations have been advanced for the cor- ....lUa:
rosive effect of higher dissolved solids and z
~ 0.10
one logical theory involves the effect of ions
in destroying the coating of corrosion prod-
ucts on the metal. High chloride concentra- "":5 0.05
tions, for example, also hinder the formation ~

of protective coatings by inhibitors, such as o.o 0 0.5 1.0 1.5 2.0 2.5 3.0 3.5
actions of corrosion increase in rate with in- Figure 32-2 • Influence of Flow Rate on Corrosion
crease in temperature. Reference to Figure
32-1 shows the accelerated corrosion at 90 F DisSIMILAR METALS. The corrosion that oc-
and 120 F, in comparison with 48 F for vari- curs from the contact of dissimilar metals is
ous oxygen concentrations. Increased tem- known as galvanic action. The tendency of a
perature decreases water viscosity and thus in- metal to dissolve and enter solution as an ion
creases the diffusion rate of oxygen. Elec- is measured by the electrode potential between
trical conductivity is increased at higher tem- the metal and its ions. The listing of the elec-
peratures, thus producing an effect similar to trode potentials of the various metals consti-
increase in dissolved solids content. Pitting tutes the familiar Electromotive Series. The
may be enhanced where adjacent metal areas electrode potential of hydrogen is defined as
vary in temperature, with the hotter sections zero. The metals higher in the table are con-

sidered able to displace those metals lower in Fig. 32-3

the table.
Galvanic Series of Metals and Alloys*
The Electromotive Series cannot be inter- CoRRODED END
preted as a precise indication of galvanic ac-
(anodic, or least noble)
tion. The potentials of the various metals can
be altered by ion concentration, temperature, Magnesium
and the corrosive environment. Consequently, Magnesium alloys
in practical service conditions, it is possible
to 'reverse the order in which some adjacent
metals are listed in the Electromotive Series. Aluminum 2S

Figure 32-3 lists a galvanic series of metals Cadmium

and alloys, based bn corrosion testing experi- Aluminum 17ST
ence. While closely allied to the Electromotive Steel or Iron
Series, it is based on practical experience Cast Iron
rather than theoretical consideration alone.
It will be noted that some of the metals are Chromium-iron (active)
grouped together. Metals so grouped possess Ni-Resist
little tendency to produce galvanic corrosion 18-8-Cr-Ni-Fe (active)
of each other and from a practical standpoint 18-8-3-Cr-Ni-Mo-Fe (active)
are relatively safe to use in contact with each
other. However, coupling of two metals from Hastelloy C
different groups and distant from each other Lead-tin solders
in the list will result in galvanic, or accele- Lead
rated, corrosion of the one higher in the list. Tin
The tendency for galvanic corrosion is in- Nickel (active)
creased when the metals coupled together are
far apart in the list. While such a galvanic lnconel (active)
series is a useful guide, it should not be in- Hastelloy A
terpreted as an absolute rule under all cor- Hastelloy B
rosive service conditions. Brasses
Corrosion of ferrous metals in the presence Copper
of moisture is a truly vast subject and it is Bronzes
possible to only touch upon the various influ- Copper-nickel alloys
encing factors in this discussion. Many other Titanium
variables will affect the corrosion rate, such Monel
as the homogeneity of the metal, condition of Silver solder
the metal surfaces, ratio of surface areas ex-
posed by two dissimilar metals, static and cy- Nickel (passive)
cling stresses, oxygen concentration cells, etc. lnconel (passive)
Water characteristics also are influenced by Chromium-iron (passive)
factors such as hydrogen sulfide, chlorine, and 18-8 Cr-Ni-Fe (passive)
biological influences which can markedly in- 18-8-3 Cr-Ni-Mo-Fe (passive)
fluence the corrosion rate. The technical lit- Silver
erature on corrosion is quite voluminous. In Graphite
addition, several excellent books on this spe-
cific subject are included among the chapter PROTECTED END

references and should be consulted by those (cathodic, or most noble)

wishing to delve more deeply into this com- • "Corrosion Testing Bulletin," International Nickel Com-
plex subject. pany, 1944.

PREVENTION OF CORROSION duces an oversaturated condition at the point

of higher temperature. If the calcium carbo-
The prevention of corrosion in once-through nate balance is so set th(l.t only a small
systems is accomplished either by forming a amount of deposit will form at the higher
calcium carbonate scale on the metal surface temperature, none will form at the lower
to act as a protective coating against corro- temperature and no protection against cor-
sion, by elimination of the corrosive oxygen rosion will be provided at that point. In ad-
itself from the water, or by the use of rela- dition, within the limits of plant control, it
tively small quantities of organic and inor- is exceedingly difficult to control scale forma-
ganic corrosion inhibitors. tion to such an exact point as to provide pro-
CALCIUM CARBONATE PRoTECTIVE ScALE. The tection against corrosion and yet not interfere
Langelier Saturation Index is useful in pre- with heat transfer.
dicting the tendency of a water to deposit or The controlled formation of calcium car-
dissolve calcium carbonate. The Saturation bonate scale as a means of corrosion preven-
... Index can be used as a guide in the controlled tion is successfully employed in some water
deposition of calcium carbonate scale. If such works distribution systems where increase in
a scale can be deposited uniformly over the temperature is not encountered. For indus-
metal surfaces throughout the system, it will trial cooling water systems, however, where
act as a physical barrier, preventing contact heat transfer is involved, this method of cor-
of the water with the metal surface. The cor- rosion prevention is impractical, except under
rosive oxygen content of the water cannot ideal control conditions.
contact the metal and corrosion is prevented. MECHANICAL DEAERATION. The corrosive
In effect, a protective coating similar to qualities of a water can usually be practically
cement lining of the pipe has been forn!ed. eliminated by oxygen removal by deaeration.
In order to develop the positive Saturation Vacuum deaeration for once-through cooling
Index required for deposition of calcium car- water systems has been applied successfully
bonate, it is necessary to adjust the pH, to a number of installations. Where oxygen
alkalinity or calcium content of the water so removal is not complete, the remaining traces
as to raise the pH slightly above the calcu- of oxygen can be removed chemically with
lated pH of saturation. The materials em- catalyzed sodium sulfite.
ployed for this purpose may be soda ash, Of course, there is considerable investment
caustic soda and lime (calcium hydroxide) . cost, as well as continued operating cost, in-
Lime is usually the most economical alkali volved with vacuum deaeration. These fac-
to employ since it raises the calcium content tors have hindered the more widespread use
as well as the alkalinity and pH. A lesser of this method. Under certain circumstances,
amount is thus required to develop the pH however, vacuum deaeration is unquestion-
of saturation. ably the most effective means of corrosion
Theoretically, it is possible to control the control and gradual extension of its use is to
deposition of calcium carbonate scale to such be expected in the future.
a point that protection against corrosion is CHEMICAL DEAERATION. Sodium sulfite is
provided, and the scale film is sufficiently thin commonly used in the treatment of boiler
to be non-objectionable. However, such a feedwater as a chemical deaerant and reacts
treatment method is more theoretical than with dissolved oxygen as follows:
practical inasmuch as a definite positive Sat-
uration Index of a certain value pertains only Na 2 S03 + 0 Naz$04
for the temperature for which it has been
sodium sulfite + oxygen = sodium sulfate
calculated. With different temperatures This reaction proceeds rapidly at tempera-
throughout a cooling water system, the main- tures of 212 F and above. However, at cold
tenance of a protective film on the metal sur- water temperatures, the reaction between so-
face at the point of lowest temperature pro- dium sulfite and dissolved oxygen is quite