Power Plant Water Chemistry A Practical Guide

Ali Sadeghi Digital Library
00fm.qxd 8/8/06 1:49 PM Page i
P O W E R P L A N T
W A T E R
CHEMISTRY
A PRACTICAL GUIDE

00fm.qxd 8/8/06 1:49 PM Page ii

00fm.qxd 8/8/06 1:49 PM Page iii
Condensate/Feedwater Chemistry iii
P O W E R P L A N T
A PRACTICAL GUIDE
by
Brad Buecker

Disclaimer: The recommendations, advice, descriptions, and the methods in this
book are presented solely for educational purposes. The author and publisher assume
no liability whatsoever for any loss or damage that results from the use of any of the
material in this book. Use of the material in this book is solely at the risk of the user.
Copyright© 2006 by
PennWell Corporation
1421 South Sheridan Road
Tulsa, Oklahoma 74112-6600 USA
800.752.9764
+1.918.831.9421
sales@pennwell.com
www.pennwellbooks.com
www.pennwell.com
Marketing Manager: Julie Simmons
National Account Executive: Barbara McGee
Director: Mary McGee

Managing Editor: Marla Patterson
Production Manager: Sheila Brock
Library of Congress Cataloging-in-Publication Data
Buecker, Brad.
Power plant water chemistry : a practical guide / Brad Buecker.
p. cm.
Includes bibliographical references and index.
ISBN 0-87814-619-9
ISBN 978-0-87814-619-2
1. Steam power plants. 2. Chemical engineering I. Title
TJ403.B82 1997
621.31’2132--dc21 97-35403
CIP
All rights reserved. No part of this book may be reproduced, stored in a retrieval system,
or transcribed in any form or by any means, electronic or mechanical, including photo-
copying and recording, without the prior written permission of the publisher.
Printed in the United States of America.
4 5 6 7 8 12 11 10 09 08
IV.indd 1 8/24/08 12:00:05 PM

00fm.qxd 8/8/06 1:49 PM Page v
Dedication
To Nancy and Alyssa, each of whose creativity easily exceeds my own,
and to my parents whose own lives have shown me the value
and rewards of hard work.

00fm.qxd 8/8/06 1:49 PM Page vi

00fm.qxd 8/8/06 1:49 PM Page vii
Table of
Contents
Figures and Tables
Acknowledgments xi
Preface xii
1 INTRODUCTION TO STEAM GENERATION WATER

CHEMISTRY SYSTEMS 1
2 CONDENSATE FEEDWATER CHEMISTRY 5

Introduction 5
Condensate/Feedwater System Construction Materials 6
Condensate Chemistry Dissolved Oxygen 6
Carbonic Acid Corrosion in Industrial Steam Generating Systems 8
Mechanical Removal of Dissolved Oxygen and Other Gases 10
Chemical Control of Dissolved Oxygen 13
Oxygen Scavengers 13
Safe Hydrazine Feed Systems 17
Industrial Oxygen Scavenging 18
Flow-Accelerated Corrosion 19
Copper-Alloy Corrosion and Preventive Measures 19
Additional Corrosion Control Requirements 20
Oxygenated Treatment 21
Feedwater Chemistry Guidelines 31
Chemical Feed Systems 31
Monitoring and Control of Condensate Contaminants 32
Condensate Polishing and Treatment of Condensate Return to
Industrial Boilers 33
High-Pressure Boiler Condensate Polishing 34
Deep-Bed Condensate Polishing 35
Powdered-Resin Condensate Polishing 37
Industrial Plant Condensate Polishing 38
Conclusion 39

00fm.qxd 8/8/06 1:49 PM Page viii
viii Table of Contents
Supplement 2-1: Why Condenser Performance is Important 41

Supplement 2-2: BASIC Program for Monitoring Condenser
Performance 42
Case Histories 46
3 BOILER WATER CHEMISTRY 51

Introduction 51
Drum-Type Boiler Design 51
Package Drum Boilers 52
Field-Erected Drum Units 52
Circulating Fluidized Bed Boilers and Heat Recovery
Steam Generators 53
Once-Through Steam Generation 54
Boiler Water Contamination 55
Iron Oxide Deposition 55
Condenser Inleakage 56
Boiler Water Treatment Programs 60
Early Boiler Water Treatment 61
Coordinated and Congruent Phosphate Treatment 62
Phosphate Hideout 64
Alternative Phosphate and Nonphosphate Programs 66
Equilibrium Phosphate Treatment 66
Phosphate Treatment 67
Chelants and Polymer Treatments 67
Caustic Treatment 69
All-Volatile Treatment 69
Oxygenated Treatment 70
Heat Recovery Steam Generators 70
Sampling 72
Layup and Off-Line Corrosion Protection 72
Conclusion 73
Supplement 3-1: BASIC Program for Calculating
Sodium-to-Phosphate Ratios of Boiler Water 75
Case Histories 76
4 STEAM CHEMISTRY 83
Introduction 83
Primary Carryover Products 84
Copper 84
Sodium Hydroxide 85

00fm.qxd 8/8/06 1:49 PM Page ix
Table of Contents ix
Chloride and Sulfate 86

Iron Oxides 86
Silica 87
Sodium Phosphates 87
Organics 87
Mechanical Carryover 88
Vaporous Carryover 90
Solids Introduction by Contaminated Attemperator Water 91
Superheater Exfoliation 91
Water Chemistry Limits to Prevent Steam Contamination by Carryover 92
Boiler Water Chemistry Guidelines for Control of Steam Chemistry 92
Prevention of Contamination via the Attemperator System 93
Control of Superheater and Reheater Exfoliation 93
Steam Chemistry Monitoring 94
Steam Chemistry Issues at Industrial Plants without Turbines 94
Conclusion 95
Case Histories 97
5 HIGH-PURITY MAKEUP WATER TREATMENT 101

Introduction 101
Pretreatment 101
Microbiocide Feed 102
Clarification and Softening 103
Filtration 105
Activated Carbon Filtration 106
Additional Pretreatment Methods 106
High-Purity Makeup Treatment Methods 106
Ion Exchange 107
Exchange Groups 108
Strong Acid Cation Resins 109
Weak Acid Cation Resins 109
Strong Base Anion Resins 109
Weak Base Anion Resins 109
Demineralizer Configurations and Mixed-Bed Exchangers 110
Degasifiers 111
Regeneration and Co-Current/Countercurrent Systems 111
Strong Base Anion Regeneration 113
Weak Acid and Weak Base Exchangers 113
Demineralizer Performance Calculations 115
Strong Acid Cation Exchanger Calculations 115

00fm.qxd 8/8/06 1:49 PM Page x
x Table of Contents
Strong Base Anion Exchanger Calculations 121

Weak Acid and Weak Base Performance 123
System Design Calculations 126
Resin Volume and Vessel Diameter Calculations 128
Backwash Requirements 128
Regeneration Requirements 129
Rinsing 129
Mixed-Bed Polishing 130
Monitoring Performance of Ion Exchanger Vessels 131
Demineralizer Component Fundamentals 132
Distributors 133
Vessels 133
Valves 133
Materials 134
Packed-Bed Demineralizers 135
Other Makeup Technologies 136
Reverse Osmosis 136
RO Membrane Design 137
RO Membrane Material 137
RO Pretreatment 139
RO Design 140
RO Components 142
RO Flow Control and Monitoring 143
RO Alarms 145
Size of a Reverse Osmosis System 145
RO Cleaning 146
Electrodialysis and Electrodialysis Reversal 146
Electrodeionization 147
Conclusion 148
Supplement 5-1: UV Light Disinfection 149
Supplement 5-2: Multimedia Filtration 150
Supplement 5-3: Activated Carbon 151
Supplement 5-4: Sulfuric Acid and Caustic Specifications 151
Supplement 5-5: Silt Density Index 152
6 COOLING WATER CHEMISTRY 155

Introduction 155
Cooling Systems 156
Once-Through Systems 156
Open Recirculating Systems 157
Cooling Towers 157

00fm.qxd 8/8/06 1:49 PM Page xi
Acknowledgements xi
Cooling Tower Calculations 160

Closed Cooling Systems 165
Cooling Water Corrosion, Scale, and Deposit Mechanisms 166
Corrosion 166
Corrosion Influencing Mechanisms 169
Nonmetallic Corrosion 170
Corrosion Inhibitors 170
Scale 173
Calcium Carbonate 173
Other Scales 177
Scale Control 177
Acid Feed 178
Alkaline Treatment Methods 179
Fouling 181
Nonmicrobiological Fouling Control Methods 182
Microbiological Fouling 183
Control of Microbiological Organisms 184
Chlorine 184
Nonoxidizing Biocides 189
Biocide Discharge Limits—Future Trends 190
Macrofouling 190
Zebra Mussels 191
Conclusion 193
Supplement 6-1: Cooling Tower BASIC Program 195
Supplement 6-2: LSI and RSI BASIC Program 196
Supplement 6-3: Acid Feed Calculations 198
Supplement 6-4: Sizing a Sidestream Filter 199
Supplement 6-5: Specifications for Sodium Hypochlorite 200
Case Histories 200
7 SAMPLING 205
Introduction 205
The Need for Sampling 206
Sample Point Selection 206
Makeup System Effluent 206
Condensate Storage Tank Effluent 211
Condensate Pump Discharge 211
Condensate Polisher Effluent 212
Deaerator Inlet 212
Deaerator Outlet/Boiler Feed Pump Suction 212

00fm.qxd 8/8/06 1:49 PM Page xii
xii Preface
Feedwater or Economizer Inlet 212

Boiler Water 213
Saturated Steam 213
Main/Reheat Steam 214
Cogeneration/Combined-Cycle/Industrial Plant Sampling 215
Techniques to Obtain Representative Samples 216
Sample Nozzle Design 217
Sample Nozzle Installation 218
Sample Point Location 218
Primary Sample Conditioning 218
Sample Flow Rate and Line Size 219
Final Sample Conditioning 219
Data Acquisition 222
Conclusion 223
Case Histories 227
Bibliography 229
Index 233

00fm.qxd 8/8/06 1:49 PM Page xiii
Acknowledgments xiii
Acknowledgments
No book of any substance can be written without assistance from others. I
would especially like to thank Ray Post of BetzDearborn for his insightful com-
ments on cooling water chemistry. They immeasurably helped me with the cool-
ing water chapter. I would also like to thank Nissen Cahan and the people at
Purolite for allowing me to use data about their ion exchange resins. The data
was enormously useful. Similar thanks go to Dr. Barry Dooley and his staff at the
Electric Power Research Institute for allowing me to use illustrations from some
of their technical reports.
Others who provided very helpful data, illustrations, photos, or advice
include Terry Dwyer at the Marley Cooling Tower Company, Don Walter from
Osmonics, Tom Svoboda of the Sentry Equipment Corporation, Lee Machemer
from Jonas & Consultants, Phil Di Vietro from ASME, Jim King from DB Riley,
and several people from U.S. Filter.
I would also like to thank COADE Engineering Software of Houston, Texas
for allowing me to make use of their unit conversion program in preparing this
book. I have used this program many times over the last several years. I would
also like to thank John Meinders of the Kansas City, Kansas Board of Public
Utilities for allowing my colleague John Wofford and I to take photos at his plant.
A special thanks also to John Wofford for being able to answer virtually every
question I had about boilers while preparing this book.
Last, but not least, I would like to acknowledge my friends and former co-
workers at City Water, Light & Power in Springfield, Illinois, and in particular
Tom Bee, Ellis Loper, Ed Riordan, Dave Arnold, and the late Charles Hartman.
During my 12-year
career at the utility, I was given the opportunity to work on a wide variety
of projects related to analytical chemistry, water and wastewater treatment, flue
gas desulfurization, and plant engineering. Without this experience I could never
have even attempted to write this book.

00fm.qxd 8/8/06 1:49 PM Page xiv
xiv Preface
Preface
When PennWell Publishing asked me to write this book, I was unsure at first
how to arrange the contents. Two concepts emerged as I began to gather mater-
ial and put my thoughts in writing. One was to provide as much practical infor-
mation as possible regarding the core areas of steam generation chemistry. It is
my hope that utility and industrial personnel, and especially those who may be
somewhat new to this type of work, can open the book and find solid guidelines
and examples to follow. I included a number of case histories from my own expe-
riences and those of several colleagues to illustrate common, and a few not-so-
common, difficulties faced by steam generating personnel.
Additionally, I tried to include new information regarding steam generation
chemistry, makeup water techniques, and boiler water chemistry to give the read-
er an overview of trends within these areas. For instance, coordinated/congruent
phosphate boiler water treatment is losing ground to other programs; reverse
osmosis has grabbed a big share of the makeup water treatment market, and
steam chemistry guidelines keep tightening as researchers learn more about the
effects of steam contaminants on turbine components. Steam generation person-
nel need to stay abreast of these developments.
Much of the data in this book represents the latest ideas regarding steam
generation chemistry. However, this book makes no guarantees regarding unit
performance, and I encourage readers to use the data as a guideline and not as
an absolute for their system(s). Even sister units exhibit different operating char-
acteristics, so treatment and operating chemistry must, to some extent, be spe-
cific to each unit.
I eagerly invite readers of this book to send me suggestions, comments, and
even their own case histories for inclusion in a possible future edition. I firmly
believe that the best teacher is experience, and I would be very pleased to hear
from those of you who deal with steam generation chemistry on a daily basis.

01.qxd 8/8/06 1:50 PM Page 1
Chapter 1
Introduction to
Steam Generation
Water Chemistry
Systems
Steam generation, whether it be for power production or industrial process

use, is a complex process. A steam generating plant is filled with pumps, piping,
valves, electrical wiring, instruments, and of course, one or more boilers. All
must work together to generate the desired product, steam, which may range
from saturated conditions at less than 100 psig to supercritical steam at 1050˚F.
The primary water systems in a steam plant include:

● Feedwater/boiler/afterboiler circuit
● Makeup system
● Condenser cooling
● Closed cooling water
● Ash sluicing at coal-fired plants
The contrasting nature between differing systems requires water of varying

quality, and also requires different treatment methods. For instance, in a well-
sealed, closed cooling water system, a simple corrosion inhibitor may be the only
chemical needed to protect system components. In a boiler, where water tem-
Ali Sadeghi Digital Library 1

01.qxd 8/8/06 1:50 PM Page 2
2 Power Plant Water Chemistry: A Practical Guide
peratures can reach 600˚F or higher and steam temperatures 1050˚F, highly puri-
fied feedwater, dosed with carefully controlled treatment chemicals, is required
if the boiler is to operate properly. Table 1–1 illustrates some of the effects that
contaminants have on boiler water systems.
The vastly different conditions between water systems, and the complexity
of a steam generating system, make the chemist’s job very lively. This book pro-
vides practical examples of water chemistry issues and problems for steam gen-
erating systems, and illustrates techniques and methods to control chemistry. It
also provides details on many of the latest trends, findings, and developments in
the areas of boiler water chemistry, steam sampling, and makeup water produc-
tion. Utility chemists and researchers have made many discoveries and improve-
ments to steam generation chemistry within the last decade. A number of these
have challenged traditional ideas. Some of the developments that industrial or
utility steam generating personnel should be aware of include:
● Boiler water treatment has undergone many changes. For years, coordi-
nated or congruent phosphate treatment was popular for many boilers.
These programs have been found to have some serious deficiencies and
are being replaced with alternative phosphate programs.
● Oxygenated treatment (OT), where oxygen is deliberately injected into the
boiler feedwater, is becoming very popular in once-through units in the
United States. OT, which was developed in Europe, has been shown to
greatly reduce iron transport from the feedwater system to the boiler.
● Ion exchange is no longer the only reliable method for producing high-puri-
ty water. Other techniques such as reverse osmosis (RO) and electrodialysis
are available for this process. Often, a combination of these techniques, such
as RO plus ion exchange, may be the most economical arrangement.
● Diverse opinions exist regarding chemical oxygen scavenging in boiler
feedwater systems. The reducing environment produced by oxygen scav-
engers is known to influence flow-accelerated corrosion (FAC), in which
the pipe wall gradually erodes. Several catastrophic failures, some of
which have caused fatalities, have occurred in recent years due to FAC.
Yet, the same reducing environment greatly lowers copper dissolution and
transport in those systems that have copper-alloy feedwater heaters.
Hydrazine, the most common and effective oxygen scavenger for many
years, is now listed as a hazardous chemical. However, alternative organ-
ic scavengers (and pH-controlling amines) can break down in boiler sys-
tems to produce organic acids and carbon dioxide, which in turn can
cause corrosion of afterboiler components including turbine blades.
● Combined-cycle or cogeneration systems with heat recovery steam gener-
ators (HRSGs) have become very popular. HRSGs, however, are often

01.qxd 8/8/06 1:50 PM Page 3
Introduction to Steam Generation Water Chemistry Systems 3
designed with two or three steam generating circuits, all at different pres-
sures. Chemical treatment requirements for the various circuits are also
different and may be dependent not only upon the pressure of the circuits,
but also upon the configuration of the HRSG.
These are but some of the issues that I have addressed in this book. Research
still continues on these and many other items, and our knowledge of steam gen-
eration chemistry will only improve in the future.
Table 1-1
Common Steam Generating System Contaminants
Compound Effect on Plant Equipment and Operation
Oxygen Oxygen is often the principal corrodent in water systems. It causes pitting and failures
of pipes and heat exchangers. Oxygen corrosion in boiler systems generates particulates
that travel to the boiler where they precipitate and cause further problems.
Calcium Calcium can combine with a number of anions to form deposits and scales. In cool-
ing water systems the most common deposits include calcium carbonate, calcium
phosphate, and calcium sulfate. These scales retard heat transfer in condensers and
other heat exchangers, and may cause underdeposit corrosion. Calcium scale is even
more problematic in boilers, as the high temperatures greatly accelerate deposition and
corrosion mechanisms.
Magnesium Magnesium will react with carbonates and silicates to form compounds of low solubility.
Magnesium salts that leak into a boiler can react at high temperatures with water to
produce acid. The corrosiveness of acidic solutions is greatly increased at the high
temperatures found in boilers.
Silica Silica combines with a wide variety of elements to produce silicates, or it may form
deposits on its own. Silicates form tenacious deposits in cooling water systems, boiler
tubes, and on turbine blades. The scales are inert to most chemical cleaning solutions
with the exception of hydrofluoric acid. This is an extremely dangerous compound,
and makes prevention of silica deposition even more important.
Organics Organics are usually found in surface waters and are the result of decaying vegetation
or farm runoff. Organics break down in the boiler to form organic acids. The resultant
low pH can be quite deleterious. Organic acids and carbon dioxide produced by
decomposition can carry over to steam turbines and corrode the blades. Organics may
also be found in the condensate return at industrial and cogeneration facilities. These
organics are usually much shorter chained than surface water organics and may require
different treatments.
Suspended
solids Suspended solids, which are also generally found in surface waters, will foul makeup
treatment equipment including reverse osmosis units and ion exchangers. They will
also form deposits in cooling towers and cooling water heat exchangers, a process that
is exacerbated by the presence of microbiological organisms.
Microbes Microbiological fouling is principally troublesome exchanger tubes and cooling tower film
fill. The slime produced by microbiological organisms will trap silt and suspended solids,
further aggravating the situation. Microbes are a leading cause of under-deposit corrosion.

01.qxd 8/8/06 1:50 PM Page 4

02.qxd 8/8/06 1:51 PM Page 5
Chapter 2
Condensate/
Feedwater
Chemistry
Introduction
The preboiler system of a typical utility steam generating unit (Fig. 2–1)
includes a steam surface condenser, several closed tube-in-shell feedwater
heaters, a deaerating feedwater heater, and sometimes an economizer. For indus-
trial systems, feedwater heaters, with the exception of the deaerator, are often
omitted unless the steam drives a turbine. The preboiler circuit condenses the
turbine exhaust steam and prepares the condensate for return to the boiler. The
condensation process significantly improves the efficiency of a unit, as is outlined
in greater detail in Supplement 2–1 at the end of this chapter.
The potential for contaminant introduction to a steam generating plant is
greatest in the preboiler system, especially at the condenser or via condensate
return from an industrial process. This makes chemistry control and monitoring
of condensate and feedwater extremely important.

02.qxd 8/8/06 1:51 PM Page 6
Figure 2-1
>
MAIN STEAM TO TURBINE
SUPERHEATER ECONOMIZER
TURBINE
BOILER
CONDENSER
VENT
>
>
>
STEAM >
DEAERATOR
FEEDWATER HEATERS FEEDWATER HEATERS
Preboiler/Boiler/Afterboiler Schematic
Condensate/Feedwater System
Construction Materials
A variety of materials have been used for feedwater heater and condenser
tubes. The most common include 90-10 and 70-30 copper nickel, 304 stainless
steel, carbon steel, Admiralty metal (70% copper and 29% to 30% zinc depend-
ing on whether 1% arsenic or tin has been added for increased corrosion resis-
tance), and Monel (70% nickel, 30% copper). In previous years, copper alloys
were widely selected due to the excellent heat transfer properties of these mate-
rials. Recently, as copper corrosion and its effects on downstream components
have become better recognized, the use of copper alloys has declined. Stainless
steel is becoming favored for feedwater heater tubes; and stainless steel, titanium,
or even the duplex alloys are being used for new condensers.
The behavior of these materials in solution significantly influences conden-
sate/feedwater chemistry guidelines and treatment methods. Regardless of the
material, one of the toughest challenges for a plant chemist is control of dissolved
oxygen.
Condensate Chemistry—Dissolved Oxygen

When a boiler is first placed in operation, the carbon steel feedwater pipe
walls develop a tightly bound layer of magnetic iron oxide (magnetite, Fe3O4) via
the following reaction:

02.qxd 8/8/06 1:51 PM Page 7
Condensate/Feedwater Chemistry 7
3Fe + 4H2O → Fe3O4 + 4H2 ↑ (2.1)
Magnetite is very dark, and in a properly treated system will appear as a

dense, black layer upon the pipe or boiler tube surface. Copper alloys also devel-
op a protective film when placed into service, which consists of a layer of
cuprous oxide (Cu2O):
2Cu + H2O → Cu2O + H2 ↑ (2.2)
Oxygen that enters the condensate system will oxidize the protective layers
on iron and copper to films that are much less stable. In the case of mild steel,
oxygen converts magnetite to ferric oxide (Fe2O3), which is not protective.
2Fe3O4 + ¹/₂O2 → 3Fe2O3 (2.3)
Fe2O3 is rust, and, like rust produced by atmospheric corrosion of iron, is

brownish-red in color and completely unprotective of the base metal. Oxygen
attack of carbon steel is quite detrimental and can cause a variety of problems.
Nodules of corrosion products and pits may form at the corrosion sites. Corrosion
products will enter the solution and be transported downstream to the boiler,
where higher heat loads cause the particles to precipitate. The deposits can set up
corrosion cells on the boiler tubes. The deposits also inhibit heat transfer across
the tube boundary, which reduces boiler efficiency and can even shorten tube life
due to overheating. The frequency of boiler chemical cleaning is, in most cases,
influenced more by iron oxide deposition than any other mechanism.
Oxygen converts cuprous oxide to cupric oxide (CuO).
Cu2O + ¹/₂O2 → CuO (2.4)
In the presence of a complexing agent such as ammonia, the oxidized cop-

per will be solvated and removed.
Cu+2 + 4NH3 → Cu(NH3)4+2 (2.5)
Other complexing agents include chloride and sulfate, however, these are
usually at extremely low concentrations and do not participate in copper corro-
sion mechanisms. Oxygen/ammonia attack is much more common, primarily
because ammonia or organic amines are the preferential choice for feedwater pH
conditioning.

02.qxd 8/8/06 1:51 PM Page 8
Oxygen is an aggressive corrodent. Control of inleakage and removal of dissolved

oxygen are very important, although, as will be shown, a treatment program
based on the deliberate injection of oxygen is becoming quite popular in certain
systems.
Carbonic Acid Corrosion in

Industrial Steam Generating Systems
In industrial systems, carbonic acid corrosion of condensate lines is often a
more serious problem. This is due to the formation of carbon dioxide in the boil-
er. CO2 formation primarily occurs when partial demineralization (softening) is
employed for makeup water production to remove the hardness ions, calcium,
and magnesium. The makeup system does not remove bicarbonates, which
decompose in the boiler to produce carbon dioxide:
HCO3- + heat → CO2↑ + OH- (2.6)
Carbon dioxide carries over with the steam and then redissolves in conden-
sate return lines. Inorganic chemists debate over whether CO2 actually forms an
acid or whether it exists as a discrete hydrated molecule, but the net effect of this
dissolution is shown in the following equation:
CO2 + H2O ⇔ H2CO3 (2.7)
The absorption of carbon dioxide into the condensate drives the pH down-
ward. The acid solution directly attacks the pipe walls via the formation of fer-
rous bicarbonate:
Fe + 2H2CO3 → Fe(HCO3)2 + H2↑ (2.8)
Any oxygen present in the feedwater will oxidize the ferrous bicarbonate to
ferric oxide and liberate carbon dioxide to repeat the process.
Oxygen, carbon dioxide, and ammonia are the most common corrodents in
steam generating systems. Methods to remove or treat these compounds are out-
lined in the following sections.

02.qxd 8/8/06 1:51 PM Page 9
Figure 2-2
Condenser Tube Map

02.qxd 8/8/06 1:51 PM Page 10
Figure 2-3
Condensate Dissolved Oxygen Concentrations vs. Air Inleakage Copyright © 1986. Electric Power Research
Institute. EPRI CS-4629. Interim Consensus Guidelines on Fossil Plant Cycle Chemistry. Reprinted with permission.
Mechanical Removal of Dissolved Oxygen

and Other Gases
Air primarily enters a steam generating system at locations around the steam
surface condenser or in makeup that comes from an atmospherically vented stor-
age tank. Condensers operate under a very strong vacuum, which in winter may
be as low as 1in(Hg) above absolute. Air inleakage is virtually impossible to pre-
vent. Prime spots for air inleakage include the expansion joint between the tur-
bine and condenser, penetrations of heater drips lines into the condenser shell,
turbine seals and explosion diaphragms, and condensate pump seals.
Although air inleakage cannot be totally eliminated, the effects can be con-
trolled under normal conditions and a combination of mechanical and chemical
means are employed for this purpose. The first step in the treatment is extraction
of dissolved gases in the condenser. Condensers are normally equipped with one
or more air removal compartments. An air removal compartment consists of a
metal shroud, open at the bottom, that extends from the front to the rear tube
sheet and encloses a group of the condenser tubes. (Fig. 2–2 illustrates the con-
denser tube map that plant personnel at my former utility use to keep track of
plugged tubes. The map shows an end view of the air removal section.) A vacu-
um, generated by external pumps or steam-jet ejectors, is applied within the

02.qxd 8/8/06 1:51 PM Page 11
Figure 2-4
Change in Dissolved Oxygen Concentration with Unit Load Copyright © 1986. Electric Power Research
Institute. EPRI CS-4629. Interim Consensus Guidelines on Fossil Plant Cycle Chemistry. Reprinted with permission.
shroud to pull gases into the compartment, upon which they are extracted and
vented to atmosphere. The shroud is situated around a block of tubes to con-
dense steam that may enter.
The Heat Exchange Institute (HEI) has established a design guideline of 7
ppb dissolved oxygen in condensate from a properly operating condenser.
However, excess air inleakage can overload the air removal system, which, as
Figure 2–3 shows, dramatically increases the dissolved oxygen concentration.
Several techniques are available to determine if excess air inleakage is a prob-
lem. Simple flow monitoring, wherein the extraction line from the air removal sec-
tion is equipped with an air flow rotameter, is one such method. A guideline estab-
lished by the HEI suggests that under normal conditions, air removal through the
extraction system should average about 1 standard cubic foot per minute (SCFM)
per 100 MW of capacity. Significantly higher values than this may indicate prob-
lems, although each unit should be evaluated on a case-by-case basis. It is more
important to establish baseline conditions, which can then be used for comparison
during times of suspected upsets.
Plant chemists can often detect excess inleakage by analyses of dissolved
oxygen (DO) levels in the condensate pump discharge. However, this data must

02.qxd 8/8/06 1:51 PM Page 12
Figure 2-5
External View of a Deaerator. Photo taken with permission of the Kansas City, Kansas, Board of Public Utilities,
Quindaro Power Station.
be carefully evaluated, because unit load affects condensate DO concentrations.

As unit load decreases, the performance of the air removal system changes and
actually causes an increase in dissolved oxygen. Figure 2–4 illustrates this effect.
If the condenser air removal system becomes overloaded or malfunctions,
the excess air will form a film on the condenser tubes. The air acts as an insula-
tor, just as if the tubes were fouled on the water side. Plant personnel who regu-
larly monitor condenser operation should be able to detect the decline in per-
formance, because reduced heat transfer will cause an increase in turbine back-
pressure. The BASIC program outlined in Supplement 2–2 provides an excellent
method for detecting loss of heat transfer, including those caused by excess air
inleakage. Case Histories 2–1 and 2–2 illustrate unique air inleakage problems
that this program detected.
The air removal compartment concentrates other compounds besides oxy-
gen, one of which is ammonia. Corrosion of copper-alloy tubes by the reactions
shown in Equations 2.4 and 2.5 is a common occurrence. Case History 2–3 out-
lines a classic example of this type of corrosion.
Downstream from the condenser, the deaerator (DA) provides additional
mechanical protection from the effects of dissolved gases. A DA (Fig. 2–5) con-
sists of a steam scrubbing vessel and storage tank. Tray-type scrubbing vessels are
very common. In this design, the scrubbing compartment contains a stacked
series of perforated trays. Condensate is introduced into the top of the compart-
ment and flows downward through the trays, while steam is injected into the

02.qxd 8/8/06 1:51 PM Page 13
condensate spray. The steam scrubs the

Table 2–1
condensate and raises its temperature, lib-
Oxygen
erating dissolved gases including oxygen, Concentration
whose solubility decreases with increasing Temperature (˚F) (cc per liter)
30 10.1
temperature (Table 2–1). The liberated 50 7.8
gases are vented through the top of the 70 6.2
deaerator. 90 5.1
110 4.4
Proper deaerator performance is 130 3.8
dependent upon several factors including 150 3.1
correct alignment of the trays, evenly dis- 170 2.4
190 1.5
tributed condensate and steam flow, and 210 0.1
sufficient venting. The vent valves are set Dissolved Oxygen Concentrations as a
to remove the maximum amount of dis- Function of Temperature.
solved gases while minimizing loss of steam. A rule-of-thumb guideline suggests
that steam losses should be about 0.1% of condensate flow, and that the visible
plume from the deaerator vent(s) should be about 1 to 2 feet in length. The con-
densate in the storage tank should be within 2° or 3°F of saturation temperature.
Deaerators are typically designed to reduce dissolved oxygen levels to 7 ppb and
to store 10 minutes worth of supply water to the boiler. Conditions out of spec-
ification indicate poor performance in the deaerating chamber. A common prob-
lem in tray-type deaerators is misalignment of the trays. Low steam flow will
affect gas removal and heating efficiency in tray-type and spray-type deaerators.
Too little venting inhibits gas removal, while too much allows excess steam to
escape.
Chemical Control of Dissolved Oxygen

Mechanical methods are not the sole means employed to control dissolved
oxygen; chemicals act as a supplement to the mechanical systems. These chemi-
cals are known as oxygen scavengers.
Oxygen Scavengers
One of the first practical chemicals to be used was sodium sulfite (Na2SO3).
Sodium sulfite reacts with oxygen to produce sodium sulfate:
2Na2SO3 + O2 → 2Na2SO4 (2.9)
Sodium sulfite has a molecular weight almost four times higher than that of
oxygen and reacts in a 2 to 1 molar ratio, so theoretically 8 parts per million
(ppm) of Na2O3 are needed to remove 1 ppm of oxygen. However, sulfite resid-
uals are often maintained at 30 ppm or higher to provide adequate protection.

02.qxd 8/8/06 1:51 PM Page 14
Figure 2-6a
Reaction Time of Hydrazine with Oxygen Copyright © 1986. Electric Power Research Institute. EPRI CS-
4629. Interim Consensus Guidelines on Fossil Plant Cycle Chemistry. Reprinted with permission.
The primary advantages of sodium sulfite are that it is a common and easi-
ly obtained chemical, is nontoxic, and can be used for water treatment where
steam is extracted for food processing or other FDA-regulated applications.
Sodium sulfite is primarily used in low-pressure industrial boilers (<600 psig)
because it adds too many dissolved solids to high-pressure boiler water. Also, in
boilers that operate above 900 psig, sodium sulfite will thermally decompose to
produce hydrogen sulfide (H2S) and sulfur dioxide (SO2), both of which are
quite corrosive.
Na2SO3 + H2O → 2NaOH + SO2↑ (2.10)

4Na2SO3 + 2H2O → 3Na2SO4 + 2NaOH + H2S↑ (2.11)
For utility and industrial boilers that operate at pressures above 900 psig,
alternative chemicals are more suitable for oxygen scavenging. The workhorse for
many years has been hydrazine (N2H4), which reacts with oxygen as follows:
N2H4 + O2 → 2H2O + N2↑ (2.12)
Hydrazine proved advantageous because it does not add any dissolved

solids to the feedwater, it reacts with oxygen in a one-to-one weight ratio, and it
is supplied in liquid form at 35% concentration.
A primary benefit of hydrazine is that it will passivate oxidized areas of pip-
ing and tube materials as follows:
N2H4 + 6Fe2O3 → 4Fe3O4 + N2↑ + 2H2O (2.13)

02.qxd 8/8/06 1:51 PM Page 15
Figure 2-6b
Effect of pH on the Hydrazine-Oxygen Reaction for a Reaction Time of 0.13 Minutes at 300˚F in
Carbon-Steel Tubing Copyright © 1986. Electric Power Research Institute. EPRI CS-4629. Interim Consensus
Guidelines on Fossil Plant Cycle Chemistry. Reprinted with permission.
N2H4 + 4CuO → 2Cu2O + N2↑ + 2H2O (2.14)
Based on molecular weights, 1 part of N2H4 will theoretically passivate 30

parts of Fe2O3 and 10 parts of CuO. Hydrazine residuals are maintained at much
lower levels than sodium sulfite, typically in a range from 20 to 100 parts per bil-
lion (ppb). (This is provided that the hydrazine does not remove all traces of dis-
solved oxygen. A slight oxygen residual is needed to prevent flow-accelerated
corrosion, which is described later in this chapter.)
Round-the-clock feed of hydrazine, or an alternative oxygen scavenger, is
very important to properly protect feedwater and boiler components. Gaps in
coverage can allow oxygen corrosion. Hydrazine performance is greatly
enhanced by increased temperature and pH (Figs. 2–6a, 2–6b). However, pre-
cautions must be taken to prevent overfeed of the chemical, as excess hydrazine
will begin to decompose at temperatures above 400°F to form ammonia:
3N2H4 → N2↑ + 4NH3 (2.15)
This can interfere with the supplemental ammonia or amine compounds

that are used for pH control. Breakdown of excess hydrazine can raise ammonia
concentrations to levels that significantly increase copper-alloy corrosion.
The hydrazine reactions shown in Equations 2.9, 2.10, and 2.11 are
enhanced by the addition of a catalyst, which increases the reaction rate at lower

02.qxd 8/8/06 1:51 PM Page 16
Figure 2-7 temperatures. The most common catalyst is hydroquinone

(C6H6O2, Fig. 2–7), which is itself an oxygen scavenger. When
OH hydrazine catalysts were introduced, performance improved
enough that many plant personnel switched the injection
+ + + from
point + the deaerator storage tank or DA outlet to the con-
densate pump discharge. This provides protection to the con-
OH densate piping and low-pressure feedwater heaters. The great-
Structure of est drawback to injection at the condensate pump discharge is
Hydroquinone.
that some of the hydrazine is removed in the deaerator.
Given the simplified chemistry of hydrazine, it
Figure 2-8
would appear to be the ideal oxygen scavenger.
Unfortunately, hydrazine is considered to be a potential O
carcinogen and is now registered as a hazardous com-
+ H3 N2 C N2 H 3 +
pound. Handling procedures have become much more
stringent. This difficulty has in part led to the develop-
Structure of
ment of other treatment chemicals. Many of the major Carbohydrazide.
water treatment companies provide alternative oxygen
scavengers, the principal ones being carbohydrazide (N4H6CO, Fig. 2–8), hydro-
quinone, and methyl ethyl ketoxime (C4H9NOH, Fig. 2–9). All of these products
are also metal passivators.
Carbohydrazide is a derivative of hydrazine and
Figure 2-9 actually breaks down to hydrazine and carbon dioxide
as it is heated in the feedwater system. The chief
+ H3 C + of carbohydrazide is that it is not considered
advantage
C NOH
H5 C 2 to be hazardous like hydrazine and can be handled
more easily. A drawback is that carbon dioxide is pro-
Structure of Methyl
duced when carbohydrazide reacts with oxygen or pas-
Ethyl Ketoxime.
sivates metals:
N4H6CO + 2O2 → 2N2↑ + 3H2O + CO2↑ (2.16)

N4H6CO + 12 Fe2O3 → 8Fe3O4 + 2N2↑ + 3H2O + CO2↑(2.17)
N4H6CO + 8CuO → 4Cu2O + 2N2↑ + 3H2O + CO2↑ (2.18)
Carbohydrazide has performed well in some units, and is reported to be a

better metal passivator than uncatalyzed hydrazine at low temperatures. Some
facility operators use carbohydrazide for wet layups of boilers due to its effec-
tiveness at ambient temperatures.
Hydroquinone is a benzene-derivative compound that has been used as a
stand-alone oxygen scavenger and as a catalyst for hydrazine. Hydroquinone
reacts with oxygen to produce benzoquinone.

02.qxd 8/8/06 1:51 PM Page 17
C6H6O2 + ¹/₂O2 → C6H4O2 + H2O (2.19)
Due to its organic structure, hydroquinone and its oxidized products can
break down to a variety of organic acids and carbon dioxide in a boiler system.
Methyl ethyl ketoxime is an organic scavenger that reacts with oxygen to
produce methyl ethyl ketone (MEK).
C4H9NOH + ¹/₂O2 → C4H9O + N2O↑ + H2O (2.20)
Reported advantages of methyl ethyl ketoxime include a higher thermal sta-

bility than other scavengers and a much higher distribution ratio in steam. The
latter allows the compound to cycle through the system. However, methyl ethyl
ketoxime and MEK can break down to smaller organics, carbon dioxide, and
nitrogen compounds.
All of these products (hydrazine, carbohydrazide, hydroquinone, and
methyl ethyl ketoxime) have performed well in some applications, and less well
in others. A clear cut choice is not always possible. For instance, one full-scale
test of all four compounds showed that the relative scavenging or passivating
properties of each chemical varied depending on location in the condensate/feed-
water system. One single product did not out-perform another in all locations or
for all functions.
Some experts in the electric utility industry oppose the use of organic oxy-
gen scavengers because of the organic acid and carbon dioxide decomposition
products. This is a reasonable view, as organic acids and CO2 are known to cause
corrosion of turbine blades and afterboiler components. However, the concen-
tration at which the attack becomes serious is less well known. Researchers are
continuing their efforts to determine the species, quantities, and effects of organ-
ic acids and CO2 in the turbine. The introduction of organics by oxygen scav-
engers may be slight in comparison to those introduced from pH-conditioning
amines, which are typically maintained at higher concentrations in the feedwater
than the oxygen scavenger.
Safe Hydrazine Feed Systems

It is possible to set up a hydrazine feed system that does not expose work-
ers to the compound. In one such system at a western utility, the hydrazine solu-
tion is supplied in portable containers. The container is connected to a perma-
nent metering pump and distribution line, which transport precise dosages of the
chemical in neat form to the feedwater system. A variation of this concept is a
system in which the primary feed tank supplies a closed day tank, which is vent-
ed to the outside atmosphere. A known volume of hydrazine is introduced to the

02.qxd 8/8/06 1:51 PM Page 18
Figure 2-10
External View of a High-Pressure Feedwater Heater. Photo taken with permission of the Kansas City,
Kansas, Board of Public Utilities, Quindaro Power Station.
day tank, is diluted with water, and is then pumped to the system.
The portable feed concept is gaining great acceptance. The water treatment
firm is responsible for delivering full tanks to the plant site and hauling away the
empty vessels. Plant personnel do not have to deal with drums and related haz-
ardous waste disposal requirements. The capacity of portable containers is typi-
cally within a range of 200 to 400 gallons. A method to minimize handling is to
stack one portable feed container on top of a primary feed container. The top
container is plumbed such that it drains into the bottom vessel. When the top
container empties, it is replaced with a full vessel. This arrangement provides two
distinct advantages. First, the bottom container can be permanently piped to the
metering pump or connected to a day tank. Second, the top container drains
completely without compromising the performance of the system. No residual
chemical remains in the vessel.
Industrial Oxygen Scavenging

In the food, beverage, and pharmaceutical industries, process steam comes
under FDA regulations. Besides sodium sulfite, an approved oxygen scavenger
for these applications is erythorbic acid (C6H8O6). Erythorbic acid is an isomer
of Vitamin C, and is reported to react very quickly with oxygen. Erythorbic acid
has been used in electric utility boilers, but like the other organic oxygen scav-
engers, it will break down at high temperatures to form organic acids and CO2.

02.qxd 8/8/06 1:51 PM Page 19
Flow-Accelerated Corrosion
A relatively new phenomenon, at least in terms of it being communicated to
steam generation personnel, is flow-accelerated corrosion (FAC). FAC has raised
additional questions about oxygen scavenging. FAC is produced by a combina-
tion of chemical and mechanical factors, and occurs in high turbulence regions
of the feedwater system with strongly reducing environments. These locations
include elbows in feedwater lines, economizer inlet headers, and high-pressure
heater drains. In simple terms, FAC develops when oxygen is completely
removed by the scavenger. This inhibits formation of the protective magnetite
layer on the pipe surface. In high turbulence zones, the flow gradually
erodes/corrodes the base metal until the wall becomes too thin to withstand the
fluid pressure. Several catastrophic failures have occurred within the past few
years and in some cases have caused fatalities. This is quite serious, and it is my
hope that this book will alert readers to this potential problem. The Electric
Power Research Institute (EPRI) is making a strong effort to inform their mem-
bers and others about FAC. Table 2-2
Several methods have
Relative Basicity
been proposed to combat Kb x 10
FAC in scavenger-treated
Compound 72˚F 298˚F 338˚F
feedwater systems. One Ammonia 20.6 6.9 4.6
method is to reduce the oxy- Cyclohexylamine 489 61 32
gen-scavenger feed, such that Diethylaminoethanol 68 11.3 9.2
a 1 to 2 ppb oxygen residual Morpholine 3.4 4.9 3.8
remains in the feedwater. This Basicities of Various Feedwater pH Conditioners.
helps preserve the magnetite Source: Betz Handbook of Industrial Water. Conditioning, Ninth
Edition. BetzDearborn, Inc., Horsham, PA.
layer without subjecting the
system to corrosive levels of oxygen. To inhibit FAC of high-pressure heater
drains, EPRI recommends that high-pressure feedwater heater (Fig. 2–10) vents
be closed. This, however, has caused problems at some utilities and must be eval-
uated on a case-by-case basis.
The balance between oxygen scavenging and prevention of FAC presents
one of the greatest difficulties in mixed-metallurgy feedwater systems, where
copper-alloy corrosion is greatly dependent on the oxidation reduction potential
(ORP) of the solution. This particular aspect of feedwater chemistry is discussed
below.
Copper-Alloy Corrosion and Preventive Measures

Corrosion of copper alloys and the effects on steam generating systems have
been receiving a great deal of attention. Foremost is that cuprous and cupric
oxides carry over with boiler steam and deposit in the afterboiler, and in partic-

02.qxd 8/8/06 1:51 PM Page 20
ular on the high-pressure turbine blades. Reports abound of electric utility units
whose output decreased over an operating period because of copper deposition
on the turbine blades. Copper corrosion and deposition also cause other diffi-
culties. For example, copper deposition in boiler tubes can greatly increase the
difficulty of a boiler chemical clean. The tube deposits may become layered with
copper and magnetite, which can require a multistaged chemical cleaning
process for complete removal. Copper inclusion in deposits can also increase the
potential for under-deposit corrosion.
As Equation 2.4 illustrated, copper corrosion is initiated when copper metal
or the protective cuprous oxide film is oxidized to cupric oxide. For this to hap-
pen, the oxidation potential (as measured in millivolts or volts) of the solution
must be greater than zero. Continual feed of the oxygen scavenger will maintain
a reducing environment and keep the ORP well below zero. However, a reducing
environment is the primary triggering mechanism for flow-accelerated corrosion.
Copper corrosion is exacerbated in cycling units. Several mechanisms may
be at work here. Mechanical stresses imposed during frequent load changes or
startups and shutdowns will fracture the cuprous and cupric oxide films on
heater tubes and cause corrosion and spalling. Oxygen corrosion of unprotected
feedwater heaters during an outage has caused severe corrosion and exfoliation
of heater tubes. In fact, this problem can be so serious that the heater steam-side
should be blanketed with steam or nitrogen during extended outages. More
details about off-line corrosion protection are presented in chapter 3.
The upshot of these difficulties with copper corrosion is that many experts
now recommend that feedwater heater tubes be fabricated from stainless steel, or
that existing copper-alloy heaters be replaced. Sometimes the latter recommen-
dation may not be economical, at least in the short term. Good water treatment
practices are a fallback for such situations.
Additional Corrosion Control Requirements

While oxygen and carbon dioxide are primary corrodents in utility and
industrial steam generating systems, the pH of the feedwater also is an important
factor. Feedwater piping and heat exchanger tubes exhibit minimal corrosion at
a mildly alkaline pH. For a system containing all steel metallurgy the optimum
pH range is 9.0 to 9.6, while for a system with a mixed metallurgy of iron and
copper, the recommended pH range is 8.8 to 9.1. Ammonia or volatile amines
provide this needed basicity. Table 2–2 lists some of the most common neutral-
izing agents. The choice of which compound to use depends upon the nature
and pressure of the steam generating system, and the characteristics of the pH-
conditioning chemical. Of the compounds listed in the table, cyclohexylamine is
most basic, while morpholine is significantly less basic. The basicity of diethy-

02.qxd 8/8/06 1:51 PM Page 21
laminoethanol and ammonia lies in between. Thus, the chemist has a number of
choices for finding a product that maintains condensate/feedwater pH within rec-
ommended guidelines.
Amines will decompose to produce ammonia in feedwater. Whether the
ammonia is generated by direct ammonia feed or amine decomposition, a gener-
al rule of thumb suggests that ammonia concentrations be limited to 0.5 ppm in
systems with copper-alloy tubed heaters. Even this level may be too high where
frequent inleakage of air is a problem.
In high-pressure utility boilers, where the steam is quite pure, decomposi-
tion of amines can potentially introduce unwanted organic acids and CO2 to the
turbine. For this reason, ammonia is widely recommended as the best pH-con-
ditioning chemical. The situation may be vastly different in an industrial boiler
where bicarbonate decomposition causes heavy carryover of CO2 to condensate
lines. Amines can be a good product for neutralization of the carbon dioxide. A
particular aspect of importance is the amine distribution ratio. This is the per-
centage of amine that carries over with the steam versus that which remains in
the boiler water. The distribution ratio varies with the product and with the pres-
sure of the boiler (Fig. 2–11). Selection of the neutralizing chemical based on dis-
tribution ratio can be very important. If protection of afterboiler condensate lines
is required, an amine with a high steam-to-liquid distribution ratio is best.
Where corrosion prevention in the boiler is more critical, a more appropriate
neutralizing amine is one whose distribution ratio allows most of it to remain in
the boiler water. Often, a blend of two amines is used to provide universal pro-
tection.
This section would not be complete without a brief discussion of filming
amines. These products are long chain amines (18 to 20 carbon atoms) that do
not neutralize acids but rather form a film on the pipe wall. The amines provide
a physical barrier between the material surface and the process liquid. The amine
group attaches to the metal surface while the nonpolar organic portion acts as the
film. Octadecylamine [CH3(CH2)16CH2NH2] is the most popular of the filming
amines. Filming amines are primarily used for protection of condensate return
lines at industrial facilities. They must be employed with caution, however.
Overfeed can introduce excess organics to the boiler water. Additionally, filming
amines do not work well in high velocity condensate return lines, because they
can be easily stripped from the metal surface.
Oxygenated Treatment
With all of the previous discussion about the deleterious effect of dissolved
oxygen, it may come as a surprise to some readers that a feedwater treatment
technology has emerged in which oxygen is deliberately injected into the con-

02.qxd 8/8/06 1:51 PM Page 22
Figure 2-11
Distribution Ratio of Two Common Neutralizing Amines. Reprinted with permission of

BetzDearborn, Inc., Horsham, PA.
densate/feedwater system. This is known as oxygenated treatment (OT). OT pro-

grams in which the feedwater is not conditioned to raise the pH are known as
neutral water treatment (NWT). If ammonia is added for pH control, the pro-
gram is designated as combined water treatment (CWT).
OT was developed in Germany about 25 years ago for replacement of all-
volatile treatment (hydrazine/ammonia) in once-through steam generating units.
The program was adopted by other European utilities, and now is gaining great
acceptance at once-through utilities in the United States. The treatment requires
the controlled injection of oxygen or hydrogen peroxide into the condensate/feed-
water system. Figure 2–12 illustrates common injection points. In CWT pro-
grams, which are most common in the United States, oxygen is dosed to maintain

02.qxd 8/8/06 1:51 PM Page 23
Figure 2-12
Condensate Condenser
High Pressure Low Pressure Polisher
Heaters Deaerator Heaters
BFP
To Boiler V
Oxygen Oxygen
Typical Oxygen Injection Points for an Oxygenated Treatment Program.
Figure 2-13
FeOOH
Fe3O4 Fe3O4
BASE METAL BASE METAL
DEOXYGENATED OXYGENATED
TREATMENT TREATMENT
Oxide Layer Formed in an Oxygenated Treatment Program.
a 50 to 150 ppb residual. Ammonia is added to raise the pH within a range of 8.0
to 8.5. Typically, 20 to 70 ppb of ammonia will produce this pH.
The chemistry of oxygenated treatment is very interesting and explains why
the program has become popular. In conventional AVT units, even with very
good chemistry programs, some iron oxides are still generated in the condensate
or feedwater system. The iron oxides are transported to the boiler, where they
precipitate on the tube walls. With controlled oxygen injection however, the base
layer of magnetite becomes overlayed and interspersed with an even tighter film
of ferric oxide hydrate (FeOOH). (See Fig. 2–13). This compact layer is more sta-
ble than magnetite and releases very little dissolved iron or suspended iron-oxide
particles to the fluid. Some utilities that switched from AVT to OT have reported
that dissolved feedwater iron concentrations, which were often 10 ppb or high-
er on oxygen-scavenger programs, dropped to as low as 1 to 2 ppb once the OT
program was fully established.
The key to an OT program is control. Makeup water must be highly puri-
fied because deposits would cause differential oxygen cells to form, which would
in turn cause under-deposit corrosion and pitting. Very pure feedwater in once-
through units is usually a given, as these steam generating systems are always
equipped with condensate polishers. OT cannot be used in systems that contain
copper-alloy feedwater heater tubes, as the corrosion would be too great.
One benefit of OT is that the feedwater pH is a unit or unit and a half lower
than that in AVT programs. This is due to the stability of the FeOOH film. The
lower pH can greatly increase condensate polisher run lengths.

02.qxd 8/8/06 1:51 PM Page 24
Table 2-3
Suggested Water Chemistry Limits

Industrial Watertube, High Duty,
Primary Fuel Fired, Drum Type
Makeup water percentage: Up to 100% of feedwater
Conditions: Includes superheater, turbine drives, or process restriction on steam purity
Saturated steam purity target: See tabulated values below.
Drum Operation psig 0-300 301-450 451-600

Pressure (1)(11) (MPa) (0-2.07) (2.08-3.10) (3.11-4.14)
Feedwater (7)
Dissolved oxygen ppm (mg/1) O
- measured before chemical
oxygen scavenger addition (B) <0.007 <0.007 <0.007
Total iron ppm (mg/l) Fe ≤0.1 ≤0.05 ≤0.03
Total copper ppm (mg/l) Cu ≤0.05 ≤0.025 ≤0.02
Total hardness ppm (mg/l)* ≤0.3 ≤0.3 ≤0.2
pH @ 25˚C 8.3-10.0 8.3-10.0 8.3-10.0
Chemicals for preboiler system
protection NS NS NS
Nonbolatile TOC ppm (mg/l) C (6) <1 <1 <0.5
Oily matter ppm (mg/l) <1 <1 <0.5
Boiler Water
Silica ppm (mg/l) SiO2 ≤150 ≤90 ≤40
Total alkalinity ppm (mg/l)* <350(3) <300(3) <250(3)
Free OH alkalinity ppm (mg/l)* (2) NS NS NS
Specific conductance (12)
mhos/cm (s/cm) 25˚C
without neutralization 5400-1100(5) 4600-900(5) 3800-800(5)
Total Dissolved Solids in Steam (9)

TDS (maximum) ppm (mg/l) 1.0-0.2 1.0-0.2 1.0-0.2
*as CaO3
NS = not specified
ND = not detectable
VAM = Use only volatile alkaline materials upstream of attemperation water source (10)
Reproduced from Consensus on Operating Practices for the Control of Feedwater and Boiler Water
Chemistry in Modern Industrial Boilers with permission from the American Society of Mechanical Engineers.
601-750 751-900 901-1000 1001-1500 1501-2000

(4.15-5.17) (5.18-6.21) (6.22-6.89) (6.90-10.34) (10.35-13.79)
<0.007 <0.007 <0.007 <0.007 <0.007

≤0.025 ≤0.02 ≤0.02 ≤0.01 ≤0.01
≤0.02 ≤0.015 ≤0.01 ≤0.01 ≤0.01
≤0.2 ≤0.1 ≤0.05 ND ND
8.3-10.0 8.3-10.0 8.8-9.6 8.8-9.6 8.8-9.6
NS NS VAM VAM VAM

<0.5 <0.5 <0.2 <0.2 <0.2
<0.5 <0.5 <0.2 <0.2 <0.2
≤30 ≤20 ≤8 ≤2 ≤1
<200(3) <150(3) <100(3) NS(4) NS(4)
NS NS NS ND(4) ND(4)
1500-300(5) 1200-200(5) 1000-200(5) ≤150 ≤80
0.5-0.1 0.5-0.1 0.5-0.1 0.1 0.1
Untitled-1.indd 25 4/30/08 1:48:56 PM

02.qxd 8/8/06 1:51 PM Page 26
Table 2-3 Notes

(1) With local heat fluxes >1.5 x 105 Btu/hr/ft2 (>473.2 kW/m2), use valves for at least
the next higher pressure range.
(2) Minimum hydroxide alkalinity concentrations in boilers below 900 psig (6.21 MPa)
must be individually specified by a qualified water treatment consultant with regard
to silica solubility and other components of internal treatment.
(3) Maximum total alkalinity consistent with acceptable steam purity. If necessary,
should override conductance as blowdown control parameter. If makeup is dem-
ineralized quality water and boiler operates at less than 1000 psig (6.89 MPa) drum
pressure, the boiler water conductance should be that in table for 1001-1500 psig
(6.9-10.34 MPa) range. In this case, the necessary continuous blowdown will usual-
ly keep these parameters below the tabulated maximum values. Alkalinity values in
excess of 10% of specific conductance values may cause foaming.
(4) Not detectable in these cases refers to free sodium or potassium hydroxide alkalini-
ty. Some small variable amount of total alkalinity will be present and measurable with
the assumed congruent or coordinated phosphate-pH control or volatile treatment
employed at these high pressure ranges.
(5) Maximum values are often not achievable without exceeding maximum total alkalini-
ty values, especially in boilers below 900 psig (6.21 MPa) with >20% makeup of water
whose total alkalinity is >20% of TDS naturally or after pretreatment by lime-soda, or
sodium cycle ion exchange softening. Actual permissible conductance values to achieve
any desired steam purity must be established for each case by careful steam purity mea-
surements. Relationship between conductance and steam purity is affected by too many
variables to allow its reduction to a simple list of tabulated values.
(6) Nonvolatile TOC is that organic carbon not intentionally added as part of the water
treatment regime.
(7) Boilers below 900 psig (6.21 MPa) with large furnaces, large steam release space, and
internal chelant, polymer, and/or antifoam treatment can sometimes tolerate higher lev-
els of feedwater impurities that those in the table and still achieve adequate solution.
Alternatives must be evaluated as to practicality and economics in each individual case.
(8) Values in the table assume existence of a deaerator.
(9) Achievable steam purity depends on many variables, including boiler water total
alkalinity and specific conductance as well as design of boiler steam drum internals
and operating conditions [Note (5)]. Since boilers in this category require a relative-
ly high degree of steam purity for protection of the superheaters and turbines, more
stringent steam purity requirements such as process steam restrictions on individual
chemical species or restrictions more stringent than 0.1 ppm (mg/l) TDS turbine
steam purity must be addressed specifically.
(10) As a general rule, the requirements for attemperation spray water quality are the same
as those for steam purity. In some cases boiler feedwater is suitable; however, frequently
additional purification is required. In all cases the spray water should be obtained from
a source that is free of deposit forming and corrosive chemicals such as sodium hydrox-
ide, sodium sulfite, sodium phosphate, iron, and copper. The suggested limits for spray
water quality are < 30 ppb (g/l) TDS maximum, < 10 ppb (g/l) Na maximum, <20
ppb (g/l) SiO2 maximum, and it should be essentially oxygen free.
(11) Low pressure boilers frequently use feedwater that is suitable for user in higher pres-
sure boilers. In these cases the boiler water chemistry limits should be based on the
pressure range that is mos consistent with the feedwater quality.
(12) Conversion from ppm (mg/l) TDS values in the ABMA standards [12] used a factor
of 0.65.
02.qxd 8/8/06 1:51 PM Page 27
Table 2-4

Primary Fuel Fired, Drum Type
Conditions: No superheater, turbine drives, or process restriction on steam purity
Steam purity (7): 1.0 ppm (mg/l) TDS maximum
Drum Operation psig 0-300 301-600

Pressure (MPa) (0-2.07) (2.08-4.14)
Feedwater (7)
Dissolved oxygen ppm (mg/1) O2
- measured before chemical
oxygen scavenger addition (1)(2) <0.007 <0.007
Total iron ppm (mg/l) Fe ≤0.1 ≤0.05
Total copper ppm (mg/l) Cu ≤0.05 ≤0.025
Total hardness ppm (mg/l)* ≤0.3 ≤0.3
pH @ 25˚C 8.3-10.5 8.3-10.5
Nonbolatile TOC ppm (mg/l) C (6) <1 <1

Oily matter ppm (mg/l) <1 <1
Boiler Water
Silica ppm (mg/l) SiO2 ≤150 ≤90
Total alkalinity ppm (mg/l)* <1000(5) <850(5)
Free OH alkalinity ppm (mg/l)* (2) NS NS
Specific conductance (12)
mhos/cm (s/cm) 25˚C
without neutralization <7000(5) <5500(5)
*as CaO3
NS = not specified

02.qxd 8/8/06 1:51 PM Page 28
Table 2-4 Notes

(1) Values in the table assume existence of a deaerator.
(2) Chemical deaeration should be provided in all cases, especially if mechanical deaer-
ation is nonexistent or inefficient.
(3) Boilers with relatively large furnaces, large steam release space and internal chelant,
polymer, an/or antifoam treatment can often tolerate higher levels of feedwater impu-
rities by external pretreatment is always a more positive solution. Alternatives must
be evaluated as to practicality and economics in each individual case. The use of
some dispersant and antifoam internal treatment is typical in this type of boiler oper-
ation; therefore, it can tolerate higher feedwater hardness than the boilers in Table 1.
(4) Minimum and maximum hydroxide alkalinities must be individually specified by a
qualified water treatment consultant with regard to silica solubility and other com-
ponents of internal treatment.
(5) Alkalinity and conductance values are consistent with steam purity limits in the same
table. Practical limits above or below tabulated values should be individually estab-
lished by careful steam purity measurements.
treatment program.
(7) This limit presents steam purity that should be achievable if other tabulated water
quality values are maintained. The limit is not intended to be nor should it be con-
strued to represent a boiler performance guarantee.

02.qxd 8/8/06 1:51 PM Page 29
Table 2-5

Primary Fuel Fired
Conditions: No superheater, turbine drives, or process restriction on steam purity
Steam purity (7): 1.0 ppm (mg/l) TDS maximum
Drum Operation psig 0-300

Pressure 0-2.07 MPa
Feedwater (3)
Dissolved oxygen ppm (mg/1) O2 - measured
before chemical oxygen scavenger addition (1)(2) <0.007
Total iron ppm (mg/l) Fe <0.1
Total copper ppm (mg/l) Cu <0.05
Total hardness ppm (mg/l)* <1.0
pH @ 25˚C 8.3-10.5
Nonbolatile TOC ppm (mg/l) C (6) <10

Oily matter ppm (mg/l) <1
Boiler Water
Silica ppm (mg/l) SiO2 ≤150
Total alkalinity ppm (mg/l)* <700(5)
Free OH alkalinity ppm (mg/l)* (4) NS
Specific conductance (12) mhos/cm (s/cm)
25˚C without neutralization <7000(5)
*as CaO3
NS = not specified

02.qxd 8/8/06 1:51 PM Page 30
Table 2-5 Notes

(1) Values in the table assume existence of a deareator.
(2) Chemical deaeration should be provided in all cases, especially if mechanical
deareation is nonexistent or inefficient.
(3) Foretube boilers of conservative design, with internal chelant, polymer, and/or
antifoam treatment can often tolerate higher levels of feedwater impurities than those
in the table [≤0.5 ppm (mg/l) Fe, ≤0.2 ppm (mg/l) Cu, ≤ 10 ppm (mg/l) total hard-
ness] and still achieve adequate deposition control and steam purity. Removal of these
impurities by external pretreatment is always a more positive solution. Alternatives
must be evaluated as to practicality and economics in each individual case.
(4) Minimum and maximum levels of hydroxide alkalinity must be individually speci-
fied by a qualified water treatment consultant with regard to silica solubility and
other components in each individual case.
(5) Alkalinity and conductance guidelines are consistent with steam purity target.
Practical limits above or below tabulated values should be individually established
for each case by careful steam purity measurements.
treatment program.
(7) Target value represents steam purity that should be achievable if other tabulated
water quality values are maintained. The target is not intended to be nor should it be
construed to represent a boiler performance guarantee.
Table 2-6
Feedwater Volatile–Treatment Volatile–Treatment Oxygenated

Parameter (Hydrazine/Ammonia) (Hydrazine/Ammonia) Treatment
Mixed Metallurgy All Ferrous Systems
pH 8.8 - 9.1 9.2 - 9.6 8.0 - 8.5

Cation Conductivity <0.2 <0.2 <0.15
Iron (ppb) <10 <5 <5
Copper (ppb) <2 – –
Dissolved Oxygen (ppb) <5 1 - 10 30 - 150
Utility Boiler Feedwater Guidelines. Source: Electric Power Research Institute.

02.qxd 8/8/06 1:51 PM Page 31
Figure 2-14
<
>
>
Chemical Feed Schematic
Feedwater Chemistry Guidelines

The preceding sections have presented feedwater chemistry issues and vari-
ous treatment concepts. This section summarizes feedwater chemistry guidelines.
Tables 2–3 through 2–5 illustrate water chemistry guidelines developed by
the American Society of Mechanical Engineers for drum boilers at various pres-
sures. Even more stringent guidelines have been developed by organizations
such as EPRI for high-pressure utility boilers. A compilation of these values is
outlined in Table 2–6. As the tables indicate, feedwater purity guidelines become
increasingly strict as boiler pressures rise. The guidelines are difficult to imple-
ment unless the plant is equipped with proper feed and monitoring systems.
Chemical Feed Systems

As the preceding examples have illustrated, chemical treatment of feedwater
is very important. This places emphasis upon the design and operation of chem-
ical feed systems. Figure 2–14 outlines a chemical feed scheme in which the
chemical, an oxygen-scavenger/pH-conditioner blend, is fed to the process from
a portable storage tank. Key components of the system include:
● Isolation valves on the pump suction and discharge

● Suction-side strainers
● Suction-side calibration column
● Redundant diaphragm metering pumps
● Discharge line check valves
● Discharge line relief valves
● Pulsation dampeners
Many of these items are standard (check valves, isolation valves, and relief
valves) to protect the equipment and personnel from the process fluid, and to
allow the maintenance staff to work on the equipment safely. Other items, while

02.qxd 8/8/06 1:51 PM Page 32
still not standard, are becoming more common. One example is the increasing
use of diaphragm metering pumps in place of piston pumps. Diaphragm pumps
are reportedly more accurate and can be regulated more precisely than piston
pumps. At least one major metering pump manufacturer is phasing out piston
pumps in favor of the diaphragm type.
Another increasingly popular item is the pulsation dampener. This device
converts the pulsed bursts from the metering pump to a more steady-state flow.
Pulsation dampeners are becoming more common for other chemical feed appli-
cations including those to cooling tower basins.
Chemical feed rates are usually very low, and the time for a dose of reagent
to leave the pump discharge and arrive at the injection point may be quite long.
For instance, consider the following actual example. The hydrazine/pH-condi-
tioner feed pump on a 2400 psig unit delivers the solution at 12.3 gallons per
hour (GPH). Flow is through 3/8 in. OD (0.245 in. ID) tubing. The linear flow
rate is 1.4 ft/sec. For a line length of 100 feet, it takes over a minute for a dose
to leave the pump discharge and reach the injection point. If the line length had
been 500 feet, the time would increase to almost six minutes. Long lag times can
seriously affect the responsiveness of the chemical feed system to signals for
increased or decreased dosages. The closer the feed system is to the injection
point, the more responsive it will be.
Regarding oxygenated feed systems, various arrangements are available for
delivery of oxygen to condensate. Most popular is a simple system in which oxy-
gen cylinders, like those used by welders, are manifolded together. The discharge
is routed through a flow regulator and on to the injection point. Valves are
arranged such that when one cylinder empties, another can be put into service
immediately. This type of system can be easily placed near the injection location.
Some plant personnel prefer liquid oxygen as the source of supply, with a flow
regulated system. At least one U.S. utility simply uses the dissolved oxygen in the
makeup water as the supply. Results to date appear to be favorable, although it
seems that this arrangement allows less control than the others mentioned above.
Monitoring and Control of Condensate

Contaminants
Utility boilers and high-pressure industrial boilers can tolerate very little
contamination. Yet the potential for contamination is high at the steam surface
condenser. Cooling water to the condenser is typically supplied from a river,
lake, wells, or the ocean. If this water is cycled in a cooling tower, solids con-
centrations increase severalfold. Thus, a leaking condenser tube will introduce
many unwanted contaminants to the pure condensate of the feedwater system.
As chapter 3 describes in more detail, these contaminants can cause severe upsets

02.qxd 8/8/06 1:51 PM Page 33
in boilers. It is highly important that plant chemists closely monitor feedwater

chemistry, ideally with on-line continuous instrumentation supported by grab
sample analyses.
Typical sample points in a condensate/feedwater system include the con-
densate pump discharge (CPD), condensate polisher outlet (CPO) deaerator inlet
(DAI), deaerator outlet (DAO) and economizer inlet (EI) or, if no economizer is
present, boiler feedwater (FW) after the final high-pressure heater. The analyses
of these samples should, of course, reflect the most important parameters in that
area of the system. For instance, analyses of the feedwater/economizer inlet sam-
ples are based upon the feedwater guidelines previously mentioned, and usually
consist of pH, dissolved oxygen, oxygen scavenger concentration, conductivity,
and perhaps ammonia. Dissolved oxygen in the DAO is usually the only analysis
of this sample. If DO concentrations are less than 7 ppb, the deaerator is operat-
ing properly. Dissolved oxygen and oxygen scavenger samples from the DAI indi-
cate if condensate chemistry is under control.
The CPD sample, and CPO sample where applicable, along with boiler
water and main/reheat steam, are the most important of all since it is at the con-
denser discharge that gross contamination will usually be detected. Two parame-
ters that are most often measured are sodium and cation conductivity because
they are easily detectable at low concentrations. Even a seemingly minor con-
denser leak or demineralizer upset will cause a significant change in one of these
variables, which can be readily observed. The current guidelines suggest that
sodium be within a range of 3 to 10 ppb or less, with the lower level specified
for boilers that can least tolerate contamination. These include once-through
boilers and boilers on AVT with reheaters. Phosphate treated boilers are slightly
more tolerant of contamination due to the buffering action of the phosphate.
Likewise, guidelines for cation conductivity suggest a range of 0.2 to 0.3 µS/cm
for drum-type units and 0.15 µS/cm for units on oxygenated treatment.
Another recommended analyses for the condensate pump discharge is total
organic carbon (TOC). Current utility boiler guidelines suggest a TOC limit of
0.200 ppm in the condensate pump discharge.
Sodium analyses can be particularly effective for detecting condenser leaks.
With a tight condenser fed by brackish water, sodium concentrations should be
below 1 ppb for a unit whose makeup system includes a mixed-bed ion exchang-
er for final polishing. At my former utility, sodium concentrations are usually
below this level. The operators are given a visual alarm on the control room water
chemistry monitoring panel if sodium levels exceed 2 ppb, and are given an audi-
ble alarm at 5 ppb. Even though the units have no condensate polishers, sodium
levels remain consistently below 1 ppb, so that even the slight increase to 2 ppb
indicates potential problems. (Soot blowing, and its attendant effects on steam
flow patterns, sometimes cause the sodium levels to increase above 1 ppb for a

02.qxd 8/8/06 1:51 PM Page 34
couple of hours. The plant chemists will begin to check for a condenser leak if
the concentration remains above 1 ppb for a more extended period of time. Case
History 2–4 outlines an interesting situation involving an intermittent condenser
tube leak.)
Condenser tube leaks may occur for several reasons. Waterside fouling, dis-
cussed in chapter 6, can cause pitting and tube failure. Steamside corrosion, an
example of which has been outlined in Case History 2–3, is a frequent problem.
Condenser tubes are usually rolled into the tube sheet, and leaks may occur at
these locations. Whatever the problem, even small leaks are cause for concern.
Supplement 2–3 provides a simple BASIC program for calculating the quantity of
a leak based on an analysis of sodium. It can give plant chemists an idea of the
seriousness of the leak and the effect it may have on boiler water chemistry. The
program can be used check leakage volumes based on other ions such as chloride.
Sometimes, contaminants may be introduced to the condensate from unlike-
ly locations. Off-line contamination is well documented. In two cases involving
sandblasting of a condenser and deaerator storage tank, respectively, plant per-
sonnel forgot to tell the contractor that an inert material such as ground walnut
shells should be used as the blasting media. Instead, the contractor, who would
not be expected to know boiler water chemistry, used a form of silica-based mate-
rial for blasting. Upon unit startup, silica gradually leached from sandy residue
remaining in the vessels and affected water chemistry. Contaminants can also be
introduced from other unsuspected locations, and Case History 2–5 illustrates a
unique example of one such occurrence.
Condensate Polishing and Treatment of Condensate

Return to Industrial Boilers
Once-through boilers and boilers on AVT cannot tolerate dissolved solids.
Historically, these units have been equipped with condensate polishers to remove
dissolved solids and particulates from condensate. The need for condensate pol-
ishing is not necessarily limited to these units. Condensate polishing can extend
the life of almost any boiler by protecting the unit from condenser inleakage and
particulate transport to the boiler.
High-Pressure Boiler Condensate Polishing

Two types of condensate polishers are preferred for high-purity applications.
These are deep-bed or powdered-resin ion exchangers. They may be either full-
flow or partial-flow, and in the case of powdered-resin polishers, can be installed
upstream of low-pressure feedwater heaters. The different characteristics of each
allows flexibility in the selection of a system for a particular application.

02.qxd 8/8/06 1:51 PM Page 35
Deep-Bed Condensate Polishing

Deep-bed condensate polishers are quite similar to makeup demineralizers
in that water is treated by a bed of ion exchange resin. When deep-bed conden-
sate polishing was first introduced several decades ago, filters were placed ahead
of the polisher to remove particulates. Polisher technology has evolved such that
in most modern deep-bed applications, the influent to the ion exchanger is unfil-
tered and the bed is allowed to remove both dissolved and particulate solids.
This is known as “naked” polishing.
Various configurations are possible for deep-bed polishing, just as they are
for makeup demineralization. One fairly common arrangement is a cation
exchange unit followed by a mixed-bed unit. More common is stand-alone,
hydrogen-hydroxide form, mixed-bed polishing.
Several factors about mixed-bed condensate polishing stand out. First, oper-
ating pressure is typically greater than that in a makeup system, and this must be
considered in the design. Second, because the polisher treats already purified
water, flow rates can be as high as 50 GPM/ft2 and sometimes even higher. Third,
due to the thermal instability of anion resins, deep-bed polishers must be located
ahead of any low-pressure feedwater heaters. Fourth, since the polisher removes
particulates, run length may be dependent on pressure differential rather than the
approach of bed exhaustion. Fifth, ammonia or amines are often used for feed-
water pH-conditioning. Any of these products that carry over with steam will be
absorbed by the polisher. Exhaustion of a polisher is first indicated by ammonia
breakthrough. Sixth, since polishers are in continuous service, some redundancy
is required. An arrangement with three 50% units is quite common.
Run lengths of a deep-bed polisher can range from several days to a month
or more. Regenerations may be based on a calculated throughput or when the
resin reaches exhaustion as signified by the ammonia break. Where the deep-bed
removes many particulates, backwash and regeneration are initiated when differ-
ential pressure across the bed reaches a certain limit.
The operating pressure of a condensate system influences several aspects of
polisher design that are not quite as critical for makeup demineralizers. The
higher pressure requires a higher pressure vessel rating. System upsets may dis-
turb the resin and place it under more mechanical stress. Design of the influent
distribution system is very important to prevent channeling. In some cases, a
spherical design is used for the vessels to improve hydraulics and water flow
through the bed.
Regeneration is the most critical aspect of condensate polisher operation.
The process is usually performed externally to prevent the possible leakage of
contaminants or regenerant chemicals into the condensate. Various external
regeneration arrangements have been developed over the years, but most popu-

02.qxd 8/8/06 1:51 PM Page 36
lar is either a two- or three-vessel system. In one common design, the resin is
sluiced to a backwash vessel for cleaning and hydraulic classification of the
cation and anion resin. The anion resin is then transferred to a separate vessel,
upon which each resin is regenerated individually. The regenerated resins are
then pumped to a third vessel where they are mixed and stored for recharge to
the condensate polisher.
A problem formerly encountered with this arrangement was cross-contami-
nation of each resin by off-size or broken beads of the other. Thus, cation resin
that ended up in the anion regeneration system would come out in the sodium
form, and anion resin accidentally introduced to the cation regeneration system
would be converted to the sulfate form. These ions would then leak into the con-
densate once the resin was returned to service.
Newer technologies have been developed that minimize the cross-contami-
nation of resins. The use of uniformly sized cation resins and uniformly sized
anion resins allows better separation during the backwash cycle. Some polishing
systems are designed to accommodate resins containing a small quantity of inert
beads. The inert resin has a density between that of the cation and anion resins,
and keeps them physically separated. Other polishers, such as the Conesep sys-
tem by Graver, extract the cation resin rather than the anion resin. Very good
resin separation has been achieved with these types of systems.
Deep-bed units can be operated in the ammonia form. In this arrangement,
the polisher releases ammonium ions during the exchange process. The ammo-
nia produced by the unit helps condition the condensate and feedwater, and per-
forms the same function as ammonia directly injected to the condensate.
Operation of a polisher in the ammonium cycle is more complicated, however.
Regeneration of the cation resin becomes a two-stage process, wherein the resin
is first regenerated with acid and then with ammonia. During operation of an
ammonia-hydroxide cycle polisher, leakage of other cations, particularly sodium,
may be greater than in a hydrogen-hydroxide cycle unit. Control of condensate
pH is also more difficult. For these reasons, ammonia-hydroxide cycle polishers
are not strongly recommended for high-pressure boilers.
An aspect of importance for hydrogen-hydroxide deep-bed polisher opera-
tion is ammonia concentration of the influent. The higher the ammonia concen-
tration, the more quickly the cation resin exhausts. The ammonia concentration
is of course related to the dosage used for pH control of the feedwater. Polisher
run lengths can become particularly short as the condensate pH rises towards the
9.5 level. Once-through plant personnel who switched feedwater treatments
from AVT to OT have noticed a severalfold increase in polisher run lengths. This
is due to the change in pH from a range of 9.0–9.6 to 8.0–8.5.

02.qxd 8/8/06 1:51 PM Page 37
Powdered-Resin Condensate Polishing

Where particulate removal is of primary concern, powdered-resin polishers
may be more effective. This type of polisher is comprised of filter elements to
which a coat of finely grained, mixed cation-anion resin is applied. The
filter/resin combination performs the dual purpose of straining particulates and
removing dissolved solids. This type of resin is nonregenerable, so the resin must
be replaced periodically, perhaps every month or two. One of the advantages of
a precoat system is that it can be located downstream of the low-pressure feed-
water heaters, and thus can capture any iron oxides and other particulates that
may have been produced in the condensate system.
Precoat systems operate at much
Figure 2-15 lower cross-sectional flow rates (2 to
4 GPM/ft2) than deep-bed polishers.
However, a relatively large cross-sec-
tional area can be provided. A com-
mon design is the candle-type filtra-
tion system illustrated in Figure
2–15. The filter elements may be
constructed of a variety of materials
including wire mesh, nylon string, or
fabric. In the precoat process, the
cation and anion resins are mixed
with water in a separate tank and are
then pumped to the polisher vessel.
As water passes through the filter
housings, the resins collect on the fil-
ter media. The water is recirculated to
the mix tank until all of the resin has
been transferred. Common resin
depths range from 1/8 inch to 1/4
inch. It is important to follow all
manufacturer's procedures when
>
>
installing the precoat, since a uniform

resin depth on all filters is required.
Powdered-resin condensate pol-
ishers will remove particles smaller
than 10 microns. Run lengths are
>
based on pressure differential and are

terminated when the differential
Simplified Schematic of a Candle-Type, Powdered-
Resin Condensate Polisher reaches a set limit. This may be as

02.qxd 8/8/06 1:51 PM Page 38
high as 25 psi, although lower differentials are more common.

Ammoniated cation resin can be used in these systems, but again sodium leak-
age may be greater than that from acid regenerated resin. One nice aspect of pow-
dered-resin systems is that the ratio of cation to anion resin can be modified dur-
ing a precoat cycle to produce a blend of resin that gives the best quality effluent.
Condensate polishers are rather expensive items whose costs may range from
a few hundred thousand dollars for a relatively small system to well over a million
dollars or more for a large unit. The costs must be weighed against the potential
advantages. Single incidents involving condenser tube leaks have cost utilities mil-
lions of dollars in material replacement and lost power production. A polisher
would have paid for itself in the prevention of just one of these lone upsets.
Industrial Plant Condensate Polishing

Many industrial steam boilers do not operate at pressures that would nor-
mally require condensate polishing. However, in many cases a great portion of
the steam used for the industrial process is returned to the boiler. The conden-
sate return often contains impurities. Some of these may be iron or metal oxides
generated in the heat exchanger and condensate return piping system. Others are
introduced by leaking heat exchangers, and consist of chemicals produced by the
plant. Condensate return may be contaminated with acids, inorganics, organics,
oils, and other compounds. These will have many of the same effects on boiler
water and steam chemistry that a condenser tube leak or oil contamination caus-
es in a utility boiler. (Case History 4-2 in chapter 4 outlines an interesting exam-
ple of organics contamination and its effect of package boiler steam chemistry.)
Treatment methods for condensate return are quite varied. One simple pro-
cedure is to dump the condensate if conductivity limits are exceeded. The con-
densate is not returned to service until the problem is corrected. Other analyses
can be used for the same purpose. For example, condensate returns at petro-
chemical plants are often monitored for total organic carbon (TOC). A return can
be dumped if TOC exceeds a predetermined limit.
Condensate polishing at industrial plants may take on several forms. Sodium-
cycle condensate polishers are suitable for systems in which only hardness needs
to be removed. Multimedia, cartridge or backwashable filters are choices for treat-
ment of condensate return containing suspended solids. Removal of oil and
organics presents special difficulties. Large, complex organics can be removed by
activated carbon; however, smaller organics and oils may partially pass through
the carbon bed. One possible treatment for these compounds is organic scavenger
resins. These are formulated like regular ion exchange resins, except no exchange
sites are added. Rather, the resin beads are synthesized with many pores to trap
the organics. The resins are operated to exhaustion and are then replaced.

02.qxd 8/8/06 1:51 PM Page 39
Conclusion
Control of chemistry in the condensate/feedwater system is extremely criti-
cal for proper boiler operation and safety. This chapter is intended to give
chemists readily available information regarding monitoring and control of feed-
water chemistry. (More details on sampling and monitoring are provided in chap-
ter 7, Sampling.) The case histories outlined in this and other chapters will hope-
fully illustrate that creative thinking is important when solving feedwater, boiler
water, and steam chemistry problems.

02.qxd 8/8/06 1:51 PM Page 40

02.qxd 8/8/06 1:51 PM Page 41
Supplement 2-1
Why Condenser Performance Is Important
I have been asked on a number of occasions, “Why isn't the turbine exhaust
steam returned ‘as is’ to the boiler? Wouldn't that save a lot of money by elimi-
nating the condenser and other equipment?” The answer is thermodynamically
related.
Consider what happens in a condenser. Ideally, when steam leaves the tur-
bine it has used all of its available heat for work and is at a saturated condition.
The condenser removes the steam's latent heat and converts it to condensate,
which produces a very strong vacuum. As the following example shows, this is
quite important thermodynamically.
Let us examine the following hypothetical situation.
1. Steam, at a pressure of 1000 psig and temperature of 950°F enters a

turbine
2. The turbine is adiabatic and reversible, and has no bleed lines for
feedwater heating

02.qxd 8/8/06 1:51 PM Page 42
The steam tables show that the enthalpy of the turbine inlet steam is 1477
BTU/lbm and the entropy is 1.6325 BTU/lbm °R. Because the turbine is adiabat-
ic and reversible, the entropy of the exhaust steam is the same as the entropy of
the throttle steam since no heat is transferred during the reversible expansion.
Accordingly, the enthalpy of the exhaust steam can be calculated for various con-
ditions. In the hypothetical turbine, if the steam were taken from the turbine
exhaust at atmospheric pressure, its enthalpy would be almost 1070 BTU/lbm.
Thus, 407 BTU/lbm of heat would be available for work. Because the enthalpy of
the condensate at these conditions is 180 BTU/lbm, the turbine efficiency would
equal ([1477 - 180] - 1080) / 1477 - 100 = 31%. If, however, all incoming steam
exhausts into a vacuum, the situation changes noticeably. With relatively cool cir-
culating water, absolute condenser pressures can reach as low as 0.5 inHg. The
enthalpy of steam exhausting into these conditions is near 846 BTU/lbm. Thus,
the amount of heat available for work increases to 631 BTU/lbm, giving a turbine
efficiency of 44%. This is an enormous improvement over the previous scenario.
(Note: Too much cooling can cause a loss of efficiency. This phenomenon, known
as condensate subcooling, will not be discussed in this book. Condensate sub-
cooling does not cause nearly the efficiency losses as do other factors such as tube
fouling or excess air inleakage.)
Obviously, the hypothetical example just presented is exaggerated since no
utility turbine is reversible or is designed to exhaust to atmospheric conditions.
Also, in most circumstances some steam is extracted from the turbine for feed-
water heating, so an increase in condensate temperature would decrease the
quantity of steam needed for feedwater heating. Thus, higher condensate tem-
peratures would alter the heat transfer in the feedwater heaters. However, the
quantity of steam extracted for feedwater heating is much smaller than that
which passes through the condenser, so the hypothetical example listed above
still clearly illustrates the thermodynamic importance of the condensing process.
Supplement 2-2
BASIC Program for Monitoring
Condenser Performance
Poor condenser performance can subtract 1% or more from unit efficiency.
This can add up to a considerable loss of money in fuel costs alone. The prima-
ry causes of poor performance are fouling of the tubes on the waterside or excess
air inleakage into the steamside. Each can drastically reduce condenser perfor-

02.qxd 8/8/06 1:51 PM Page 43
mance, as is described in more detail in this book. Utility personnel employ var-
ious methods to track condenser conditions, including monitoring of condenser
backpressure and waterside inlet and outlet differential pressures. However, the
attached BASIC program provides a very effective method for tracking condenser
efficiency.
Using the Program

Thirteen measurements or values are needed to run the program: cooling
water inlet and outlet temperatures; cooling water density; condenser steam tem-
perature; cooling water flow rate; correction factors for the condenser tube mate-
rial and inlet water data; data about the condenser and tubes (number of passes,
tube length, number of tubes, tube inside diameter, and tube outside diameter);
and an empirical constant known as C, which has been derived by the Heat
Exchange Institute. Because most of the dimensional quantities are constant for
any particular condenser, the number of inputs can be reduced by almost half if
constants for tube length, ID, OD, number of passes, and number of tubes are
written into the program. In addition, the density of water decreases by only
0.5% over the temperature range of 40°F to 90°F, so a constant value for water
density can be incorporated into the program without noticeably affecting the
calculations. Tables for condenser tube and cooling water correction factors are
included with this program, and they too can be entered as constants.
Program Output
The program calculates an ideal heat transfer (Ui), which is then compared
to the actual heat transfer (Ua). This value is known as the cleanliness factor.
When condenser tubes are placed in service they quickly develop an oxide coat-
ing. This coating, which actually protects the metal substrate, retards heat trans-
fer. Thus, a condenser free from mineral or microbiological deposits will still only
achieve about 85% of the ideal heat transfer, so a cleanliness factor of 85% indi-
cates totally clean tubes. (Factors above 80% are quite good and indicate that the
condenser is achieving good heat transfer.) It is very important that accurate
temperature readings are available because slight variations in temperature can
significantly affect results. However, even if absolute values are not totally accu-
rate, the program is still good for indicating changes in condenser performance.

02.qxd 8/8/06 1:51 PM Page 44
CONDPERF.BAS
10 CLS:LOCATE 5,1
20 PRINT “CONDENSER CLEANLINESS FACTOR PROGRAM”
30 PRINT
40 PRINT “ENTER THE CIRCULATING WATER INLET TEMPERATURE (¯F)”
50 INPUT TIN
60 PRINT
70 PRINT “ENTER THE DENSITY OF THE INLET COOLING WATER (LB/FT^2).”
80 PRINT “A VALUE OF 62.3 IS SUITABLE FOR ALL CONDITIONS.”
90 INPUT RHO
100 PRINT
110 PRINT “ENTER THE CIRCULATING WATER OUTLET TEMPERATURE (¯F)”
120 INPUT TOUT
130 PRINT
140 PRINT “ENTER THE CONDENSER STEAM TEMPERATURE”
150 INPUT TSAT
160 PRINT
170 PRINT “ENTER THE CIRCULATING WATER FLOW RATE (GPM)”
180 INPUT FLOW
190 PRINT
200 PRINT “ENTER THE CIRCULATING WATER CORRECTION FACTOR AS”
210 PRINT “OUTLINED IN SUPPLEMENT 2-2.”
220 INPUT CWCF
230 PRINT
240 PRINT “ENTER THE CONDENSER TUBE CORRECTION FACTOR AS”
250 PRINT “OUTLINED IN SUPPLEMENT 2-2.”
260 INPUT CTCF
270 PRINT
280 PRINT “ENTER THE NUMBER OF CONDENSER TUBES”
290 INPUT NT
300 PRINT
310 PRINT “ENTER THE NUMBER OF CONDENSER TUBE PASSES”
320 INPUT NP
330 PRINT
340 PRINT “ENTER THE INSIDE TUBE DIAMETER (IN.)”
350 INPUT ID
360 PRINT
370 PRINT “ENTER THE OUTSIDE TUBE DIAMETER (IN.)”
380 INPUT OD
390 PRINT
400 PRINT “ENTER THE TUBE LENGTH (FEET)”
410 INPUT L
420 PRINT
430 PRINT “IN THE CALCULATIONS WHICH FOLLOW, THE IDEAL HEAT TRANSFER”
440 PRINT “IS DETERMINED THROUGH THE USE OF AN EMPIRICAL CONSTANT”
450 PRINT “DETERMINED BY THE HEAT EXCHANGE INSTITUTE. THIS CONSTANT”
460 PRINT “KNOWN AS C, VARIES DEPENDING ON THE OUTSIDE TUBE DIAMETER.”
470 PRINT “THE VALUES OF C ARE AS FOLLOWS:”
480 PRINT “5/8 & 3/4 INCHES, C=267”
490 PRINT “7/8 & 1 INCHES, C=263”
500 PRINT “1-1/8 & 1-1/4 INCHES, C=259”
510 PRINT “1-3/8 & 1-1/2 INCHES, C=255”
520 PRINT “1-5/8 & 1-3/4 INCHES, C=251”
530 PRINT “1-7/8 & 2 INCHES, C=247”
540 PRINT
550 PRINT “ENTER A VALUE FOR C”
560 INPUT C
570 PRINT

02.qxd 8/8/06 1:51 PM Page 45
580 REM Q = HEAT TRANSFER (BTU/HR)

590 REM DTLM = LOG MEAN TEMPERATURE
600 REM SAOD = SURFACE AREA OF OUTSIDE TUBE SURFACE (FT^2/FT)
610 REM UA = ACTUAL HEAT TRANSFER COEFFICIENT
620 REM VL = LINEAR VELOCITY OF WATER THROUGH THE TUBES (FT/SEC)
630 REM UI = IDEAL HEAT TRANSFER COEFFICIENT
640 REM UD = DESIGN HEAT TRANSFER COEFFICIENT
650 REM CF = CLEANLINESS FACTOR
660 Q=8.0203*RHO*FLOW(TOUT-TIN)
670 DTLM=(TOUT-TIN) /LOG((TSAT-TIN) / (TSAT-TOUT))
680 SAOD=3.14159* (OD/12)
690 UA=Q/ (DTLM*SAOD*L*NT)
700 VL=(.002228*FLOW) / ((((ID/24)^2) * (3.14159*NT/NP)))
710 UI=C*SQR (VL)
720 UD=UI*CWCF*CTCF
730 CF=(UA/UD) * 100
740 PRINT “THE CLEANLINESS FACTOR =”;:PRINT USING“###.#”;CF
Supplement 2-3
BASIC Program for Calculating the Rate of
Cooling Water Leakage into a Condenser
When a leak develops in a condenser the flow rate of the leak can be deter-
mined by comparing the concentration of a particular contaminant in the cool-
ing water to that in the condensate. Sodium is a good parameter for this mea-
surement. The following BASIC program will calculate the flow rate of the leak
based on the difference in sodium concentrations. Plant personnel can then
determine the potential effects of the leak on unit chemistry, and decide whether
the size of the leak calls for immediate action.
CONDLEAK.BAS
10 CLS:LOCATE 5,1
20 PRINT “CONDENSER INLEAKAGE CALCULATION PROGRAM”
30 PRINT
40 REM QFW = FEEDWATER FLOW IN POUNDS PER HOUR
50 REM QFWGPM = FEEDWATER FLOW IN GPM
60 REM NAFW = SODIUM CONCENTRATION IN THE FEEDWATER (PPM)
70 REM NACW = SODIUM CONCENTRATION IN THE COOLING WATER (PPM)
80 REM QCWL = COOLING WATER LEAK INTO CONDENSER (GPM)
90 INPUT “ENTER THE FEEDWATER FLOW (POUNDS PER HOUR)”;QFW
100 PRINT
110 INPUT “ENTER THE SODIUM CONCENTRATION IN THE FEEDWATER (PPM)”;NAFW
120 PRINT
130 INPUT “ENTER THE SODIUM CONCENTRATION IN THE COOLING WATER (PPM)”;NACW
140 PRINT
150 QFWGPM=QFW/(8.34*60)

02.qxd 8/8/06 1:51 PM Page 46
160 QCWL=(QFWGPM*NAFW)/(NACW-NAFW)
170 PRINT “THE RATE OF COOLING WATER FLOW INTO THE CONDENSER = “;:PRINT
USING “###.##”;QCWL;:PRINT “ GPM”
Case History 2-1

Conditions: Two-pass condenser
Copper-nickel tubing
I had been performing thrice-weekly cleanliness factor analyses on the con-
denser. The values remained very steady in the mid-70% range for several
months, but suddenly within two days dropped to 45%. Waterside fouling does
not occur nearly this rapidly. Such changes are more indicative of air inleakage
and accumulation of gasses on the steam-side condenser tube surfaces. The
maintenance staff was so notified. When the maintenance crew inspected the
condenser, they quickly discovered a crack in the condenser shell where a heater
drip line entered. Once they welded this crack, the cleanliness factors returned
to previous values, where they remained for another two months until suddenly
dropping again. The weld had failed. The maintenance crew then welded a col-
lar to the shell, which totally sealed the crack. This cured the problem.
Case History 2-2

Admiralty tubing
I had been collecting thrice-weekly cleanliness factor readings on this con-
denser as well. Rather suddenly, the condenser began performing erratically. At
high unit loads, the cleanliness factor ranged between 70% to 75%, but at low
unit loads it sometimes dropped as low as 18%. Again, such fluctuations could
not have been the result of waterside tube fouling. All evidence pointed to air
inleakage, but the source could not be located. Utility managers brought in a leak
detection firm, whose personnel used helium leak detection to completely check
the condenser and turbine. They classified leaks as large, medium, and small,
and found well over a dozen leaks including two large ones, one of which was at
the expansion joint between the turbine and the condenser. All leaks were
repaired by plant maintenance crews, but this did not solve the problem. Finally,
an operator discovered that a trap on the condensate return line from the gland
steam exhauster was sticking open at low loads. The trap and line were designed
to return condensed gland steam from the subcooler to the condenser, but allow

02.qxd 8/8/06 1:51 PM Page 47
any gases to vent to the atmosphere. When the trap stuck open, the strong con-
denser vacuum pulled outside air in through the vent. Maintenance personnel
quickly replaced the trap and the condenser performance problems disappeared.
Case History 2-3

Admiralty tubing
The condenser operated for 17 years with few tube failures. Without warn-
ing, periodic tube failures began to occur in the air removal section and caused
a number of forced outages. During one of these forced outages, maintenance
personnel pulled out four of the previously failed tubes. All of these tubes exhib-
ited circumferential gouges at points where the tubes passed through tube sup-
port plates. A through-wall penetration had occurred at one of these gouges in
each tube. The gouged metal contained numerous small cracks.
The corrosion was identified as ammonia/oxygen attack of the copper alloy.
The air removal system concentrates both gases, but because ammonia is highly
soluble in water, much of it redissolves and runs down the tube support plates.
The high concentration of ammonia in a relatively high oxygen environment
caused this localized corrosion. In addition, admiralty brass is more susceptible
to ammonia/oxygen corrosion than virtually all of the other copper alloys. The
air removal section was retubed with 70-30 copper nickel, which is more resis-
tant to this attack.
Interestingly, the air removal section of the condenser on a sister unit con-
tained stainless steel tubes. It had originally been designed with admiralty tubes,
but these had failed within one year of startup. Plant personnel recalled that
when the unit was first placed in service, the drips from both stages of the air
removal steam jet ejectors had been returned to the condenser. These drips typ-
ically contain several ppm or more of ammonia. Although no written records
exist, it is quite likely that the ammonia introduced by the drips caused ammo-
nia grooving at the tube support plates. Many years later, tubes directly below the
air removal section began to fail by this mechanism. At one point, these failures
caused four forced outages in five months. When maintenance crews pulled
some of the failed tubes for examination, the failures had occurred at ammonia
grooves corresponding to the junction of the tubes with the tube support plates.

02.qxd 8/8/06 1:51 PM Page 48
Case History 2-4

Admiralty tubes
Once-through cooling system surface water supply
Sodium concentrations in the condensate pump discharge typically average
between 0.5 and 0.9 ppb. Once, however, during approximately a three-week
stretch of essentially baseload operation, the sodium levels periodically increased
to a range of 1 to 2 ppb, where they remained for several hours or perhaps up to
a day before returning to normal. The sodium increase could not be traced to any
operational factors such as load change or soot blowing, so plant chemists con-
cluded that a condenser tube leak had developed. Maintenance personnel did
indeed find a pinhole leak in one tube. When they plugged the tube, the prob-
lem was cured. However, the question still remained as to why the contamina-
tion appeared then disappeared with regularity. What seems most likely is that
the hole frequently became plugged with debris that entered with the cooling
water. No problems were noted with the analyzer, which had successfully detect-
ed leaks before and after this incident. On-line sodium analysis proved extreme-
ly valuable in tracking this leak and its peculiar behavior.
Case History 2-5

Conditions: 175 MW unit
2400 psig operating pressure
Following an outage at a Midwestern utility, plant chemists discovered that
higher than normal levels of silica continued to persist in the feedwater and boil-
er water long after the silica concentrations normally decreased to acceptable
ranges. Excess silica in boiler water is sometimes caused by upsets in the raw
water supply, which introduce colloidal silica that breaks down to reactive silica
as it is heated. However, plant chemists thoroughly analyzed the makeup water
and demineralizer product water but found no evidence of abnormal concentra-
tions of ionic or colloidal silica. Following additional investigation, the chemists
learned that when maintenance personnel closed the deaerator storage tank door
for unit startup, they had used a silicon-containing compound to seal the door.
Silica was leaching from the sealant into the feedwater.

02.qxd 8/8/06 1:51 PM Page 49
Table 2-7
Condenser Tube Correction Factors
Tube Wall Gauge - BWG
Tube Material 24 22 20 18 16 14 12
Admiralty Metal 1.06 1.04 1.02 1.00 0.96 0.92 0.87
Arsenical Copper 1.06 1.04 1.02 1.00 0.96 0.92 0.87
Copper Iron 194 1.06 1.04 1.02 1.00 0.96 0.92 0.87
Aluminum Brass 1.03 1.02 1.00 0.97 0.94 0.90 0.84
Aluminum Bronze 1.03 1.02 1.00 0.97 0.94 0.90 0.84
90–10 Cu–Ni 0.99 0.97 0.94 0.90 0.85 0.80 0.74
70–30 Cu–Ni 0.93 0.90 0.87 0.82 0.77 0.71 0.64
Cold Rolled Low
Carbon Steel 1.00 0.98 0.95 0.91 0.86 0.80 0.74
Stainless Steels
Type 304/316 0.83 0.79 0.75 0.69 0.63 0.56 0.49
Titanium 0.85 0.81 0.77 0.71 – – –
Condenser Tube Correction Factors. Source: Heat Exchange Institute. Adapted from “Computer program pre-
dicts condenser cleanliness factors,” Power Engineering, June 1992.
Table 2-8
Cooling Water Correction Factors
Inlet CWCF Inlet CWCF Inlet CWCF

Temperature (˚F) Temperature (˚F) Temperature (˚F)
30 0.550 54 0.855 78 1.037
31 0.562 55 0.865 79 1.041
32 0.574 56 0.875 80 1.045
33 0.586 57 0.885 81 1.048
34 0.601 58 0.895 82 1.051
35 0.615 59 0.905 83 1.054
36 0.628 60 0.915 84 1.057
37 0.641 61 0.925 85 1.060
38 0.655 62 0.934 86 1.063
39 0.668 63 0.942 87 1.066
40 0.683 64 0.951 88 1.069
41 0.694 65 0.960 89 1.072
42 0.707 66 0.970 90 1.03
43 0.720 67 0.978 91 1.085
44 0.733 68 0.986 92 1.088
45 0.747 69 0.993 93 1.090
46 0.760 70 1.00 94 1.092
47 0.772 71 1.005 95 1.095
48 0.785 72 1.010 96 1.097
49 0.797 73 1.015 97 1.100
50 0.810 74 1.020 98
51 0.822 75 1.025 99
52 0.833 76 1.029 100
53 0.844 77 1.033
Cooling Water Correction Factors. Source: Heat Exchange Institute. Adapted from “Computer program pre-
dicts condenser cleanliness factors,” Power Engineering, June 1992.
02.qxd 8/8/06 1:51 PM Page 50

03.qxd 8/8/06 1:52 PM Page 51
Chapter 3
Boiler Water
Chemistry
Introduction
Deposition of contaminants and corrosion are common in boilers and are
the cause of many forced outages. Material damage and lost power generation or
product output due to corrosion cost electric utilities and industries billions of
dollars per year. This chapter outlines the principal mechanisms of boiler chem-
istry upsets and discusses treatment methods to minimize them. Preceding this
discussion is a brief look at some of the common boiler types used in the fossil-
fired steam generating industry.
Drum-Type Boiler Design

The most popular boilers at industrial facilities and electric utilities are of
the drum variety. They are so called because of the presence of a steam drum,
which gives the boiler some unique operating characteristics. Large industrial
and utility drum boilers must be erected in the field, but smaller boilers are often
shipped and installed as package units.

03.qxd 8/8/06 1:52 PM Page 52
Figure 3-1 Package Drum
STEAM
Boilers
DRUM
Many industrial plants
generate steam only for process
use or to drive small turbines.
For these applications, precon-
structed, package boilers are
often sufficient. Figures 3-1
through 3-3 illustrate the sim-
plified circuitry of three of the
most common package boiler
designs, the “A,” “D,” and “O”
types. Natural gas or oil is the
MUD MUD principal fuel for these boilers.
DRUM DRUM Steaming rates typically range
General Circuitry of an A Type Boiler from 40,000 to 200,000
pounds per hour, with
100,000 pounds per hour being very common. As the figures indicate, each of
the boilers contains one or more mud drums. The mud drum(s) serves as a col-
lection site for precipitates formed by chemical treatment programs.
The greatest advantage of package boilers is that they can be added in stages
as plant capacity increases. Package boilers are relatively small in size, so site and
construction requirements are far more simple than for field-erected units. The
boilers can be equipped with superheaters to produce steam of the quality and
temperature needed to drive turbines or provide adequate heat for high temper-
ature industrial processes. Package boilers are typically rather short in height,
where their length may be modified depending on steaming requirements.
Field-Erected Drum Units

Field-erected units are the norm for the higher steam quantities or pressures
needed at electric utilities and industrial plants that produce significant amounts
of power. Custom-built units utilize long, vertical waterwall tubes and are thus
much larger and taller (Figs. 3-4 and 3-5) than package units. Although gas and
oil-fired boilers are not uncommon, the predominant fuel is coal. Several boiler
designs have evolved over the years. During the heyday of large boiler construc-
tion, the most common types included Cyclone, wall-fired, and pulverized-coal
(PC), tangentially fired boilers. The steam drum is located near the top of the
unit, and the boiler either contains a mud drum or lower headers for collection
of precipitates. High temperature superheater and reheater tubes are hung verti-

03.qxd 8/8/06 1:52 PM Page 53
Boiler Water Chemistry 53
cally just before or in the gas passage from the boiler, although some horizontal-
ly arranged superheater and reheater tubes may be located further downstream
in the gas passage. Larger units are often equipped with an economizer that is
placed in the gas passage near the low-temperature superheaters and reheaters.
Circulating Fluidized Bed Boilers and Heat Recovery

Steam Generators
Due to the changing nature of the electric utility industry, and the effects of
more stringent environmental regulations, other steam generating technologies
have emerged in the United States. These include circulating fluidized bed (CFB)
boilers and heat recovery steam generators (HRSG). In a CFB boiler, the fuel is
combusted in an air-mixed (fluidized) bed within the furnace compartment.
Combustion temperatures in this process usually range between 1500˚F and
1700˚F, and are much lower than the 2700˚F to 3000˚F temperatures produced
in more conventional boilers. The lower temperatures greatly reduce the forma-
tion of nitrogen oxides (NOx). Other advantages of CFBs include:
•A sulfur dioxide reactant (lime-

Figure 3-2
stone) is directly blended with the
fuel, and it removes SO2 in the fur-
STEAM nace.
DRUM •A wider variety of coals can be used
in a CFB unit than in other boilers.
The steam generation process in

a CFB is based upon the waterwall
tube/drum arrangement, so water
treatment programs are similar to
those for other drum-type boilers.
HRSGs have become popular as
the second component of combined-
cycle or cogeneration systems in
which waste heat from a combustion
turbine produces steam in the
HRSG. The features of combined-
cycle operation that have proven
MUD most appealing include the fact that
DRUM
unit efficiencies for electrical produc-
tion are approaching 60%; the sys-
General Circuitry of a D Type Boiler
tems are fired with natural gas (and

03.qxd 8/8/06 1:52 PM Page 54
Figure 3-3 sometimes oil); a unit is relative-

ly inexpensive; and a combined-
cycle plant can be erected much
STEAM more quickly than a coal-fired
DRUM
unit. HRSGs are unique because
the steam generator may con-
tain two or three successively
higher pressure circuits, each
with its own drum. Water
chemistry varies from one cir-
cuit to another and makes water
treatment somewhat complex.
The aspect of drum-type
units that greatly affects chem-
istry is the continual circulation
of water from the steam drum
through relatively cooler down-
comers to the waterwall tubes
and back (Fig. 3-6). Steam pro-
MUD
DRUM duced in the waterwalls is col-
lected in the upper portion of
the drum for distribution. The
General Circuitry of an O Type Boiler
circulation/evaporation process
concentrates solids in the boiler
water. Drum boiler chemistry treatment programs are designed to minimize the
scaling or corrosive effects of contaminants, and also to prevent excessive accu-
mulation of solids, which may then carry over with the steam.
Once-Through Steam Generation

As chapter 4 discusses in more detail, the difference in density between
steam and water allows for their separation in a boiler drum. For units at or
above supercritical pressure (3205 psig), the fluid becomes a single phase, thus
eliminating drums from consideration. All of the incoming feedwater is convert-
ed to steam in the waterwall tubes, which is then collected in steam headers for
distribution. Since the water does not circulate, precipitating treatment programs
are unacceptable, but rather zero-solids programs are necessary and feedwater
chemistry limits take precedence. Zero solids treatment also applies to subcriti-
cal once-through boilers and has been the treatment of choice for drum boilers
that operate at 2800 psig or above.

03.qxd 8/8/06 1:52 PM Page 55
Figure 3-4
Utility Boiler. Photo taken with permission of the Kansas City, Kansas, Board of Public Utilities,
Quindaro Power Station
Boiler Water Contamination

Boiler water treatment programs are designed to minimize corrosion and
also to mitigate the effects of contaminants, which can enter the boiler from a
number of different sources.
Iron Oxide Deposition

As was discussed in chapter 2, the boiler waterside tube walls develop a
tightly bound layer of magnetite when the unit is placed in operation. Over the
passage of time, however, the protective film becomes overlaid with more porous
deposits of magnetite, which consist of iron-oxide corrosion products transport-
ed from the feedwater system. These porous deposits have much lower heat
transfer coefficients than the tube metal and can eventually reduce boiler effi-
ciency and cause localized overheating of tubes. Porous deposits can also act as
concentration sites for potentially corrosive chemicals, which normally are at low
concentrations in the bulk boiler water. Water that penetrates the deposit is sub-
jected to greater heat as it reaches the tube surface. It may then boil off through
other pores in the deposit (Fig. 3-7), leaving concentrated species behind.
Iron oxides usually constitute the bulk of boiler tube deposits and are most
responsible for the need to chemically clean the boiler. A large amount of iron
oxide particles are generated during boiler startups. These can be detected very
easily by a filtration process. In this method, a sample of boiler water is passed
through a 0.45 micron filter. At initial startup, the filter will often turn black with
magnetite particles as the sample passes through. Succeeding filtrations over the

03.qxd 8/8/06 1:52 PM Page 56
Figure 3-5
Common Utility Boiler Outline Drawing provided by DB Riley, Inc.
course of the day will gradually change from black to gray to nonexistent as the
particles are either removed by the boiler blowdown or precipitate on the tube
walls.
One particular difficulty with iron oxide deposition is that the particles pre-
cipitate more heavily on the hot side of the tubes. This is the worst location, as
higher temperatures increase the potential for corrosion. Under-deposit corro-
sion mechanisms will proceed more rapidly on the hot side of the tubes.
Condenser Inleakage
A condenser tube leak allows many unwanted contaminants to enter the
steam generating system. These contaminants, upon reaching the boiler, react to

03.qxd 8/8/06 1:52 PM Page 57
Figure 3-6
SATURATED
STEAM
>
STEAM
DRUM
DOWNCOMERS WATERWALL
TUBES
>
FLOW > FLOW
HEAT >
>
>
LOWER
HEADER
Flow Pattern in a Drum-Type Boiler
form scale, acid, or other deleterious compounds. The following equations illus-
trate some of the principal reactions.
Ca+2 + 2HCO3- → CaCO3↓ + CO2↑ + H2O (3.1)

Ca+2 (or Mg+2) + SiO3-2 → CaSiO3↓ (or MgSiO3↓) (3.2)
MgCl2 + 2H2O →Mg(OH)2↓ + 2HCl (3.3)
Equations 3.1 and 3.2 are typical scale-forming reactions. Many hardness
compounds exhibit retrograde solubility as temperatures rise (Fig. 3-8) and will
precipitate on boiler internals. The effect of such deposition on heat transfer is
outlined in Table 3-1, which shows thermal conductivities for several boiler met-
als and scales. Even a relatively thin deposit layer will significantly reduce heat
transfer, and a boiler must be fired harder to achieve the same level of steam pro-

03.qxd 8/8/06 1:52 PM Page 58
Figure 3-7
Illustration of Chimney Boiling in a Deposit. Illustration by Alyssa Buecker.
duction. This in turn can lead to overheating of the boiler tubes, which will
shorten tube life. In cases of severe deposition, heat transfer is so restricted that
overheating may occur very rapidly resulting in catastrophic tube failure.
Deposits are also the precursor to another phenomenon known as underde-
posit corrosion, which may occur via several different mechanisms. Boiler water
is normally maintained in an alkaline range (pH of 9 to 11 depending on boiler
pressure and type) to prevent acidic attack (Fig. 3-9). This pH range is moder-
ately basic, and small amounts of caustic alkalinity (OH-) may be present in the
bulk boiler water. In one form of caustic corrosion, water that enters a porous
deposit through some openings will boil off through other channels, leaving
sodium hydroxide behind. The concentrated NaOH then attacks the boiler metal
and protective magnetite film via the following reactions:
Fe + 2NaOH → Na2FeO2 + H2↑ (3.4)

Fe3O4 + 4NaOH → 2NaFeO2 + Na2FeO2 + 2H2O (3.5)
The localized attack may cause tube damage within a relatively short period
of time.
Deposits can also cause the concentration of other species including acid
chlorides, which are generated by reactions similar to that outlined in Equation
3.3. Not only does the acid corrode the boiler metal, but the reaction generates
hydrogen, which can lead to hydrogen damage of the tubes. In this mechanism,
hydrogen gas molecules, which are very small, penetrate into the metal wall
where they then react with carbon atoms in the steel to generate methane (CH4):

03.qxd 8/8/06 1:52 PM Page 59
Figure 3-8
Solubility of Various Scale-Forming Compounds as a Function of Temperature Copyright © 1986. Electric

Power Research Institute. EPRI CS-4629. Interim Consensus Guidelines on Fossil Plant Cycle Chemistry. Reprinted With
Permission.
2H2 + Fe3C → 3Fe + CH4↑ (3.6)
Formation of the very voluminous methane molecules causes cracking in

the steel, greatly weakening its strength. Hydrogen damage is very troublesome
because it cannot be easily detected. After hydrogen damage has occurred, the
plant staff may replace tubes only to find that other tubes continue to rupture.

03.qxd 8/8/06 1:52 PM Page 60
Figure 3-9
Corrosion Characteristics of Mild Steel as a Function of pH.
Whatever the mechanism of contaminant introduction or deposit formation,

if it is severe enough the consequences can be quite drastic, as Case Histories 3-
1 and 3-2 at the end of this chapter illustrate. For this reason, all utility boilers
and most industrial boilers are chemically treated to mitigate the effects of cont-
aminants and to condition the boiler water to prevent corrosion.
Boiler Water Treatment Programs

Guidelines for boiler water treatment are, to a partial extent, dependent on
the effect of boiler water chemistry on steam purity. Accordingly, boiler water
guidelines will be discussed not only in this chapter, but in the next, which out-
lines steam chemistry.

03.qxd 8/8/06 1:52 PM Page 61
Table 3-1 Various treatment programs

Thermal Conductivity have been developed over the
Material 2
(BTU/ft -hr-˚F-in) years to optimize boiler water
chemistry control. For the medi-
Carbon Steel 360 um- and high-pressure units com-
Magnetite 20
mon to electric utilities and some
Calcium Carbonate 7
Porous Silica 0.6 industrial plants, the most popular
programs are coordinated-congru-
Thermal Conductiveness of Boiler Tube Metal and Two
Common Scales. ent phosphate treatment, equilib-
rium phosphate treatment (EPT),
phosphate treatment (PT), polymer treatments, caustic treatment, all-volatile treat-
ment, and oxygenated treatment.
Early Boiler Water Treatment

In the early days of steam generation, various odd-sounding treatment meth-
ods were used. Some of these treatments had operators putting sawdust or pota-
to peels into boilers. These natural products contributed organics such as lignins,
tannins, and starch to the water. The chemicals acted as sequestering agents of
hardness ions. Other chemistry programs actually allowed the formation of calci-
um carbonate on tube walls to protect the tube surface from the boiler water.
These programs were successful because boiler pressures were very low. As
boilers increased in size and pressure, more advanced treatment methods became
necessary. A big breakthrough came when phosphate treatment was introduced.
Sodium phosphate compounds (principally trisodium phosphate [Na3PO4]
blended with smaller amounts of disodium phosphate [Na2HPO4]) have been the
standard for years in many boiler water conditioning programs. The phosphate
ion is effective because it conditions boiler water in an alkaline pH range, and
reacts with scale-forming compounds to produce soft sludges. Trisodium phos-
phate provides the needed alkalinity as follows:
Na3PO4 + H2O ⇔ NaOH + Na2HPO4 (3.7)
Phosphate ions (PO4-3) can exist in the mono, di, and trihydrogen state in
aqueous solutions, and thus can give up or accept hydrogen ions. This buffering
capability makes phosphate effective in preventing wide pH swings in boiler
water due to contaminant ingress.
Phosphate’s second major function is to absorb contaminants that enter the
boiler. Phosphate reacts directly with calcium to produce calcium hydrox-
yapetite:
10Ca+2 + 6PO4-3 + 2OH- → 3Ca3(PO4)2 • Ca(OH)2↓ (3.8)

03.qxd 8/8/06 1:52 PM Page 62
Figure 3-10
Guidelines for Coordinated Phosphate Control Source: Betz Industrial Water Treatment
Seminar, BetzDearborn, Inc., Horsham, PA.
Magnesium and silica react with the alkalinity produced by phosphate to

form the nonadherent sludge, serpentine:
3Mg+2 + 2SiO3- + 2OH- → 3MgO • 2SiO2 • 2H2O↓ (3.9)
Hydroxyapetite and serpentine are much more benign and easy to remove via
blowdown than the hard scale or corrosive products which would otherwise form.
Coordinated and Congruent Phosphate Treatment

In the early days of phosphate treatment, phosphate and pH were main-
tained at fairly high ranges of 20 to 40 ppm phosphate and pH of 11 to 12. As
boilers increased in size and pressure, corrosion of the waterwall tubes began to
be a problem. Researchers determined that the high alkalinity generated caustic
corrosion via the mechanisms outlined in Equations 3.4 and 3.5. Accordingly,
chemists began to refine phosphate treatment programs to prevent caustic attack.
One of the principal programs that came out of this effort was coordinated phos-
phate treatment.
As is evident from its chemical formula, the molar ratio of sodium to phos-

03.qxd 8/8/06 1:52 PM Page 63
phate in trisodium phosphate is three to one. In a coordinated phosphate pro-

gram, enough disodium phosphate is added to maintain a Na/PO4 ratio between
2.8 and 2.2 to 1. The Na2HPO4 shifts the equilibrium of Equation 3.6 to the left,
which helps minimize the formation of free caustic. Figure 3-10 illustrates the
phosphate/pH control limits to maintain coordinated treatment at subcritical
boiler pressures.
Although coordinated treatment was a significant refinement to phosphate
treatment, the process did not stop there. Sodium phosphates are reversely solu-
ble at temperatures above about 250˚F and will begin to precipitate (hide out) on
boiler tube walls. The precipitate usually contains a lower sodium-to-phosphate
ratio than the bulk boiler water, and this can influence boiler water chemistry. As
chemists recognized this phenomenon, they modified coordinated phosphate
treatment to congruent treatment, in which the ideal Na/PO4 ratio became 2.6 to
1 with a lower limit of 2.2 to 1. Evidence at the time indicated that the Na/PO4
ratio of the deposits was around 2.6 to 1. This ratio was selected for the treat-
ment so that any phosphate, which did come out of solution, would precipitate
“congruently” and not affect chemistry. Congruent precipitation has proven to be
untrue in many cases, especially in high-pressure units, which will be discussed
in the next section.
Congruent treatment and occasionally coordinated treatment are still wide-
ly used in many low- and medium-pressure boilers, and in some high-pressure
utility boilers that do not suffer from hideout. Phosphate concentrations in a con-
gruent program are typically maintained within a 2 to 5 ppm range. The key to
the program is the sodium-to-phosphate ratio. If the ratio climbs above the rec-
ommended limit of 2.6 to 1, the water may become too alkaline. Conversely,
ratios below the recommended lower limit of 2.2 to 1 can generate acidic condi-
tions. Supplement 3-1 contains a BASIC program for calculating sodium-to-
phosphate ratios. The only inputs needed are pH and phosphate. The program
is based on standard industry calculations that were developed when coordinat-
ed and congruent programs became popular. The program is simple to use, and
can even be incorporated into on-line water chemistry data acquisition systems.
(See Case History 3-3.)
The supplement mentions a correction calculation for ammonia. This is due
to the effect ammonia (or a neutralizing amine) has on the Na/PO4 ratio calcula-
tion. Ammonia will raise boiler water pH, but obviously not change either the
sodium or phosphate concentration. When ammonia is present, the standard
Na/PO4 ratio calculation outlined in the supplement will predict a higher
Na/PO4 ratio than actually exists. This is particularly troublesome from a con-
gruency standpoint, as any phosphate precipitates may have a sodium-to-phos-
phate ratio well below the minimum recommended limit of 2.2 to 1. The chemist
will assume that phosphate dosages are correct, when in actuality acidic phos-
phates may be corroding the tube walls.

03.qxd 8/8/06 1:52 PM Page 64
Figure 3-11
Guidelines for Coordinated Phosphate Control Adjusted for Ammonia Level. Chart
Provided by BetzDearborn, Inc., Horsham, PA.
This effect must be taken into account when monitoring the phosphate
treatment program. Several methods are possible. Figure 3-11 illustrates control
curves for coordinated phosphate treatments at various ammonia levels. By mon-
itoring the ammonia concentration in the boiler water, the chemist can prepare
phosphate dosages based on the curves. Where neutralizing amines are used, the
calculation becomes more complicated since various amines exhibit different
basicities. Holland Technologies of Jamison, Pennsylvania, has developed a com-
puter program that accounts for the effects of ammonia and the common amines
that are used in feedwater systems. Finally, measurement of both sodium and
phosphate allows direct calculation of the sodium-to-phosphate ratio. If done
on-line, the results can be incorporated into data acquisition systems, such as
that already mentioned in Case History 3-3.
Within the last decade or so, more and more difficulties with congruent
chemistry in high-pressure boilers have become apparent. This is due to phos-
phate hideout and the incongruent precipitation of sodium phosphate.
Phosphate Hideout
Figure 3-12 illustrates the reverse solubility of sodium phosphate. The effect
becomes very pronounced at the boiler water temperatures common to units that

03.qxd 8/8/06 1:52 PM Page 65
Figure 3-12
Solubility of Trisodium Phosphate versus Temperature Copyright © 1986. Electric Power Research Institute.
EPRI CS-4629. Interim Consensus Guidelines on Fossil Plant Cycle Chemistry. Reprinted with permission.
operate above 2000 psi and/or are subjected to frequent load changes. Reports
by some utilities formerly using congruent treatment indicate that phosphate
concentrations would often drop well below 1 ppm when hideout was severe.
This depleted the boiler water of the chemical designed to control chemistry and
protect against contaminant introduction. Furthermore, researchers found that
the phosphate precipitated incongruently, with deposit sodium-to-phosphate
molar ratios of 2 to 1 or lower.
The effects of hideout are most graphically illustrated in a cycling unit.
Consider a boiler at low load with congruent treatment. As load is raised and
heat fluxes increase, sodium phosphate begins to precipitate onto the tube sur-
faces. Due to incongruent precipitation, the Na/PO4 ratio of the phosphate
remaining in solution rises. This increases the pH to perhaps several tenths of a
unit above congruent guidelines. The chemist may then add a di-/trisodium
blend to raise the phosphate concentration and lower the pH, but this phosphate
also hides out. When boiler load is reduced, the situation reverses and the pre-
cipitated phosphate redissolves. In this case, however, the sodium phosphates,
which have a low Na/PO4 ratio, drive the pH downward. Severe hideout and the
reverse dissolution process have been known to force boiler water pH below 7 in
cycling units. Since EPRI recommends unit shutdown if the pH drops below 8.0,

03.qxd 8/8/06 1:52 PM Page 66
such events can be most alarming. Coordinated or congruent phosphate treat-

ment may be impossible to control in boilers that exhibit this phenomenon.
Phosphate hideout appears to be further influenced by the cleanliness of the
boiler tubes, and becomes more severe with increased deposit loading, particu-
larly of iron oxides on the tubes. Sodium phosphates form a sodium-iron-phos-
phate complex with the magnetite layer. Routine boiler chemical cleanings can
potentially reduce some hideout by minimizing the formation of porous mag-
netite deposits.
The sodium-iron-phosphate reaction also causes other problems. Harvey
Herro and Robert Port of Nalco and others have found sodium-iron-phosphate
products (with few other contaminants) at boiler tube corrosion sites. This indi-
cates that sodium phosphate actually participates in corrosion reactions. This
may be due in part to the fact that the phosphate hides out at a low sodium-to-
phosphate ratio, which tends to make the deposits acidic. Acid phosphate cor-
rosion is now suspected as being a significant culprit in many boiler tube failures
that previously were thought to have been caused by other mechanisms.
The problem of phosphate hideout has become so well recognized that
many utilities formerly on congruent treatment have switched to alternate pro-
grams. Congruent control cannot be universally condemned, as it still works well
in some units. However, this number is decreasing.
Alternative Phosphate and

Nonphosphate Programs
At least two alternative phosphate programs have been developed, which
counteract the formation of acidic phosphate deposits. These are equilibrium
phosphate treatment (EPT) and phosphate treatment (PT).
Equilibrium Phosphate Treatment

EPT was pioneered by Jan Stodola of Ontario Hydro for use in the utility’s high-
pressure fossil units. In an EPT program, phosphate residuals are maintained with-
in a range of 0.2 to 2.5 ppm, and pH within a range of 9.0 to 9.7. Alkalinity is con-
trolled by the addition of trisodium phosphate with supplemental addition of caus-
tic when needed. This results in a solution with a sodium-to-phosphate ratio of 2.8
to 1 or greater, with most of the control range at 3 to 1 or above. Caustic alkalinity
is maintained from 0 to 1 ppm. Adjustments are usually made by adding trisodium
phosphate, with caustic only added at startup to raise pH.
Ontario Hydro has been using EPT for over a decade, and utility personnel
have reported excellent results. They have also reported that under-deposit caustic
corrosion has not been a problem, even in boilers with moderate deposit buildups.

03.qxd 8/8/06 1:52 PM Page 67
While equilibrium phosphate programs minimize hideout, the treatment is

less forgiving towards contaminant inleakage than coordinated or congruent pro-
grams. As Case History 3-2 illustrates, phosphate can be quickly consumed dur-
ing a large condenser leak. Thus, monitoring of contaminant inleakage, as out-
lined in chapters 2 and 7, is very important. Utility personnel at a number of
other utilities have adopted EPT, and reports of its successful use continue to
increase.
Phosphate Treatment
Phosphate treatment almost appears to be a combination of EPT and con-
gruent or coordinated treatment. Phosphate residuals are maintained between
2.5 to near 10 ppm, with a sodium-to-phosphate ratio of 2.8 to 1 or greater. As
with EPT, caustic alkalinity is held to 1 ppm maximum. Hideout in PT-condi-
tioned systems has been reported. Although such deposits are more alkaline in
nature than those in a congruent program, hideout can still create difficulties for
the chemist. Possibly for this reason, EPT appears to be the more preferred
replacement for congruent or coordinated programs.
Chelants and Polymer Treatments

Chelants are compounds that chemically bind (chelate) certain elements.
Two of the most commonly used chelants for boiler water treatment are ethyl-
enediamine-tetraacetic acid, or EDTA (Fig. 3-13), and nitrilotriacetic acid. These
chelate cations such as calcium, magnesium, iron, and copper through interac-
tion of the positively charged ions with the partially negative oxygen and nitro-
gen atoms of the chelant. The cation in a chelated structure appears as if it is in
a cage. Chelants do not form precipitates like the phosphates, but rather keep
hardness and other ions in solution so that they can be removed in the boiler
blowdown. EDTA has been effective in boilers operating at pressures up to 1500
psig, but the compound will begin to break down at a temperature of 400˚F to
produce byproducts that also chelate, but less effectively. NTA is typically limit-
ed to boilers of less than 900 psig pressure.
The chelants react on a 1 to 1 basis with cations, thus a significant residual
may be needed to tie up contaminants. An EDTA molecule has a molecular
weight of 288, while the respective molecular weights of calcium and magnesium
Figure 3-13
HOOC-CH2 CH2-COOH
N-CH2-CH2-N
HOOC-CH2 CH2-COOH
Structure of the ETDA Molecule.

03.qxd 8/8/06 1:52 PM Page 68
are 40 and 24.3. This indicates that 7.2 and 11.9 ppm of EDTA would theoreti-
cally react with 1 ppm of calcium and magnesium, respectively.
The use of chelants must be carefully considered because they can cause prob-
lems in boiler systems. For example, chelants are attacked by dissolved oxygen and
should not be added to the feedwater until after the deaerator. More importantly,
improper control and overfeed of chelants can cause dissolution of the protective
magnetite layer on the boiler tubes. Serious corrosion has been known to occur.
Polymers of the acrylate (Fig. 3-14), sulfonated styrene (Fig. 3-15), and
other families are also used for boiler water scale control. Most often the poly-
mers are employed as part of a combined program with phosphates or chelants.
The polymers modify the crystalline structure of precipitates, making them less
sticky and more easily removable by blowdown. These chemicals do not act on
a stoichiometric basis with the compounds being treated, and typical dosages
may range from 1 to 10 ppm.
One area where sequesterants such as polymethacrylate (Fig. 3-16) or sul-
fonated styrene appear to be particularly effective is iron control, especially after
unit startup. The sequesterants will bind free iron and allow it to be removed by
the boiler blowdown rather than precipitate on the tube walls.
Chelants and polymers have been
Figure 3-14 effectively used for boiler water chemistry
CH2 CH2 ~ control. However, proper selection of the
~ CH CH
correct chemical or combination of chem-
COOH COOH icals may not be easy. Monitoring of the
Structure of Polyacrylic Acid. chemical residuals is more difficult than in
phosphate programs. Many people in the
Figure 3-15
electric utility industry appear to favor the
H H H more “simple” phosphate or volatile treat-
~ C C C ~ ment programs. For low-pressure indus-
trial treatment, however, a chelant/poly-
H mer program may perform very well.
Certainly polymers should be considered
SO H SO H for iron control during unit startups.
3 3
Structure of Sulfonated Polystyrene.
Figure 3-16 Caustic Treatment

Caustic treatment, wherein sodium
CH3 CH3 hydroxide is the principal boiler water
CH2 CH2 ~ conditioning chemical, is used at a num-
~ C C
ber of overseas utilities. In a caustic treat-
COOH COOH ment program, boiler water pH is main-
tained within a range of 9.4 to 9.6 by
Structure of Polymethylmethacrylate. addition of caustic to maintain a NaOH

03.qxd 8/8/06 1:52 PM Page 69
level of 1.0 to 1.5 ppm. Feedwater chemistry in a caustic treatment program

must be well controlled to prevent excessive deposit formation in the boiler,
which might lead to underdeposit caustic corrosion. Caustic treatment of high-
pressure boilers has not caught on in the United States, although caustic control
of low-pressure boiler water chemistry was once a common technique. Sodium
nitrate is frequently added with the caustic to prevent stress corrosion cracking
of boiler tubes.
All-Volatile Treatment
All-volatile treatment (AVT) was principally developed for once-through
boilers, since these units cannot tolerate dissolved solids. Because once-through
boilers have no drum, boiler water chemistry is a function of feedwater chem-
istry. AVT chemistry guidelines for a once-through unit call for a pH range of 9.3
to 9.6 with less than 2 ppm dissolved solids. Ammonia levels may range from 1
to 2 ppm. Condensate polishers are an absolute requirement for once-through
units because a condenser leak or demineralizer upset would introduce uncon-
trollable contamination to the boiler.
AVT is also used in some very high-pressure, drum-type units where the
pressure approaches critical. As boiler pressures increase to this value, the den-
sity of steam and water approach each other. Thus, it becomes much more diffi-
cult to separate water from steam in the boiler drum internals, so boiler water
must be quite pure to prevent carryover.
AVT minimizes mechanical carryover, but it does not protect drum boilers
from contaminant introduction due to a condenser leak or other problems.
Condensate polishers are the most effective buffer against chemistry upsets. For
units without polishers, the chemical feed system should contain an emergency
phosphate unit so that phosphate can be immediately injected into the drum in
the event of an upset. This must be considered as only a temporary measure,
because the sudden increase in dissolved solids due to contaminant inleakage
and phosphate injection will greatly increase the potential for carryover and
deposition.
Several problems with AVT have become evident. First, the volatile treat-
ment chemicals will carry over into the steam. Unfortunately, they also tend to
carry chloride and sulfate with them, which then deposit on the low-pressure
turbine blades. Chlorides and sulfates are prime contributors to stress-corrosion
cracking and corrosion fatigue. Second, the volatiles are less capable of neutral-
izing acidic turbine deposits than phosphate, which, if it carries over in small
quantities, may be beneficial due to its alkaline nature. Third, where amines and
organic oxygen scavengers are used, the breakdown products may include car-
bon dioxide and organic acids. These are potential corrodents of turbine blades

03.qxd 8/8/06 1:52 PM Page 70
and afterboiler components.

AVT has lost a lot of its popularity. Oxygenated treatment has supplanted it
at many once-through utilities overseas and now in the United States.
OT, which was described in chapter 2, is basically a feedwater treatment. Its
effectiveness in boilers stems from the fact that OT greatly reduces iron transport.
This is quite important, as iron oxide usually makes up the bulk of boiler tube
deposits. European utilities that have been using OT for a long time have found
that chemical cleanings have been significantly reduced. This provides a great
economic incentive, when one considers that the recommended chemical clean-
ing frequency of a supercritical unit is every 18 months.
Case History 3-4 provides some specific details on the startup and initial
observations of an OT program at a supercritical steam generating unit. An
important point to remember is that the oxide layer formed in an OT program is
orange in color. This can come as a surprise when plant personnel open up a
feedwater system after initiating oxygenated treatment.
OT is currently being tested in one drum boiler in the United States. The
program has been in progress for over three years and results so far seem to be
promising. Iron transport has definitely been lowered. Look for this treatment to
be tested in other drum boilers.
The concept of oxygenated treatment can be difficult to accept, especially con-
sidering that for years all of the chemistry manuals emphasized removal of oxygen
from boiler feedwater. However, positive reports about the treatment are proliferat-
ing. OT has even been recommended for circuits in heat recovery steam generators.
Heat Recovery Steam Generators

HRSGs are often multistage units in which a low-pressure steam circuit
feeds water or steam as makeup to higher-pressure circuits. Figure 3-17 shows a
generic outline of a two-stage HRSG. Water treatment may take on some inter-
esting aspects, depending on the unit configuration. One common design has the
low-pressure circuit fitted with a combination drum/deaerator. Water from this
integral deaerating circuit is used for feed to the higher-pressure circuits and also
as attemperation water. Such a configuration requires high-purity makeup, zero-
solids treatment for the low-pressure circuit, and perhaps even condensate pol-
ishing of the feedwater. Other HRSG designs have separate circuits in which only
the steam from the low-pressure system is used elsewhere in the unit. Sodium
softening may be adequate for makeup treatment.

03.qxd 8/8/06 1:52 PM Page 71
Figure 3-17
Structure of a Two-Drum HRSG.

03.qxd 8/8/06 1:52 PM Page 72
Chemical treatment of HRSGs is complicated by the multipressure aspect of

the units. One combined-cycle manufacturer fabricates HRSGs with a low-pres-
sure circulating stage and a high-pressure, once-through stage. Recommended
treatments are AVT for the low-pressure circuit and OT for the high-pressure sys-
tem. Where all of the circuits are of the drum variety, phosphate treatments can
be very effective. Even the highest-pressure circuit of an HRSG is usually below
1500 psig, so hideout is minimal. Furthermore, HRSGs are not exposed to direct
flame as are conventional boilers, thus, heat fluxes are much lower. This mini-
mizes not only the overheating of tubes, but also the adverse chemical reactions
that can be caused at localized areas of high heat input.
Sampling
Sampling of boiler water is extremely important because the harsh environ-
ment in the boiler greatly magnifies the effects of corrosion and deposition mech-
anisms. The most important constituents to be monitored include pH, conduc-
tivity, silica, and phosphate. On-line monitoring is greatly preferred, with grab
sampling as a backup. For drum-type units, the sample is either taken from spe-
cial ports located in the drum, or more often from the boiler blowdown or down-
comer. Further details are outlined in chapter 7.
Layup and Off-Line Corrosion Protection

More corrosion can occur during an outage than at any other time. This is
almost exclusively due to air intrusion into the boiler and subsequent oxygen
corrosion of components. The optimum theoretical approach to prevent corro-
sion is to completely dry the boiler with warm air circulation. For long-term out-
ages, where the unit may not be needed for quite a while, this merits serious con-
sideration, although complete dryness is sometimes difficult to achieve. Most
outages, however, are only for a few days or a few weeks, with the knowledge
that the boiler will be required to operate by a specific date. Wet layups are much
more practical for these situations. However, the boiler and feedwater system
must be laid up properly or severe corrosion will result.
Layup guidelines differ depending on whether the layup will be short term,
less than four days or so, or long term, such as several weeks or a month. For
short-term layups, the boiler should be filled with condensate containing 200
ppm of hydrazine and enough ammonia or neutralizing amine to bring the pH
to 10. A nitrogen blanket should be introduced to the drum and superheater
through the vent lines as the boiler pressure decays. A 5 psig N2 pressure needs
to be maintained throughout the outage. Additionally, the feedwater system must

03.qxd 8/8/06 1:52 PM Page 73
not be neglected. If the standard operating concentrations of oxygen scavenger

and pH-conditioner are in the feedwater at time of shutdown, this solution can
remain. If not, then provisions should be made to introduce a solution contain-
ing 50 ppb of oxygen scavenger at a pH of 10 to the system. The steam-side of
the heaters should be blanketed with steam or nitrogen, as exfoliation of feed-
water heater tubes can become very severe without this protection. The deaera-
tor should also be blanketed with steam, if possible, or laid up with an oxygen-
scavenger/ammonia solution. The superheater and reheater are usually allowed
to remain dry.
For long-term outages, some modifications to the treatment are recom-
mended. For example, a solution containing 50 to 100 ppm of hydrazine is sug-
gested for the feedwater system, with pH adjustment to 9.5. Some experts also
recommend filling the superheater with the same solution as in the boiler. If this
is done, plant personnel should pin the superheater hangers, as the water will
add a great deal of weight to the pendants. A nitrogen cap should be applied at
the drum and superheater vents to prevent intrusion of air.
Where possible, the boiler layup solution should be periodically circulated.
This helps reduce stagnant zones and also allows the plant chemists to obtain
more accurate analyses of the layup chemical concentrations. Should the con-
centrations be too low, the staff can add more chemicals. Case History 3-5
describes recirculation systems that were installed on two drum boilers in
response to a unique wet-layup-induced corrosion mechanism.
Before the boiler is started up, the layup chemicals need to be removed. This
includes a flush of the feedwater system and drain and refill of the boiler. If this is
not done, excessive ammonia will be present throughout the system. This can
cause serious chemistry problems, especially if the feedwater heaters are tubed
with copper alloys. If the superheaters are drainable, they should be emptied, too.
Conclusion
A number of effective programs are available for boiler water treatment.
Plant chemists should carefully monitor boiler water conditions to determine the
effectiveness of the program in use. Single upsets have been known to cause tube
failures and forced outages that have cost utilities millions of dollars for material
replacement and purchased power. In industry, boiler failures can curtail or shut
down production units.
Off-line corrosion can also be very destructive. Proper layup procedures are
important to maintain the integrity of the steam generating unit during outages.

03.qxd 8/8/06 1:52 PM Page 74

03.qxd 8/8/06 1:52 PM Page 75
Supplement 3-1
BASIC Program for Calculating
Sodium-to-Phosphate Ratios of Boiler Water
Sodium-to-phosphate ratio monitoring is very important for boilers that are
treated with coordinated or congruent phosphate. Chemists must maintain the
ratios within relatively narrow guidelines to properly control the boiler water
chemistry. The BASIC program on page 77 provides a simple and efficient method
for determining these ratios. The only inputs needed are pH and phosphate con-
centrations (ppm), however, ammonia may significantly affect the calculations.
The calculations first appeared in the November 1986 issue of Power
Engineering magazine, and the computer program then appeared in the May
1992 issue of this magazine. Since the publication of these two articles, utility
personnel and researchers have become much more aware of the effect of ammo-
nia on boiler water pH. At the 1996 International Water Conference, George
Verib of Ohio Edison presented an excellent paper on boiler water treatment,
part of which discussed ammonia and its relation to sodium/phosphate ratios.
Ammonia can have a very significant effect on the calculations, especially in
higher pressure units where the sodium-phosphate concentrations are low. Ohio
75
03.qxd 8/8/06 1:52 PM Page 76
Edison has switched to equilibrium phosphate treatment in many of their units.

Personnel have refined the sodium/phosphate ratio calculation by determining a
value known as the actual hydroxide concentration, where:
[OH-]actual = [OH-]measured - [NH3]measured

pOHactual = -log[OH-]actual
pHactual = 14 - pOHactual
The program on page 77 can be modified to incorporate these calculations,

as follows:
65 PRINT
66 INPUT “ENTER THE AMMONIA CONCENTRATION (PPM) IN THE DRUM WATER”;NH3
85 REM THE MOLAR AMMONIA CONCENTRATION = NH3/17027
86 MPNH3=NH3/17027
120 REM OHM = THE MEASURED HYDROXYL ION CONCENTRATION
130 OHM=10^(PH-14)
135 REM OHA = THE ACTUAL HYDROXYL ION CONCENTRATION
136 OHA=OHM-MPNH3
380 MSP=H+DIPO:TSP=OHA+DISS:DSP=MPCO-MSP-TSP
Other techniques potentially offer more accurate measurement when

ammonia, or perhaps more importantly, amines are present. One is direct analy-
sis of both the boiler water sodium and phosphate.
Case History 3-1

Immediately after this boiler had been returned to service from a scheduled
autumn unit outage, lab chemists discovered a condenser leak that was intro-
ducing up to 0.75 ppm of total dissolved solids (TDS) to the condensate.
Although the lab staff requested that the boiler be taken off-line immediately, the
operations managers refused because load demand was too high. The boiler was
on congruent phosphate control, so the plant chemists increased monitoring fre-
quency and maintained phosphate and pH levels within recommended guide-
lines. Blowdown rates were increased to remove the extra sludge that was
formed. After approximately three weeks, the source of contamination was dis-
covered and repaired.* Two months after the condenser upset, boiler waterwall
tubes began to fail with alarming frequency. The unit was taken off numerous
times for boiler tube repairs, and in one instance had only been on-line for a few
hours when another tube failed. The failures happened so frequently that plant
managers determined that an emergency tube replacement was needed during

03.qxd 8/8/06 1:52 PM Page 77
the scheduled spring outage. This was completed at an approximate cost of

$2,000,000. The mechanism attributed to these failures was under-deposit cor-
rosion caused by excessive sludge and scale formation. This example provides a
graphic illustration of the problems that condenser leakage can cause.
PHOS.BAS
10 CLS:LOCATE 5,1
20 PRINT “THIS PROGRAM WILL CALCULATE THE SODIUM/PHOSPHATE RATIO IN BOILER
WATER.”
30 REM - THE CALCULATIONS FOR THIS PROGRAM MAY BE FOUND ON PAGES 31-32
40 REM - OF THE NOVEMBER 1986 ISSUE OF POWER ENGINEERING
50 PRINT
60 INPUT “ENTER THE DRUM WATER pH” ; PH
70 PRINT
80 INPUT “ENTER THE PPM OF PHOSPHATE IN THE DRUM WATER” ; PO
90 PRINT
100 REM - THE MOLAR PHOSPHATE CONCENTRATION = PO/94971
110 MCPO=PO/94971!
120 REM - H = THE HYDROGEN ION CONCENTRATION
130 H=1-^(-1*PH)
140 POH=PH-14
150 REM - OH = THE HYDROXYL ION CONCENTRATION
160 OH=10^(PH-14)
170 REM - A = THE K1 CONSTANT FOR THE PHOSPHATE SERIES
180 A=10^-2.1
190 REM - B = K1 * K2
200 B=A*10^7.2
210 REM - C = K1 * K2 * K3
220 C=B*10^-12.3
230 REM - THE CALCULATIONS IN LINES 260-280 DETERMINE THE “APPLICABLE PHOS-
PHATE
240 REM - SPECIES DISTRIBUTION USING IONIZATION FRACTION METHOD FOR TRIPRO-
TIC
250 REM - ACID.”
260 E=H*H*H : F=H*H*A : G=H*B : D=C=E=F=G
270 AA=H*H : KA=A/D : KB=B/D : KC=C/D
280 DIPO=(AA*KA)*MCPO : MONO=(H*KB)*MCPO : ZERO=KC*MCPO
290 REM - THE FOLLOWING CALCULATIONS DETERMINE THE MOLAR CONCENTRATIONS OF
300 REM - THE VARIOUS PHOSPHATE SPECIES AND SODIUM
310 MSP=H+DIPO : TSP=OH+ZERO : DSP=MCPO-MSP-TSP
320 NA=MSP+(2*DSP)=(3*TSP)
330 R=NA/MCPO
340 PRINT “THE SODIUM TO PHOSPHATE RATIO IS ” ;:PRNT USING “##.##”; R
*The inleakage of contaminants was not the result of a failed condenser

tube. The condenser hotwell is equipped with a drain line that discharges to the
cooling water outlet. During the outage, the valve was opened to drain the con-
denser steamside, but plant personnel forgot to close it before unit start-up. Once
the unit was placed in service, cooling water was pulled back into hotwell by the
condenser vacuum. When the operators discovered the problem, they simply
closed the valve. Unfortunately by that time the damage had been done.

03.qxd 8/8/06 1:52 PM Page 78
Case History 3-2

This case history illustrates the rapidity with which a condenser leak can
affect boiler water chemistry. At the time of this particular upset, the only on-line
chemistry monitoring for the entire unit consisted of a sodium analyzer at the
condensate pump discharge (CPD). The monitor was not equipped with an
alarm. On the morning of the upset the operator’s log indicated that at 7:00 a.m.,
the concentration of sodium in the CPD was less than 1 ppb. At approximately
7:45 a.m., when plant chemists were collecting morning grab samples, they dis-
covered that the sodium monitor was pegged out at a reading of over 100 ppb.
A subsequent grab sample analysis, which lab personnel conducted within 15
minutes of this discovery, revealed that the boiler water (which was on congru-
ent phosphate control) contained no phosphate and had a pH of 5.8! The chem-
istry manager immediately notified the operators, and the unit was taken out of
service as quickly as possible. During this time the plant staff injected trisodium
phosphate into the drum to stabilize the pH and establish a phosphate residual,
and the operators blew down the drum and lower headers to remove solids.
Once the unit was off-line, maintenance personnel discovered the problem.
A plug had fallen out of a topmost condenser tube, which had previously failed
due to steam erosion. The tube, which had massive failures throughout its length,
allowed great quantities of cooling water to pour into the condenser. These cont-
aminants quickly consumed the phosphate in the boiler water, and then by the
process outlined in Equation 3.3 of this chapter, caused a pH depression.
This story had a good ending. The quick action by the plant staff prevented
serious deposition of solids on the tube walls. The boiler was chemically cleaned
at the earliest available opportunity. No tube failures occurred after this upset.
However, the event proved to be a catalyst in convincing utility managers to
install a comprehensive on-line water chemistry monitoring system.
Case History 3-3

I was given project management responsibility for the on-line water chem-
istry monitoring system referenced in Case History 3–2. The system analyzes
conditions in three separate generating units. Drum analyzers include pH and
phosphate monitors, whose readings are needed for the BASIC program outlined

03.qxd 8/8/06 1:52 PM Page 79
in Supplement 3–1. The monitoring system also is equipped with a PLC that
controls valve operation and collects data from the instruments for distribution
to remote computer terminals. PLC logic includes common arithmetic functions,
so I converted the calculations in the BASIC program to ladder logic and sent the
output to PLC registers. The values were then transmitted to the remote display
screens along with other water chemistry data.
This data distribution concept could be used for other boiler water treat-
ment programs. For any future applications involving congruent chemistry, the
effects of ammonia or amines would make the calculations more complicated. A
potential solution is to monitor both sodium and phosphate on-line and set up
the PLC logic to directly calculate sodium-to-phosphate ratios.
Case History 3-4

3,500 psig (supercritical)
Plant personnel switched from AVT to OT primarily because the unit had to
be chemically cleaned every 18 months to 2 years. They selected gaseous oxygen
as the oxidant, and installed injection quills at the condensate polisher outlet and
the deaerator outlet. An oxygen flow rate of 2.2 SCFH maintained dissolved oxy-
gen levels at 50 to 150 ppb through all load conditions. When plant personnel
started up the system, they lowered ammonia feed to establish a pH range of 8.0
to 8.5 in the feedwater versus the 8.8 to 9.6 with AVT.
The staff made the following observations after OT was begun.
• The condensate piping was overlayed with the protective FeOOH film in
about one day, but it took almost a week for the feedwater piping to be
converted
• The cation conductivity of the condensate polishers remained below the
utility guideline of 0.15 MS
• Dissolved iron concentrations dropped from a range of 9.0–9.7 ppb to
around 3 ppb
• The formerly black color of the pipe surfaces (magnetite) became brown-
ish orange due to the formation of FeOOH
So far, the plant staff has been very pleased with the program.

03.qxd 8/8/06 1:52 PM Page 80
Case History 3-5

Conditions: Drum-type Cyclone boilers
Cooling-coil steam attemperators
Two 33 MW sister units were each equipped with cooling-coil steam attem-
perators located in the boiler mud drum. (See Fig. 3–18 at the end of this case
history.) In both units, the attemperators began to corrode and frequently fail
within one to two years after the boilers were converted from baseload to cycling
operation. The corrosion was confined to the inlet header and approximately the
first ten feet of the inlet tubing, which is the hottest portion of the bundles.
Various solutions to the problem were proposed by utility personnel and
consultants. These included replacing the carbon steel with higher alloy materi-
als and placing sleeves around tubes in the affected zones. The plant results staff,
including myself, looked for a simpler answer. A couple of solutions became
readily apparent after we examined the corrosion and the history of boiler oper-
ation. The corrosion appeared as exfoliation and pitting of the tube surfaces
exposed to the mud drum water. This suggested oxygen corrosion exacerbated
by possible deposit formation on the hottest portion of the tubes. Our suspicions
of oxygen corrosion were further reinforced by two pieces of data. First, the cor-
rosion did not really become problematic until the boilers were placed in cycling
operation. When this operational change was made, the boilers were usually
operated only in the summer and a few of the coldest days in winter, or when
one of the larger units tripped off-line. However, these cycling units had to be
available at all times, and thus were laid-up wet. Layup procedures prior to our
investigation consisted simply of pumping standard quantities of hydrazine and
cyclohexylamine into the drum, where the only mixing would be by diffusion of
the chemicals into the water. We installed a simple circulation system on the boil-
er to pull water from the bottom of the boiler headers and recirculate it to the
steam drum (Fig. 3-19). This drastically improved distribution of the chemicals
throughout the boiler water circuits. During our investigation we also found that
the hydrazine/amine feed systems were only capable of providing 13 and 15
hours of continuous chemical feed to the respective units. This meant that when
the boilers were in operation, hydrazine and cyclohexylamine were only being
introduced for slightly more than half a day. We redesigned the systems to allow
for 24-hour continuous feed.
These modifications seem to have cured the attemperator corrosion prob-
lems and undoubtedly improved conditions in the boiler tubes as well. (These
boilers had been plagued by frequent waterwall leaks.) No attemperator leaks
have occurred in the five years since the chemistry changes were incorporated.
The modifications were certainly less expensive than other proposed solutions.

03.qxd 8/8/06 1:52 PM Page 81
Figure 3-18
MUD DRUM
STEAM INLET
HEADER
STEAM
OUTLET
Outline of Mud Drum (Cooling Coil Attemporator Arrangement)
Figure 3-19
Flow Diagram of Layup Solution Recirculation System

03.qxd 8/8/06 1:52 PM Page 82

04.qxd 8/8/06 1:54 PM Page 83
Chapter 4
Steam
Chemistry
Introduction
The previous chapters have outlined the importance of feedwater and boil-
er water chemistry. Although control of chemistry in these systems is critical for
protection of the systems themselves, feedwater and boiler water chemistry have
a direct effect on steam purity. Deposition of solids and corrosion in superheaters
and turbines have been common problems. Even in industrial plants that have
no turbines, proper control of steam chemistry is still very important. Some of
the most intensive research in the steam generating industry is being directed
towards steam chemistry.
Steam produced in the boiling process is never completely pure, and even
at its best contains trace amounts of solids. The process by which solids are trans-
ferred to steam is known as carryover. Carryover is influenced by the following:

04.qxd 8/8/06 1:54 PM Page 84
• Virtually all solids are at least slightly soluble in steam. Solids become
more soluble as boiler pressure increases, and some, particularly silica,
carry over extensively as a vapor.
• Even with the best steam separating devices, moisture droplets still enter
the steam.
• Sulfate and chloride will enter steam as ammoniated salts.
• Improper drum level control, poor drum design, or excessive solids
buildups in drum water will increase carryover.
Contaminants also enter steam via attemperator systems. Even if the drum
is operating properly and minimizing carryover, dissolved solids still have this
direct path to the steam system. Attemperator contamination is greatly exacer-
bated during upset conditions such as a condenser leak.
Another mechanism for steam contamination is exfoliation of superheater
and reheater tube walls, which introduces iron oxide particles to the steam.
Primary Carryover Products

The most commonly found turbine deposits or corrodents include silica,
magnetite, copper, sodium chloride, hydroxide, and phosphate, organic acids,
and ammonium chloride and sulfate. Figure 4-1 shows the relative solubility of
these and other compounds in steam. As the diagram indicates, the solubility of
every compound increases with increasing steam pressure. Likewise, the solubil-
ity decreases with decreasing steam pressure. The change from high to low pres-
sure is precisely the condition in a turbine. The solubility characteristics of each
compound dictate the location where the contaminant drops out of solution.
Figures 4-2 and 4-3 illustrate the regions in the turbine where this occurs. In
general, most compounds concentrate in the low-pressure end of the turbine.
One particularly problematic area is the saturation zone. As steam passes through
the low-pressure stage of the turbine, a fraction of it condenses. The moisture
aids in the concentration of contaminants, particularly those which are water sol-
uble. Corrosion of L-1 and L-2 blades has been a common and well-document-
ed problem.
The following subsections outline the difficulties caused by the most com-
mon carryover products.
Copper
Copper tends to precipitate in the high-pressure stage of the turbine, where
it forms copper oxide or metallic copper deposits. Over time, the deposits will
affect turbine performance, and many cases of unit derating due to copper depo-

04.qxd 8/8/06 1:54 PM Page 85
Steam Chemistry 85
Figure 4-1
Steam Solids Concentration versus Steam Pressure. Copyright © 1986. Electric Power Research Institute.
sition have been recorded. This phenomenon has been under close scrutiny in
recent years, as chemists strive to understand the mechanisms and magnitude of
the problem.
Sodium Hydroxide
Sodium hydroxide is soluble in water and steam, and exists as a dissolved
compound in both phases. Sodium hydroxide is a primary culprit in stress cor-
rosion cracking (SCC), and, as Figure 4-3 shows, may cause turbine blade cor-
rosion in the area between the inlet of the low-pressure turbine and the begin-
ning of the saturation zone. Turbine blades are naturally under stress due to oper-
ating conditions, and thus are susceptible to SCC.

04.qxd 8/8/06 1:54 PM Page 86
Figure 4-2
Solubility Characteristics of Various Carryover Products in the Steam Cycle. Copyright © 1987.
Electric Power Research Institute. EPRI CS-5275. Water, Steam, and Turbine Deposit Chemistries in Phosphate-
Treated Drum Boilers. Reprinted with permission.
Chloride and Sulfate

Several compounds, including chloride and sulfate salts, concentrate in a
narrow region defined as the salt zone. Chlorides induce pitting in stainless
steels, which are the material of construction for turbine blades and rotors.
Chlorides and sulfates also cause corrosion fatigue and SCC of turbine blades.
These two corrosion phenomena cost U.S. utilities hundreds of millions of dol-
lars per year in replacement costs and lost power.
Chloride and sulfate introduction into steam is influenced by the ammonia
concentration in the boiler water, which allows the compounds to carry over as
ammonium salts. Higher ammonia levels increase chloride and sulfate carryover.
Although chlorides and sulfates are themselves harmful to turbine blades,
the salts can become acidic and cause additional corrosion. In fact, acid corro-
sion of turbine blades has been observed to be much higher in AVT units than
phosphate-treated units. Evidence indicates that in drum-type units sodium
phosphate carryover helps to neutralize acidic compounds that may form on the
turbine components.
Iron Oxides
Iron oxides transported by high-velocity steam cause solid particle erosion
(SPE) of the turbine blades. Turbine screens remove some of these particles, but
small particles may pass through the screens and erode the edges of the high-
pressure turbine blades. Exfoliation of superheater piping generates many of
these iron oxide particles.

04.qxd 8/8/06 1:54 PM Page 87
Steam Chemistry 87
Figure 4-3
Concentration Zones of Contaminants in a Turbine. Copyright © 1986. Electric Power Research Institute.
Silica
Silica will coat turbine blades and gradually cause a decrease in turbine per-
formance. Although it is sometimes possible to water wash silica from the blades,
some of the deposits may be very tenacious. Chemical cleaning is required to
remove this adherent scale.
Sodium Phosphates
Sodium phosphates will carry over and deposit on intermediate- and low-
pressure turbine blades. If the deposits are not too heavy this can actually pre-
sent an advantage, as the phosphates help neutralize acids formed by other car-
ryover products. However, excessive phosphate carryover is detrimental. Many
cases of phosphate carryover have been reported, with some causing reheater
tube failures due to sodium phosphate accumulation in reheater U-bends.
Organics
Breakdown of organics in the boiler or steam lines introduces organic acids
and carbon dioxide to the turbine. These acids can cause corrosion in the low-
pressure blades. At chemical plants, excess organics in condensate return have
been known to either carry over to steam directly or influence boiler water chem-

04.qxd 8/8/06 1:54 PM Page 88
istry and drum operation with resultant superheater deposition problems.

Carbon dioxide produced by organic decomposition also can generate carbonic
acid corrosion in condensate return systems.
This list is not exhaustive, but it provides a clear example of the problems
that carryover can cause in the turbine. Case History 4-1 outlines an extreme
example of carryover. Should corrosion or deposition become too severe, turbine
failure may result. Not only does this pose a safety hazard, but turbine outages
are very expensive. The next sections outline in greater detail the principal meth-
ods by which solids enter steam.
Mechanical Carryover
During the steam generation process in drum boilers, water droplets become
entrained in the steam. If the droplets are allowed to remain, they will introduce
dissolved solids to the superheater, reheater, and turbine. Boiler drums, there-
fore, contain steam separating devices to remove the bulk of the entrained mois-
ture. Figure 4-4 shows a simplified diagram of a common steam separating
scheme. The separation equipment relies upon the difference in density between
steam and water. In the design shown, steam first passes through cyclonic sepa-
rators, which throw water droplets to the cylinder walls where they drain back
into the drum. The steam then passes through chevron-like scrubbers that
remove additional water. Other steam drum internals may represent those shown
in Figure 4-5.
Even these devices
Figure 4-4
will not prevent steam
STEAM contamination if boiler
OUTLET
water chemistry or operat-
ing conditions are not
> < properly controlled. The
most common factors that
< SCRUBBERS >
influence mechanical car-
>
<
ryover include foaming,

priming, improper drum
< CYCLONE > water level control, poor
SEPARATORS
drum design, or operation
< >
outside of boiler design
STEAM/WATER limits.
Priming is the intro-
duction of slugs of water to
the steam. This frequently
occurs during large and
rapid load swings, when
Generic Outline of Steam Drum Internals.

04.qxd 8/8/06 1:54 PM Page 89
Steam Chemistry 89
Figure 4-5
Inside of a Steam Drum. Photo taken with permission of the Kansas City, Kansas, Board of Public Utilities,
Quindaro Power Station.
the demand for steam changes more quickly than the unit can supply it. Priming
can also be caused by poor drum level control, wherein high drum levels over-
load the steam separating equipment. Foaming results from excessive concentra-
tions of solids in the boiler water. As the foam bubbles burst, they introduce
excessive quantities of water to the steam separating devices, which may not be
able to totally remove the added moisture. Foaming may be more prevalent at
industrial facilities, where makeup treatment systems are less sophisticated and
condensate polishing is minimal or nonexistent. Case History 4-2 outlines
superheater deposition problems at a chemical plant, that were believed to have
been caused by introduction of excessive organics to the steam boilers.
Inadequate drum design will influence all of the phenomena mentioned
above. One of the most important factors in drum design is that sufficient space
be provided between the operating water level and the steam separation devices.
If the drum is too small, slugs of water or excess water vapor will enter the steam
separating devices. Operation of the boiler at higher than rated capacity can also
overload the steam separating internals.
Proper operation of the steam separating devices is critical for production of
high-purity steam. Maintenance personnel should inspect the internals during
every planned maintenance outage where the steam drum is opened and repair
those components which may be damaged or out of position.

04.qxd 8/8/06 1:54 PM Page 90
Vaporous Carryover
Figure 4-1 illustrated the increase of solids solubility with increasing steam
pressure. As unit pressure increases, boiler water chemistry must be more tight-
ly controlled to restrict vaporous carryover. Ultimate limits are reached in once-
through units, where concentrations of solids must be below steam saturation
values so that none of the solids precipitate in the waterwall tubes.
Silica is the most well-known vaporous carryover product. It is also one of
the most troublesome for several reasons. First, silica can form tenacious deposits
on turbine blades, that often can only be totally removed by chemical cleaning.
Second, the vaporous carryover potential greatly increases with increasing boiler
pressure. For example, in a 900 psig boiler without reheat, the recommended
maximum boiler water silica concentration is 7 ppm. In a 2400 psig boiler with
reheat, the recommended maximum is less than 0.2 ppm! Third, silica can easi-
ly enter a steam generating unit. It is the most weakly held ion on makeup sys-
tem anion resin and will break through first. More importantly, silica can exist in
colloidal form in water. The colloids will pass almost untouched through a dem-
ineralizer, but will break down to reactive silica from the heat of the boiler water.
Vaporous carryover of copper oxides is another area of concern. This pre-
dominantly occurs at pressures over 2300 psig and can be quite troublesome in
high-pressure units. Even a few pounds of copper on turbine blades may be
enough to decrease the turbine capacity by several megawatts. The corrosion
control procedures outlined in chapter 2 are very important for prevention of
copper transport from the feedwater system to the boiler to the turbine. This
includes off-line corrosion prevention. A number of cases of severe copper cor-
rosion of feedwater heater tubes during unit outages have been reported. In some
instances, a subsequent chemical cleaning of the boiler generated several thou-
sand pounds of copper in the cleaning solution.
Table 4-1
Recommended Steam Turbine Purity Limits

Westinghouse General Electric
Parameter
Sodium, ppb <5 <3

Cation Conductivity, S <0.3 <0.2
Chloride, ppb <5
Silica, ppb <10 *
Copper, ppb <2 *
Iron, ppb <20
*GE suggests that boiler manufacturer’s limits are satisfactory.
Steam Turbine Manufacturer’s Recommended Steam Turbine Purity Limits.

04.qxd 8/8/06 1:54 PM Page 91
Steam Chemistry 91
Solids Introduction by Contaminated

Attemperator Water
Steam temperature control is important for efficient operation of a steam
generating unit. Changes in operation, load, and firing patterns require that some
form of external steam temperature adjustment be available, especially for high-
pressure units. This is often accomplished by direct injection of water into the
superheater and reheater sections of the unit. Typically, feedwater is used for this
purpose. However, as has been illustrated in chapter 2, feedwater may be conta-
minated with iron or copper corrosion products, or compounds introduced by
condenser leaks, demineralizer overruns, or poor condensate polisher regenera-
tion. Attemperation can therefore be a direct source of contamination to the tur-
bine. It is suspected that many corrosion failures in turbines were not due to
buildups of contaminants during normal operation, but were caused by contam-
inant ingress during transient periods. A likely source for such transients is the
attemperator system, especially for units without condensate polishers.
Superheater Exfoliation
The stresses experienced by steam generating equipment during shutdown,
start-up, and cycling duty increase the mechanical degradation of the equipment.
This occurs in the superheater and reheater, where the stress and harsh environ-
ment cause exfoliation of the magnetite layer on the tube surfaces. These parti-
cles may then pass to the turbine where they can cause erosion of the turbine
blades. This effect, known as solid particle erosion, is most commonly found in
the high-pressure end of the turbine where linear velocities are highest. Case
History 4-4 illustrates some of the products that may be found in superheater
and reheater tubes due to exfoliation and mechanical carryover.
Sometimes, corrosion of afterboiler components may occur due to excess
oxygen in the steam. A prime location for this corrosion is the crossover line
between the high-pressure and low-pressure ends of the turbine, where any
slight amount of moisture helps generate the corrosion reaction. Some utilities
inject an oxygen scavenger into the crossover line to prevent oxygen attack of
carbon steel components in this area and in the low-pressure turbine. This brings
up an interesting point regarding oxygenated treatment. Research to date indi-
cates the absence of oxygen corrosion in the turbine. This is probably due to the
carefully controlled conditions of the program.

04.qxd 8/8/06 1:54 PM Page 92
Water Chemistry Limits to Prevent Steam

Contamination by Carryover
Proper control of boiler water and steam chemistry is essential to protect
turbines and other afterboiler components from corrosion and failure. This is one
area of steam generation chemistry that is being heavily researched. One target is
the chemistry of early condensate, which is the solution produced when steam
first begins to liquefy to water. Evidence indicates that contaminants in early con-
densate may concentrate severalfold over their levels in steam. This increases the
potential for corrosion.
Recommended steam chemistry limits have become tighter as more infor-
mation becomes available on carryover and the mechanisms of contaminant
deposition and corrosion. Table 4-1 shows steam purity limits established by two
of the major turbine manufacturers. These are compared to current guidelines
recommended by EPRI (Table 4-2). Many of the values are similar. The very low
values are a reflection of the deleterious effects of contaminants on turbines. EPRI
advises that the limits may be lowered as further research is conducted. Whether
lower limits are achievable remains to be seen.
Boiler Water Chemistry Guidelines for Control

of Steam Chemistry
In the absence of contaminant introduction by the attemperator system,
solids can only enter steam by carryover. Since carryover is pressure dependent,
the amount of solids that can be tolerated in boiler water decreases as pressure
increases. Table 4-3 (interpreted from EPRI graphs) illustrates maximum con-
centrations of sodium, silica, chloride, and sulfate in drum boiler water that will
still maintain steam chemistry below recommended limits. Table 4-3 reveals sev-
eral important details. First, the level of permissible chloride and sulfate is very
low. Second, silica takes the most precipitous drop of any constituent. This is
Table 4-2
Steam Equilibrium Phosphate Treatment Phosphate Treatment

Parameter All-Volatile-Treatment
Sodium 3 5
Cation Conductivity (/s/cm) <0.15 <0.3
Chloride (ppb) 3 3
Sulfate (ppb) 3 3
Silica (ppb) 10 10
TOC (ppb) 100 100
Recommended Steam Turbine Purity Limits for Utility Boilers. Source: Electric Power Research Institute.

04.qxd 8/8/06 1:54 PM Page 93
Steam Chemistry 93
Table 4-3
Recommended Boiler Water Concentrations to Meet
Steam Purity Guidelines
Boiler Pressure Sodium Silica Chloride Sulfate
(psig) (ppm) (ppm) (ppm) (ppm)
900 3.3 3.6 0.33 0.33
1100 3.0 1.9 0.24 0.24
1300 2.7 1.3 0.17 0.17
1500 2.5 0.81 0.13 0.13
1700 2.2 0.57 0.086 0.086
1900 1.9 0.39 0.065 0.065
2100 1.6 0.27 0.048 0.048
2300 1.2 0.16 0.037 0.037
2500 0.71 0.14 0.028 0.028
2700 0.44 0.085 0.020 0.020
2900 0.27 0.051 0.014 0.014
Author’s Interpretation of Graphs in the EPRI Interim Consensus Guidelines for Fossil Plant Cycle
Chemistry. Adapted from Interim Consensus Guidelines on Fossil Plant Cycle Chemistry. Electric Power Research
Institute. EPRI CS-4629. Copyright © 1986.
reflective of the vaporous carryover quality of silica. In high-pressure boilers it

can barely be tolerated.
The limits in Table 4-3 are not absolute for any particular system. Each boiler
has its own subtleties, which may require tighter, or allow slightly looser, control.
Prevention of Contamination via the Attemperator

System
The best method to prevent solids introduction to steam by contaminated
attemperator water is to polish the condensate. Many plants are not equipped
with condensate polishers, so protection of the steam system is dependent upon
careful monitoring and control of feedwater chemistry. Plant personnel should
follow the feedwater chemistry guidelines outlined in chapter 2. If a condenser
tube leak develops, it should be repaired as quickly as possible. Other potential
sources for attemperator water contamination include demineralizer overruns
and improper regeneration of condensate polishers.
Control of Superheater and Reheater Exfoliation

The stresses imposed on superheater and reheater tubes will eventually
cause cracks in the magnetite layer that forms on the internal wall surfaces.
Pieces of iron oxide will break off (exfoliate), upon which they can be carried by
the steam to the turbine. Pieces that are small enough to pass through turbine
screens then impinge upon the turbine blades and cause physical damage. Solid
particle erosion (SPE) can be readily observed when the turbine is opened for

04.qxd 8/8/06 1:54 PM Page 94
inspection. The edges of the turbine blades exhibit an abraded pattern. This
problem can become more acute in cycling units because frequent startups, shut-
downs, or load changes impart additional stress to the tubes.
Superheater/reheater exfoliation is mostly a mechanical phenomenon, and a
chemical cleaning of the tubes may be necessary to remove the fractured mag-
netite layers. Superheater chemical cleaning is not simple, and requires tempo-
rary piping, well thought out procedures, and the services of a reliable chemical
cleaning firm. However, superheater/reheater cleaning may still be cost effective
if SPE is a severe problem. For boilers with drainable superheaters, it is possible
to remove some of the exfoliated material by water washing.
Steam Chemistry Monitoring

Steam chemistry monitoring and sample analyses are quite important. Prime
sample locations include reheat steam (or main steam for boilers without
reheaters) and saturated steam. The former sample allows the plant chemist to
monitor steam conditions after the attemperator. The appropriate parameters for
analysis include those constituents outlined in Table 4-2. On-line analyses are
recommended for sodium, cation conductivity, silica, and possibly chloride, with
grab sample analyses for sulfate and TOC. Continuous sampling of saturated
steam serves as a good backup to the main/reheat sample, and is the primary
sample when main or reheat sampling is unavailable.
Steam Chemistry Issues at Industrial Plants

without Turbines
Where steam is used for industrial processes without driving turbines, car-
ryover issues are still important, but may be influenced by somewhat different
factors than those for utility units. Many of these have been discussed in chapter
2. Boiler water and steam chemistry monitoring are still important, especially for
boilers with superheaters. Carryover can easily cause severe deposition in the
superheaters with resulting tube failures. The ASME guidelines outlined previ-
ously in Tables 2–3 through 2–5 provide good guidelines for feedwater and boil-
er water chemistry. Monitoring should at least include on-line sodium or
degassed cation conductivity. Proper training is also important. I have been to
facilities where the operators knew how to perform standard chemical tests, but
did not know what the results indicated. Steam chemistry was a mess, but the
operators were taking accurate readings! Training must not only include the
“how” of water chemistry monitoring, but also the “why.”
One of the most important issues regarding industrial steam generation is
the effect of steam purity on condensate produced later on in the process. Boilers

04.qxd 8/8/06 1:54 PM Page 95
Steam Chemistry 95
using softened water for makeup can generate prodigious amounts of carbon
dioxide. When the CO2 dissolves in the condensate, the corrosivity increases.
Treatment with neutralizing or filming amines is often required to protect down-
stream condensate piping.
Conclusion
Control of steam chemistry is vitally important for protection of turbines
and other afterboiler components. Intensive research is being conducted into the
effects of various compounds in turbines, and recommended steam purity limits
continue to become lower. Utility chemists should be aware that tightening
steam chemistry guidelines will require improved boiler water chemistry control
and boiler water/steam monitoring.

04.qxd 8/8/06 1:54 PM Page 96

04.qxd 8/8/06 1:54 PM Page 97
Case History 4-1

This case history describes how greatly a condenser leak can affect carryover
when the situation is not corrected in a timely manner.
Cooling water for the condenser is supplied from a cooling tower whose
makeup comes from a saline source. Cooling water TDS concentrations can reach
20,000 ppm. The incident occurred during the night shift, when a condenser
tube began leaking. The chemist who came on duty discovered severe contami-
nation throughout the boiler. The treatment products in the boiler had been con-
sumed and the pH of the boiler water had dropped to a very low level.
The plant chemistry staff immediately notified plant operators, but they did
not take the boiler out of service until the next day. When the turbine shroud was
removed, plant personnel discovered that virtually all of the turbine blades were
covered with heavy salt deposits. The subsequent cleanup costs of the turbine
and boiler, including extra outage time needed, were around $5,000,000.
This example graphically illustrates the effects that a major chemistry upset
can have on steam generating equipment including afterboiler components.

04.qxd 8/8/06 1:54 PM Page 98
Case History 4-2

Conditions: Organic chemical plant
Four 100,000 lb/hr, 550 psig package boilers
The boilers generate superheated steam for use in the production of phenol
and phenol-derivative compounds. Approximately 80% of the condensate is
returned to the boilers. Neutralizing amines are injected into the condensate for
corrosion control. Steam is only used for process functions and does not drive
any turbines.
Sporadic, but abnormally frequent, failures have been occurring in the
superheater tubes. When a tube fails, the company removes and replaces the
entire superheater, a rather expensive procedure. Tubes taken from the removed
sections contained internal deposits whose thickness ranged from 1/8 inch to per-
haps 1/4 inch. An analysis of these deposits showed that they were primarily com-
posed of iron oxide, silica, sodium, and carbon. Each of the boiler drums con-
tains steam separating internals, and the company even added an external set of
separating devices to one of the most problematic boilers. The problems, how-
ever, still persist.
An examination of water chemistry records revealed the probable source of
the difficulties—organics. For boilers of this pressure, the recommended maxi-
mum TOC level in the boiler water is 0.5 ppm. Prior water chemistry records by
an outside vendor indicated that TOC levels in the condensate return sometimes
reached the 20 ppm level. Plant personnel admitted that other tests had shown
that TOC levels in the condensate return had risen above 200 ppm at times. This
data immediately suggested foaming as a probability, and indeed foam was evi-
dent in the boiler water sample line discharge.
Plant personnel were given the following list of recommendations to track

down and correct the superheater deposition problem.
• Upgrade the sampling system so that accurate steam samples are available.
The current sampling taps have been placed in locations that do not pro-
vide representative samples.
• Upgrade analytical procedures so that plant personnel or an outside vendor
can determine all of the major constituents in the condensate return, boiler
feedwater, boiler water, and steam. For example, of the five major organic
chemicals produced at the facility, the operators can only analyze for phenol.
The procedure involves a colorometric determination, and the operators

04.qxd 8/8/06 1:54 PM Page 99
Steam Chemistry 99
report that the sample often develops a different color than that outlined in
the analysis procedure. This indicates contamination by other compounds.
• If the upgraded sample data confirms contamination by the condensate
return, install a condensate polisher. The type and capabilities of the pol-
isher would of course be designed around the constituency of the conta-
minants in the condensate return.
Case History 4-3

Conditions: Makeup water surface supply to a demineralizer pretreated
by a municipal water system
The facility is a combined water/electric utility. Makeup water to the electric

generating plant demineralizer (cation/anion/mixed-bed system) is supplied
from the water treatment plant. The source of supply to the water treatment plant
is a manmade lake.
The incident occurred following an extraordinarily heavy rainfall. Silica in
all of the utility boilers suddenly increased. Lab chemists routinely analyzed the
demineralizer effluent and had seen no increase in silica concentrations. The
problem was traced to colloidal silica, which had escaped removal in the water
treatment plant clarifiers. Most probably, the heavy rainfall agitated the lake and
stirred up very fine sediment including colloids.
Silica will polymerize into very small colloids that behave differently than
silica in solution (reactive silica). For one, the colloids are undetectable by the
standard colorometric method for reactive silica. The colloids also do not carry
an ionic charge and will pass through anion exchange resins. However, the col-
loids break down as they are heated in the feedwater system and boiler. Thus, it
is possible to detect very little silica in the makeup treatment system effluent, but
find significant quantities in the boiler water and steam.
The colloidal silica influx in this case disappeared a few days after the rain-
fall subsided. Other utilities have not been so lucky and are saddled with water
that always contains silica colloids. A method for removing the colloids is to
include reverse osmosis in the makeup treatment scheme.

04.qxd 8/8/06 1:54 PM Page 100
Case History 4-4

Conditions: 2400 psig electric utility boiler
Reheat unit
After 10 years operating time, reheater tubes in this unit began to fail. The
failures all occurred at the U-bends of the vertically hanging reheater pendants.
When the U-bends were removed they were found to contain a mixture of black
and white deposits, that in some places were perhaps an 1/8 inch or more in
depth. The material consisted of a mixture of exfoliated iron oxide and deposit-
ed sodium phosphate and was a result of two of the steam contamination mech-
anisms listed in this chapter. The deposits had not caused corrosion of the inter-
nal surfaces of the U-bends, but rather had acted as an insulator which caused
the tubes to overheat.
Solutions for deposition of this type include improved steam sampling to
better detect and control carryover, and periodic superheater chemical cleanings
to remove loose iron oxide. Some of the newer boiler water treatment programs,
such as equilibrium phosphate treatment, operate at lower levels of sodium
phosphate in the boiler water. This reduces the amount of phosphate carryover.

05.qxd 8/8/06 1:56 PM Page 101
Chapter 5
High-Purity
Makeup Water
Treatment
Introduction
As the preceding chapters have indicated, the water supplied to a steam gen-
erating unit must be pure in order to prevent serious corrosion or scaling. Ion
exchange had been the backbone of the high-purity water treatment industry, but
with the improvement of membrane technologies, a variety of methods now
exists to produce pure water. In most cases, some form of pretreatment is need-
ed ahead of the makeup system to protect it from fouling, scaling, or microbio-
logical contamination.
Pretreatment
Water has been called the closest thing to a universal solvent because it will
dissolve, at least to some extent, most compounds. Raw waters, surface waters in

05.qxd 8/8/06 1:56 PM Page 102
particular, also contain suspended solids including colloidal particles, microor-

ganisms, and organic complexes. These compounds and suspended solids will
cause problems in high-purity treatment systems, and must be removed or
reduced before final purification. While an in-depth examination of pretreatment
methods is outside the scope of this book, an overview of these techniques
reveals their capabilities.
Figure 5-1
FEED >
>
CLARIFIER MEDIA CARBON

FILTER FILTER
BIOCIDE
Flow Schematic of a Common Pretreatment Process
Figure 5-1 shows a flow schematic of a common pretreatment process. The

methods outlined in this scheme are:
• Microbiocide feed
• Clarification/softening
• Media filtration
• Activated carbon filtration
Each provides a very important function.
Microbiocide Feed
Chlorine has been the principal microbiological control agent for many
years, and has proven to be the most economical biocide. However, use of chlo-
rine gas has been curtailed due to safety and environmental concerns. Popular
replacement feed chemicals now include sodium hypochlorite (liquid chlorine),
bromine, chlorine dioxide, and in some cases ozone. (UV light is also a possibil-
ity. Supplement 5-1 outlines a UV light arrangement at a manufacturing facility).
The mechanisms by which chlorine and other oxidizing agents disinfect water
are described in chapter 6. The oxidizing biocide should be added in great
enough strength to maintain a slight residual (0.1 to 0.5 ppm) throughout the
pretreatment process. This helps prevent the growth of microbes in equipment
downstream of the injection point. Chlorine and other oxidants will attack ion
exchange resins and some types of reverse osmosis membranes, making it nec-
essary to remove the oxidant ahead of these devices. Activated carbon is an excel-

05.qxd 8/8/06 1:56 PM Page 103
High-Purity Makeup Water Treatment 103
lent chlorine scavenger, but if the system contains no carbon filters, some other
method of oxidant removal is required. Most commonly, a dehalogenating or
reducing agent (sodium sulfite or sodium bisulfite) is injected ahead of the dem-
ineralizer or reverse osmosis unit to protect the equipment.
Clarification and Softening

Clarification and softening are often carried out in a single process unit, and
Figure 5-2 outlines a simplified version of a common clarifier configuration. The
water may first be allowed to pass through a grit or settling chamber to allow
large particles to settle on their own. Many particles, however, are truly sus-
pended and will stay in the water indefinitely. These particles tend to develop a
negative electrical charge, which keeps the solids separated and in suspension. A
two-stage process, coagulation-flocculation, is employed in the clarifier to bring
the particles together in a settleable solid. In the first step, a coagulant is added
for charge neutralization. The coagulant may be an inorganic salt such as alu-
minum sulfate [Al2(SO4)3 • 18H2O], sodium aluminate (Na2Al2O4), ferrous sul-
fate [Fe2(SO4) • 9H2O], or ferric chloride (FeCl3). These salts dissolve in water
to produce positively charged cations that neutralize the negative charge on the
suspended solids and allow them to draw together. Cationic organic polymers,
Figure 5-2
Mixer Motor
Chemical Feed
Raw Water
<
>
Clarified Water
Top of Sludge
Blanket
>
<
Sludge
Simplified Clarifier Design.

05.qxd 8/8/06 1:56 PM Page 104
such as polyquaternary amines, also serve as coagulants. The coagulant is added

into the rapid mix zone of the clarifier, which agitates the water to mix the coag-
ulant and suspended solids. Fast mixing is important to enhance the coagu-
lant/particle interaction.
As the negative charges on the suspended solids are neutralized, the parti-
cles begin to attach to one another through Van der Waals attraction. However,
the particles normally do not become large enough in this process to settle rapid-
ly. A flocculant is added to enlarge the coagulated particles and improve the set-
tling rate. Nonionic or anionic polymers are the most effective flocculants.
Polymers with an amide functional group (Fig. 5-3) are common nonionic floc-
culants, while polymers with carboxylate functional groups (Fig. 5-4) are the
anionic counterpart. The polymers collect the coagulated particles into large,
fluffy flocs.
Figure 5-3 Flocs can be destroyed by strong agitation, so the incom-
O ing water is allotted only a short period of time in the rapid
C mix zone. It flows from the rapid mix zone into more slowly
OH agitated zones. As is illustrated in Figure 5-2, the flow even-
Amide Functional
Group
tually proceeds from the mixing zone to the outer circumfer-
ence of the clarifier. A blanket or layer of previously coagulat-
Figure 5-4
ed and flocculated solids is maintained in this outer region at
O
all times. The layer traps the newly formed flocs that have
C
emerged from the mixing zone. Clarified water flows out of
NH2
Carboxylate
the top of the basin, where it is collected in a trough for dis-
Functional Group. tribution to the plant.
Hardness can also be removed in the clarifier by the addition of lime
[Ca(OH)2] and soda ash (Na2CO3). Reactions for hardness reduction are as follows:
Carbonate hardness removal

Ca(HCO3)2 + Ca(OH)2 2CaCO3 + 2H2O (5.1)
Mg(HCO3)2 + 2Ca(OH)2 Mg(OH)2 + 2CaCO3 + 2H2O(5.2)
Noncarbonate hardness removal

CaSO4 + Na2CO3 CaCO3 + Na2SO4 (5.3)
MgSO4 + Ca(OH)2 Mg(OH)2 + CaSO4 (5.4)
Silica is removed as magnesium silicate. The amount of silica that can be

taken out is dependent upon the magnesium concentration. Supplemental mag-
nesium salts such as oxide or carbonate may be added to enhance silica removal.
Clarifier/softeners are designed to give the incoming water enough time for
the coagulation/flocculation and softening reactions to reach maximum efficien-
cy. The clarifier is typically sized to produce a rise rate of the treated water with-

05.qxd 8/8/06 1:56 PM Page 105
in the range of 0.5 to 1.5 GPM/ft2. Retention times may range from 1 to 4 hours.
The effluent quality of clarified water may vary depending on influent quality, but
with proper operation a clarifier can reduce calcium to 35 to 40 ppm (as CaCO3).
Turbidities of less than 10 nephelometric turbidity units (NTU) should be
expected.
While inground circular clarifiers are often specified for large applications,
clarification/softening of water at flow rates of up to 1000 GPM or even more can
be accomplished in individual or parallel package clarifiers. These units are usu-
ally of rectangular design. The water is introduced to the rapid mix zone, flows
to a flocculation compartment, and then enters the main body of the clarifier. To
improve settling, the main clarifier compartment is equipped with a series of
inclined plates or tubes. The water is forced to flow up these tubes to the outlet.
The interaction of particles flowing upwards with heavier particles settling down-
wards enhances the contact between newly formed floc and the heavier solids.
Residence times in package clarifiers may be much shorter than in circular clarifiers.
An attractive feature of package clarifiers is cost. Package clarifiers are much
less costly than in-ground units, where installation costs may be significant.
When only pretreatment of the makeup influent is required, package units can
be quite attractive. Systems of less than 600 GPM or so are usually shipped as
one unit. Even large systems can be shipped in just a few pieces and be field
erected.
Filtration
With the proper selection of coagulants and flocculants, a clarifier will
remove most suspended solids. However, enough solids remain to potentially
clog ion exchange resins or RO membranes. These particles are removed by pass-
ing the effluent through a filter, usually containing several media. A common
multimedia filter arrangement contains anthracite, sand, and garnet in graded
layers (Supplement 5-2). Coarse material is placed at the top to remove large par-
ticles, with fine media below to remove small particles. These filters may operate
at up to 5 GPM/ft2 and produce a water with turbidities of less than 1 NTU.
Proper filter operation is critical to the performance of high-purity makeup sys-
tems, especially reverse osmosis units.
A common guideline recommends backwashing of the filter when the dif-
ferential pressure increases by 10 psi over normal. Backwash flow rates of 14 to
16 GPM/ft2 are typical, although the manufacturer’s guidelines must be followed
during this very important step. A poor backwash can generate mudballs in the
filter, which will then impair subsequent operation. Water density changes with
temperature, so backwash flow rates may have to be adjusted at various times
during the year.

05.qxd 8/8/06 1:56 PM Page 106
Activated Carbon Filtration

Activated carbon filters are placed ahead of demineralizers to remove oxi-
dizing biocides and large organics that may have passed through the clarifier.
Although the filters are very effective at removing these compounds, they also
serve as excellent breeding grounds for bacteria that survived the biocide treat-
ment. The bacteria can be a great problem. The carbon filter effluent should be
checked regularly for organic or microbiological content. One method for com-
bating this problem is to equip the filter with an auxiliary steam line taken from
a plant source. Periodic introduction of steam to the filter will sterilize the acti-
vated carbon. If sterilization of the bed is not possible, it will probably have to
be replaced on a fairly frequent interval. Six months is recommended.
Activated carbon is efficient due to its tremendous absorptive capacity. It
may be produced from a number of different materials, and one type of activat-
ed carbon may not have the same properties as another type. Supplement 5-3
offers guidelines on the selection of activated carbon media for makeup water
pretreatment.
Additional Pretreatment Methods

Other pretreatment methods are available to prepare water for purification
in a demineralizer or reverse osmosis unit. For example, if the water contains
high levels of iron and manganese, or dissolved gases including carbon dioxide,
ammonia, and hydrogen sulfide, it can be aerated to remove these contaminants.
If levels are too high, other oxidation methods such as manganese greensand may
be needed.
For hardness removal from small volume flows, hot-lime softening is a
viable treatment. This method has been fairly popular for industrial applications.
Hot-lime softeners employ the same chemistry as cold-lime units, with the
exception that the incoming water is heated with steam or some other source in
a closed vessel. This greatly enhances the chemical reactions. Hot-lime softening
can reduce hardness and silica to the following levels:
Calcium - 15 ppm
Magnesium - 5 ppm
Silica - 1 to 2 ppm
High-Purity Makeup Treatment Methods

The most important and commonly used makeup treatment methods are
ion exchange and reverse osmosis. These technologies are mature. Other tech-

05.qxd 8/8/06 1:56 PM Page 107
nologies that have begun to emerge and show promise include electrodialysis
reversal (EDR) and electrodeionization (EDI).
Ion Exchange
The ion exchange process was first truly developed in the early 1900s by a
German scientist, Gans, who used synthetic zeolites (sodium aluminum silicates)
to exchange calcium and magnesium ions for sodium. This discovery led to the
development of water softeners. Then, in the 1930s and 1940s, plastic synthetic
ion exchange resins became available, which led to the industry that we know
today.
The ion exchange process is dependent upon the efficiency with which
water can flow through the exchange media and contact exchange sites. To
obtain maximum efficiency, the media must have several properties. These
include:
• It must have the maximum number of exchange sites

• It must have structural integrity
• Easy loading of the media into the reaction vessel is required
• It must be of reasonable cost
• It must be capable of being regenerated efficiently
Small spherical plastic beads have Figure 5-5

proven to be best for incorporating C C C C C C C C C
these factors. The beads behave some-
what like a fluid, which makes them
easier to handle and load/unload into
C C C C C C
exchange vessels. By far, the most com-
C C C C C
mon material from which ion exchange
resins are fabricated is polystyrene
cross-linked with divinylbenzene (Fig.
5-5), although acrylics are used in some
Polystyrene-Divinylbenzene Structure.
instances. Beads are manufactured in an
emulsion process, which allows the polystyrene-divnylbenzene to form spheres.
The fabrication process and degree of cross-linking can be controlled to vary the
porosity within, and structural strength of, the beads. Resins are usually referred
to as either gel-type or macroporous (macroreticular). The former have a rela-
tively low divinylbenzene cross-linkage of perhaps 8% or so. The resin beads do
not have discrete pores, although in water they will swell and allow the passage
of ions through the organic structure. Macroreticular resins have divinylbenzene
cross-linkages of 12 to 20%, which impart a greater rigidity and strength to the

05.qxd 8/8/06 1:56 PM Page 108
resin and give the resin its porous structure. Resins of different porosity and
strength are useful in different environments. For example, macroporous resins
are more durable in heavy-duty applications. The gel-type resins typically con-
tain more exchange sites. Both types are heavily used in the water treatment
industry.
Macroreticular resins with large pores and no exchange sites have been
developed for removal of organic complexes from raw water. The resin functions
by allowing the organics to penetrate the bead, where they adsorb onto the poly-
mer chains. The resin is nonregenerable and is discarded when it reaches exhaus-
tion. These resins show promise for removing relatively small chain organics
from condensate, which are otherwise difficult to treat.
Exchange Groups
Once the beads have been synthesized, exchange groups are added. The
principal exchange groups are:
• Sulfonic acid (SO3-H+) strong acid cation resin

• Carboxylic acid (COO-H+) weak acid cation resin
• Quaternary amine (CH2N(CH3)3+OH-) strong base anion resin
• Primary, secondary, and tertiary amines (CH3NH2) weak base anion resin
The exchange groups give the resins their character and define their perfor-
mance in service.
Strong Acid Cation Resins

All of the exchange sites listed above can be likened in performance to their
acid or base counterparts. This is most clearly illustrated with a strong acid cation
(SAC) resin. SAC resins have an affinity for cations over hydrogen, and will
exchange hydrogen for them as water flows through the vessel.
Ca SO4 Ca++ H2SO4

Mg Cl + ~C – SO3-H+ ⇔ ~C – SO3 Mg++ + HCl (5.5)
Na HCO3 Na+ H2CO3
Because the resin behaves as a strong acid, it will split salts and separate
cations from their corresponding anions. A water containing the ions calcium,
magnesium, sodium, chloride, sulfate, bicarbonate, and silica will exit the vessel
as a dilute solution of hydrochloric, sulfuric, carbonic, and silicic acids. The
order of affinity the resin has for ions is Ca+2 > Mg+2 > Na+.

05.qxd 8/8/06 1:56 PM Page 109
Weak Acid Cation Resins

The order of affinity for ions on a weak acid cation (WAC) resin is H+ > Ca+2
> Mg+2 > Na+. Thus, WAC exchange sites tend only to release hydrogen ions to
an anion, such as bicarbonate, that has a strong affinity for H+. One may ask
then, “If a WAC resin does not remove all cations, what is its practicality?” The
answer is regeneration efficiency. Because a WAC resin has a strong affinity for
hydrogen, the resin is much more easily regenerated than a SAC resin. The
arrangement of a WAC exchanger followed by SAC exchanger can be very use-
ful for treating high-alkalinity waters.
Strong Base Anion Resins

Strong base anion (SBA) resins are the counterpart to SAC resins and will
remove virtually all anions. The resin exchanges hydroxide ions for the anions in
the preferential order SO4 = > Cl-> HCO3- > HSiO3-. The practicality of ion
exchange is illustrated by the reactions that occur when the cation exchanger
effluent flows through a SBA exchanger.
HCl Cl-
H2SO4 + CH2N(CH3)3+OH- ⇔ R+ SO4= + H2O (5.6)
H2CO3 CO3=
H2SiO3 SiO3-
The final product is water.

Two kinds of SBA resins are most common, Type I and Type II. Type I resins
contain quaternary amine exchange sites as shown in Equation 5.6. Type II con-
tains the quaternary ammonium functional group [CH2N(CH3)2CH2CH2OH].
Type I resins are more stable at higher temperatures. Type II resins are slightly
less basic and can be regenerated a bit more efficiently.
Weak Base Anion Resins

Weak base anion (WBA) resins serve a similar purpose to WAC resins; they
remove a portion of the ions so that a downstream SBA resin does not have as
heavy a load. Because the functional group of the WBA resin is a weak base (pri-
mary, secondary, or tertiary amine), the resin removes weak bases from solution,
principally chloride and sulfate. However, unlike WAC resins, WBA resins
remove the weak bases as their acid conjugates, e.g., H2SO4 and HCl. WBA resins
are efficiently regenerated and may be useful ahead of SBA exchangers if the

05.qxd 8/8/06 1:56 PM Page 110
water contains high levels of chlorides or sulfates.

WBA exchangers are also sometimes used ahead of SBA exchangers if the
water contains more than 1 or 2 ppm of natural organics. Decaying vegetation
produces complex organics that will foul anion beads. This effect is especially
pronounced for SBA resins. WBA resins will help remove these organics. The
organics regenerate more easily from WBA resins than SBA resins.
Demineralizer Configurations and Mixed-Bed

Exchangers
Table 5-1 illustrates various demineralizer vessel configurations. The reader
will note the mixed-bed (MB) demineralizer mentioned in several of the options.
Mixed-bed demineralization is the process that improved ion exchange perfor-
mance so that it became suitable for ultra-pure applications.
Water from a
Table 5–1 cation/anion system,
Common Demineralizer Configurations
although of high quality,
still contains too many
SAC/SBA SAC = Strong Acid Cation contaminants for use in
SAC/WBA/SBA SBA = Strong Base Anion high-pressure boilers or
WAC/SAC/WBA/SBA WAC = Weak Acid Cation as ultra-pure water at
SAC/SBA/SAC WBA = Weak Base Anion semiconductor manufac-
SAC/SBA/MB MB = Mixed Bed turing plants. A mixed-
bed exchanger contains
intimately intermixed SAC and SBA resins. The exchanger performs as if it con-
sisted of millions of miniature cation/anion units. A MB exchanger usually serves
as a polisher of effluent from a cation/anion system, reverse osmosis unit, or
other purification arrangement. The mixed-bed will reduce contaminants from
ppm levels to ppb levels.
Of the systems outlined in Table 5-1, the SAC/SBA/MB arrangement has
been most popular, especially for high-purity applications including feed to boil-
ers that operate above 1000 to 1500 psi. Selection of the best makeup system for
lower-pressure units is a bit more complex. A system with a MB exchanger would
obviously provide water suitable for industrial and small utility boilers, but this
represents excessive treatment. At pressures below 600 psi, sodium softening is
often adequate for makeup treatment. The softener contains SAC resin in the
sodium not hydrogen form, as only hardness ions need be removed from the
water. The great advantage of this process is that the resin can be regenerated
with a brine solution rather than more expensive acid. Sodium softeners may be
used with hardness-removing pretreatment equipment to provide makeup to
industrial boilers.

05.qxd 8/8/06 1:56 PM Page 111
Above 600 psi, higher-quality demineralized water is usually recommended,

but often without mixed-bed polishing. A strong acid cation/strong base anion
system may be quite satisfactory. SAC/SBA systems will leak a small amount (a
few ppm or less) of sodium hydroxide, but this can often be tolerated at these
pressures. As will be described, reverse osmosis is also a potential treatment
option.
Degasifiers
Alkalinity that passes through a cation bed is converted to carbonic acid
(H2CO3), which is essentially just hydrated carbon dioxide (CO2 • xH2O). The
load on the anion exchanger can be reduced if the carbon dioxide is removed
upstream. Forced-draft or vacuum degasification will accomplish this. In the for-
mer mechanism, water is cascaded down a series of trays while air is blown
upward. This process will reduce carbon dioxide concentrations to approxi-
mately 10 ppm although it does saturate the water with oxygen. A vacuum
degasifier pulls gases from the tower as the water flows down the trays. Even
lower CO2-removal efficiencies may be obtained, but a vacuum degasifier is more
expensive. The cost of a degasifier must be weighed against the value obtained
by decreasing the load on the anion exchanger. Caustic prices generally tend to
be high, so a degasification system may be warranted for even moderately alka-
line waters.
Regeneration and Co-Current/Countercurrent

Systems
The heart of the demineralization process lies in regeneration and how effi-
ciently and well the resin can be restored to optimum capacity after it becomes
exhausted.
Figure 5-6 outlines a cation exchange vessel. Feedwater is introduced
through distributors to the top of the vessel above the resin. Due to the resin’s
varying affinity for cations, calcium is preferentially removed, followed by mag-
nesium and then sodium. Relatively distinct bands of these cations form in the
resin, with the more tightly held ions forming more narrow bands. Eventually the
resin becomes exhausted, upon which sodium begins to break through. At this
point the resin must be regenerated to prevent contamination of the boiler. (In
actuality, the resin never becomes totally exhausted because pockets of unreacted
resin still remain due to flow variations through the bed. One such pocket, known
as the heel, often develops at the bottom of the vessel along the outside walls.)
When the resin exhausts, it must be regenerated with an acid solution. In
the United States, sulfuric acid is the typical regenerant, while in Europe many

05.qxd 8/8/06 1:56 PM Page 112
Figure 5-6 systems are designed to use hydrochloric

acid. The chemical process is similar in
both. Equation 5.1 shows that the equilib-
Influent rium of ion exchange during process oper-
ation is greatly shifted to the right. This is
what makes the process so effective.
During regeneration, a concentrated (rela-
Calcium tive to the ionic concentration of the
Magnesium
process water) hydrogen ion (acid) solu-
tion is introduced to shift the equilibrium
Sodium
of Equation 5.1 to the left. For SAC resins,
4% sulfuric acid is a common regenerant
Unreacted Resin concentration, although this percentage
may be altered depending upon condi-
tions. One of the most popular variations
is step-wise regeneration. A 2% sulfuric
acid solution is initially introduced to the
Effluent
bed. The low concentration prevents pre-
Cation Exchanger Showing Layers of Ionic cipitation of calcium sulfate within the
Separation.
bed. After a suitable period of time, the
concentration is raised to 4%, and near the end of the process may be raised to
6 or perhaps even 8%.
The regenerant acid solution is produced in an acid-mixing station. Two
designs are common. In the first, a known volume of concentrated acid (66˚
Baumé 93% is typical) is pumped from the acid bulk tank to a day tank, where
it is diluted with demineralized water to perhaps 25% concentration. This solu-
tion is then pumped to a mixing tee, where it is blended with demineralized
water to produce the desired regenerant solution. In the second option, the acid
is taken directly from the bulk storage tank to the mixing tee.
In the mixing process, the acid solution is set at a constant rate, and the
water flow is adjusted to maintain the concentration at the desired value. On-line
conductivity analyzers monitor the acid concentration and automatically adjust
the water flow to maintain the desired acid percentage. The accuracy of this
process is dependent on the location of the conductivity probes in the regener-
ant feed line. The probes should be installed in accordance with good fluid
mechanical procedures and be located away from fittings or other devices that
cause flow disturbances.
The critical aspect of regeneration is the efficiency of the process, and how
much of the resin can be restored to full capacity. A significant influence on
regeneration efficiency is co-current versus countercurrent regeneration. First
generation demineralizers were designed as co-current units (Fig. 5-7), in which

05.qxd 8/8/06 1:56 PM Page 113
Figure 5-7 the regenerant is introduced in the

same path as the service water.
This presents difficulties. When
Influent regenerant is introduced co-cur-
Regenerant
rently, the most strongly held ions
must flow completely through the
bed before they are eluted. Thus,
the cations can attach and then
detach from the exchange sites as
they pass through. This requires
Resin
extra regenerant to force the
cations out of the bed, and also
leaves some cations at the bottom
of the bed where they will show up
in the effluent during the next ser-
vice run.
Effluent The first technique developed
Co-Current Regeneration.
to alleviate this problem was air
mixing of the resin. Although not
an ideal solution, it did improve regeneration efficiency by destratifying the cal-
cium and magnesium layers. The most dramatic improvement came with the
introduction of countercurrent regeneration. In this process, the regenerant is
introduced in the reverse direction of the service flow (Fig. 5-8). It is important
that the bed remain intact during regeneration. Several methods were developed
to accomplish this, with two having become most popular:
•Introducing the regenerant at the bottom of the vessel and using a block-
ing flow of water at the top to keep the bed in place
• Introducing service water at the bottom of the vessel and injecting regen-
erant from the top
Countercurrent regeneration offers a very obvious advantage; the most tight-

ly held ions are eluted without having to pass through zones containing more
weakly held ions. Countercurrent regeneration can reduce chemical require-
ments, but perhaps more importantly allows a better quality water to be pro-
duced. This will be illustrated in the section on performance calculations.
Strong Base Anion Regeneration

The performance and regeneration of a SBA exchanger is very similar to the
SAC process mentioned above. The SBA exchanger also develops bands of

Figure 5-8 exhausted resin in the order

shown in Figure 5-9. When the
resin reaches complete exhaus-
tion, silica is the first ion to break
through. SAC beds are regener-
ated with sodium hydroxide,
usually in a 4% concentration.
The dilute solution is prepared
by blending 50% caustic with
demineralized water in processes
similar to that described for the
SAC exchanger.
Step-wise regeneration of an
anion bed is not needed because
the caustic does not form a pre-
cipitate with any of the anions. As
with cation exchangers, counter-
current regeneration is more effi-
Countercurrent Regeneration cient than co-current regenera-
tion. Silica removal is greatly
enhanced by heating the regenerant water. For Type I resins, the optimum tem-
perature is 120˚F. Type II resins are not as thermally stable and should not be
heated above 105˚F.
Proper regenerations require high-grade regenerant chemicals so as not to
contaminate the resin during the process. Supplement 5-4 outlines specifications
for sulfuric acid and caustic that meet these requirements.
Weak Acid and Weak Base Exchangers

The primary purpose for WAC and WBA exchangers is to reduce the load
on downstream SAC and SBA resins. This is economical because of the superb
regeneration efficiency of these resins. Whereas SAC and SBA resins may require
regenerant with three times the hydrogen or hydroxide capacity of the ionic load-
ing, weak beds can often be restored with near stoichiometric amounts of regen-
erant. In fact, in many systems with weak acid or weak base exchangers, the
waste regenerant from the SAC or SBA exchanger is used as the regenerant solu-
tion for the weak bed. This obviously reduces chemical consumption over that
which would be used to regenerated the beds separately.
Untitled-2.indd 114 5/1/08 2:45:34 PM

05.qxd 8/8/06 1:56 PM Page 115
Demineralizer Figure 5-9

Performance
Calculations
In order to better obtain an Influent
understanding of how deminer-

alizers are sized, resin volumes
calculated, regenerant dosages
determined, and how the con-
stituents of the feed water affect
the quality of the effluent, we
Resin
will perform design calculations
for a SAC/SBA demineralizer
using resins supplied by a lead-
ing manufacturer. The deminer-
alizer will be sized to produce
100 GPM of effluent at a 16- Regenerant
hour service run time.

Effluent
Anion Exchanger Showing Layers of Ionic Separation.
Strong Acid
Cation Exchanger
Calculations
Before starting, a point should be made about the terms used for ion
exchange capacities. In the United States, the common value selected for
exchange capacities is kilograins per cubic foot, while in Europe the capacities
are expressed as equivalents per liter or milliequivalents per milliliter. Some of
the following charts indicate both. The example calculations listed below are all
based on grains and kilograins.
Assume that influent water to the demineralizer has the composition out-
lined in Table 5-2. Silica is 10 ppm. (Note: Debate seems to exist over whether
silica should or should not be considered an anion in water balance calculations.
I have left it out of the table, but dissolved silica behaves as an anion in anion
Table 5-2 exchanger beds
Cations Concentration Anions Concentration
and is considered
(ppm as CaCO3) (ppm as CaCO3)
as such in the fol-
lowing anion
Sodium 100 Bicarbonate 80
exchanger calcu-
Calcium 50 Chloride 70
lations.)
Magnesium 50 Sulfate 50
At this point,
Total 200 200
the manufactur-

05.qxd 8/8/06 1:56 PM Page 116
Figure 5-10
Strong Acid Cation Exchanger Specifications
Polymer Structure – Polystyrene cross-linked with divinylbenzene
Total Capacity Hydrogen Form – 39.2 kgr/ft3
Swelling (NaH) – 5%
Standard Operating Conditions
Operation Rate Solution Minutes Amount
Service 1–5 GPM/ft3 Influent water
Backwash 3–5 GPM/ft3 Influent water 5–20 10–25 gals/ft3
Regeneration 0.2–0.8 GPM/ft3 0.5–5% H2SO4 30 4–10 lbs/ft3
Slow rinse 0.2–0.8 GPM/ft3 Decationized 60 20 gals/ft3
Fast rinse 1–5 GPM/ft3 Decationized 60 30 gals/ft3
Backwash expansion 50–75%
*Data courtesy of The Purolite Company, Bala Cynwyd, Pennsylvania
er’s literature must be used to determine resin capacity for the system. As Figure
5-10 shows, the capacity of the resin as shipped is 39.2 kgr/ft3. However, due to
the equilibrium that develops between hydrogen ions and cations in an exchange
bed, once the resin has been used its recoverable operating capacity is well below
the preoperational value. Resin manufacturers supply graphs and charts to cal-
culate resin performance, and these are used in the following steps.
The first step is to calculate the normal operating capacity of the resin. The
factors which affect this value are regenerant dosage, the percentage of sodium in
the feed, and the ratio of calcium and magnesium in the feedwater. All can be
determined from Figure 5-11. Although regenerant concentrations are often
talked about in percent, it is the actual mass of acid introduced to the bed that
determines regeneration capacity. If more acid is supplied, more cations will be
removed. The units used for this calculation are pounds of acid per cubic foot of
resin. Although any dosage may be used, the optimum range is typically within
5 to 6 lbs/ft3. For this example, 5 lbs/ft3 is the regenerant dosage.
The header at the top of the graph in Figure 5-11 shows that the chart
applies to waters with 0 to 50% sodium. The ratio of sodium to other cations in
the water is important, because sodium elutes from the bed more easily than cal-
cium and magnesium. This effect is double-edged. As the sodium concentration
of the influent increases at the expense of hardness ions, the available capacity of
the resin increases because sodium is more easily regenerated. However, sodium
leakage during the service run will increase, because sodium’s weaker affinity for
the resin allows some of it to continually escape. Countercurrent regeneration
can mitigate this effect, as will be seen.
The graph also shows lines for magnesium and calcium. Regeneration effi-
ciency is dependent on this ratio, because magnesium elutes more easily than cal-

05.qxd 8/8/06 1:56 PM Page 117
Figure 5-11 Figure 5-12
SAC Resin, Base Operating Capacity: 0–50% SAC Resin, Base Operating Capacity: 85% Sodium
Sodium in Influent. Courtesy of the Purolite Company, in Influent. Courtesy of the Purolite Company, Bala
Bala Cynwyd, PA. Cynwyd, PA.
cium. The lines indicate resin capacities for various magnesium/calcium ratios.
The base operating capacity can now be predicted. Influent sodium makes
up 50% of the cations, the ratio of calcium to magnesium is 1 to 1, and the regen-
erant dosage is 5 lbs/ft3. The point on the y-axis that corresponds to these con-
ditions is 14.2 kgr/ft3. Had the ratio of magnesium to calcium been higher, the
base operating capacity would have been greater. Similarly, Figure 5-12 reveals
that if the sodium ratio been higher, with all other factors the same, the predict-
ed capacity would have increased. (In this case, if sodium made up 85% of the
cations the base operating capacity rises to 15.7 kgr/ft3.)
Correction now must be made for the alkalinity of the influent water. As
water passes through a cation bed, it acquires hydrogen ions as the cations are
exchanged. If the cations are associated with weak bases such as chloride and
sulfate, then the hydrogen ions behave as a strong acid and tend to regenerate the
lower portion of the bed. If, however, the cations are associated with alkalinity,
the hydrogen ions combine appreciably with this stronger base and lose most
regenerant capabilities. This has an effect on capacity calculations, and is taken

05.qxd 8/8/06 1:56 PM Page 118
Figure 5-13 into account by the graph

shown in Figure 5-13.
Alkalinity makes up 40% of the
anions. At 50% sodium, the
correction factor becomes 1.07.
Temperature also has an
effect on the process (as it does
with any reaction). Figure 5-14
illustrates the temperature cor-
rection graph for this resin. As
is evident, the reaction is slow-
er at lower temperatures. For
this example, the temperature
of the water is assumed to be
63˚F. This gives a correction fac-
tor for 50% sodium water of
about 0.97.
The overall predicted resin
capacity is equal to the calculat-
ed base operating capacity mul-
tiplied by the alkalinity correc-
tion factor and temperature
SAC Resin Alkalinity Correction Factor. Courtesy of the
Purolite Company, Bala Cynwyd, PA.
correction factor. The value for
this example is:
14.2 kgr/ft3 x 1.07 x 0.97 = 15.3 kgr/ft3 (5.7)

Figure 5-14
The sodium leakage from the
bed can also be determined.
Previously, it was shown that alkalin-
ity had an effect on the operating
capacity. Alkalinity, and its ratio to
other cations, principally chloride
and sulfate, also affects sodium leak-
age. If little alkalinity is present, the
hydrogen ions that exchange for
cations are associated with weak
bases. The solution acts as a regener-
ant, albeit a weak one, and elutes
sodium ions, which appear in the
SAC Resin Temperature Correction Factor. Courtesy of
effluent. As the alkalinity of the influ- the Purolite Company, Bala Cynwyd, PA.

05.qxd 8/8/06 1:56 PM Page 119
SAC Resin Sodium Leakage, Co-Current SAC Resin Sodium Leakage, Co-Current
Regeneration, 5 lbs/ft2. Courtesy of the Regeneration, 6 lbs/ft2. Courtesy of the
Purolite Company, Bala Cynwyd, PA. Purolite Company, Bala Cynwyd, PA.
ent increases, this effect diminishes because the hydrogen ions are more attracted
to the stronger base. Sodium leakage, therefore, becomes a function of the influ-
ent sodium and alkalinity concentrations. Figure 5-15 can be used to calculate the
projected sodium leakage from the cation bed. With sodium comprising 50% of
the cations and alkalinity 40% of the anions, and a regenerant dosage of 5 lbs/ft3,
the projected sodium leakage during the service run is approximately 2.2 ppm.
Figure 5-16 shows that with a regenerant dosage of 6 lbs/ft3, the sodium leakage
could be reduced to around 1.7 ppm.
These calculations were all based on co-current regeneration. Figures 5-17
and 5-18 show the base capacity and projected sodium leakage from the same
resin with countercurrent regeneration. Base operating capacity is increased from
14.2 kgr/ft3 to 15.2 kgr/ft3. Using the previously established correction factors for
alkalinity and temperature, the operating cation exchanger capacity becomes

05.qxd 8/8/06 1:56 PM Page 120
Figure 5-17
SAC Resin Base Operating Capacity, Countercurrent Regeneration. Courtesy

of the Purolite Company, Bala Cynwyd, PA.
15.8 kgr/ft3. This value will be used in subsequent cation exchanger vessel siz-
ing. More important is the effect that countercurrent regeneration has on sodium
leakage. The countercurrent process reduces sodium leakage during the service
run from 2.2 ppm to less than 40 ppb. This can have a significant impact on
anion exchanger operation.
The reduction in sodium leakage due to countercurrent regeneration is eas-
ily explained. The sodium band forms in the lower portion of the resin. During
co-current regeneration, sodium not removed by the regenerant remains at the
bottom of the vessel. Therefore, some sodium ions that detach from exchange
sites have only a short distance to travel before reaching the exchanger effluent.
In countercurrent systems, the reverse flow of the regenerant moves sodium ions
further up in the bed. The bottom of the bed becomes more fully regenerated.
Upon resumption of the service cycle, sodium leakage is low.

05.qxd 8/8/06 1:56 PM Page 121
Figure 5-18
SAC Resin Sodium Leakage, Countercurrent Regeneration. Courtesy of the

Figure 5-19
Strong Base Anion Exchanger Specifications
Polymer Structure – Polystyrene cross-linked with divinylbenzene
Total Capacity Chloride Form – 28.3 kgr/ft3
Swelling (ClOH) – 20%
Standard Operating Conditions – Countercurrent Regeneration
Service 1–5 GPM/ft3 Decationized water
Backwash 2–3 GPM/ft3 Decationized water 5–20 10–25 gals/ft3
Regeneration 0.25–0.5 GPM/ft3 3–6% NaOH 30–60 3–10 lbs/ft3
Slow rinse 0.25–0.5 GPM/ft3 Decationized water 30 (approx) 15–30 gals/ft3
Fast rinse 1–5 GPM/ft3 Decationized water 20 (approx) 25-45 gals/ft3
Backwash
expansion 50–75%
*Data Courtesy of The Purolite Company, Bala Cynwyd, Pennsylvania
Strong Base Anion

Exchanger Calculations
Anion resin calculations are similar in many respects to cation calculations,
because the quality of the influent and regenerant dosage directly affect the per-
formance of the exchanger. Figure 5-19 gives the design values for a strong base
resin. The following operating calculations will be based on countercurrent
regeneration to match the final calculations and graphs presented for the cation

05.qxd 8/8/06 1:56 PM Page 122
Figure 5-20
45˚ C (120˚ F)
SBA Resin Base Operating Capacity: 120˚F Regenerant. Courtesy of the Purolite Company, Bala
Cynwyd, PA.
exchanger. An important point to remember is that that the solubility of silica

increases with temperature. Many demineralizers are equipped with heaters to
warm the SBA regenerant dilution water. The optimum temperature is 120˚F. The
calculations below are all based on regeneration at 120˚F.
The ratio of anions has a direct impact on the base operating capacity of the
resin. Figure 5-20 is a graph of the SBA resin operating capacity for various influ-
ent conditions. From the water quality being used for this example, the base
operating capacity indicated by the graph is 14.4 kgr/ft3.
Correction factors must also be applied to SBA capacity calculations, but
they are governed by silica leakage. Since silica is such a harmful contaminant in
boiler systems, it is desirable to shut down the demineralizer if silica leakage even
slightly begins to exceed design values. A common value for rise in end-point sil-
ica leakage is 0.1 ppm (100 ppb). Figure 5-21 shows the correction factors for
end-point silica leakage. The correction factor at 100 ppb is 0.96. The overall
operating capacity of the resin then becomes:

05.qxd 8/8/06 1:56 PM Page 123
14.4 kgr/ft3 x 0.96 = 13.8 kgr/ft3 (5.8)
The actual silica leakage can now be determined from Figures 5-22, 5-23,
and 5-24. Silica makes up 2.5% of the total anions. At a caustic regenerant level
of 5 lbs/ft3, the base silica leakage is projected to be 5 ppb or less. (Since the
graph is not explicit at this low level, 5 ppb will be assumed.) The actual silica
leakage is then determined by the equation:
Silica leakage = Base silica leakage/(CF1 x CF2), (5.7)
where CF1 is a correction factor for sodium leakage from the cation bed and
CF2 is a temperature correction factor.
At this point it starts to become apparent how sodium leakage from a cation
exchanger can affect silica leakage. The previous cation exchanger calculations
showed that sodium leakage from a co-currently regenerated bed would be
approximately 2.2 ppm, while from a countercurrently regenerated bed would be
under 40 ppb. Figure 5-23 outlines correction factors for silica leakage based on
influent sodium. At 2.2 ppm sodium, the correction factor is 0.75, but at 40 ppb
of sodium leakage the correction factor rises to nearly 0.96.
The temperature correction factor as outlined in Figure 5-24 gives a value of
0.9 for 120˚F regenerant solution. The final projected silica leakage is:
5 ppb (0.96 x 0.9) = 6 ppb (5.9)
Figure 5-21 Weak Acid and Weak

Base Performance
Weak acid and weak base
resins perform differently than
strong acid and strong base resins,
and thus different calculations
must be used. The affinity for ions
of a WAC resin is in the order H+
> Ca++ > Mg++ > Na+. WAC
resins will remove cations associ-
ated with bicarbonate alkalinity,
but will not remove those associ-
ated with weak bases—most
SAC Resin Correction Factor for Silica End Point notably chloride and sulfate. For
Leakage. Courtesy of the Purolite Company, Bala Cynwyd, PA.
the water quality used in these

05.qxd 8/8/06 1:56 PM Page 124
Figure 5-22 examples, a WAC resin would

remove all of the calcium plus
30 ppm of the magnesium.
These are the cations associat-
ed with the 80 ppm of alkalin-
ity in the water. The remaining
magnesium and the sodium
would pass through.
Figures 5-25, 5-26, 5-27,
and 5-28 show the design data
and operating capacity graphs
of a WAC resin. The resin
capacity is determined by cal-
culating the operating capacity
and multiplying it by the tem-
perature correction factor and
flow rate correction factor. The
operating capacity curve is
SAC Resin Silica Leakage, Countercurrent
Regeneration. Courtesy of the Purolite Company, Bala
unlike that for SAC resins, and
Cynwyd, PA. reflects a WAC resin’s much
stronger preference for divalent cations than monovalent cations. In our exam-
ple, the ratio of hardness to alkalinity is 100 to 80, which gives an alpha factor
of 1.25. This corresponds to a capacity of approximately 72 kgr/ft3. If the com-
monly recommended value of 2 GPM/ft3 is selected for volumetric flow rate at a
temperature of 68˚F, the actual operating capacity remains 72 kgr/ft3. This is
SAC Resin Sodium Leakage Correction Factor. SAC Resin Regenerant Temperature Correction
Courtesy of the Purolite Company, Bala Cynwyd, PA. Factor. Courtesy of the Purolite Company, Bala
Cynwyd, PA.
05.qxd 8/8/06 1:56 PM Page 125
Figure 5-25
Weak Acid Cation Exchanger Specifications
Polymer Structure - Acrylic divinylbenzene
Total Capacity Hydrogen Form - 91.6 kgr/ft3
Swelling (H Na) - 90%
(H Ca) - 20%
Standard Operating Conditions
Service 1–5 GPM/ft3 Influent water
Backwash 3–5 GPM/ft3 Influent water 5–20 15–60 gals/ft3
Regeneration 1–2.5 GPM/ft3 0.5–1% H2SO4 30–45 100–120% of theory
Slow rinse 1–2.5 GPM/ft 3 Decationized 15 15–30 gals/ft3
Fast rinse 1.25–5 GPM/ft Decationized
3 22–45 gals/ft 3
Backwash Expansion 50–75%

Minimum Bed Depth - 30 inches
*Data courtesy of The Purolite Company, Bala Cynwyd, Pennsylvania.
WAC Resin Base Operating Capacity. Courtesy of WAC Resin Temperature Correction Factor.
the Purolite Company, Bala Cynwyd, PA. Courtesy of the Purolite Company, Bala Cynwyd, PA.
much closer to the design capacity than is the case with SAC resins, and reflects
the efficiency with which WAC resins may be regenerated.
Weak base anion resins remove weak bases as their acid conjugates (sulfu-
ric acid, hydrochloric acid). The resins do not remove alkalinity. For this exam-
ple then, a WBA exchanger would remove the 120 ppm of chloride and sulfate.

05.qxd 8/8/06 1:56 PM Page 126
Figure 5-28 Figures 5-29 and 5-30 show the

design data and operating capacity
for a WBA resin. The principal fac-
tor that affects exchange capacity
is the regenerant dosage. Because
WBA resins can be regenerated
very efficiently, sodium carbonate
or even ammonium hydroxide
may be used in lieu of caustic. The
base operating capacity of this
resin with a 5 lbs/ft3 caustic regen-
erant is 26.6 kgr/ft3 (1.22 eq/l).
Assuming, as was done with the
WAC resin, a flow rate correction
factor of 1.0, the overall operating
capacity remains at 26.6 kgr/ft3.
WAC Resin Flow Rate Correction Factor. Courtesy of the Again, this is very close to the
design capacity.
System Design Calculations

Once the resin volume is known, vessel sizes can be calculated. These sizes
are not only dependent upon the service rate required but also include such fac-
tors as backwashing, regeneration, and good hydraulic design.
Figure 5-29
Resin VolumeWeak
and Base Anion Exchanger
Vessel Diameter Specifications
Calculations
Polymer Structure - Polystyrene cross-linked with divinylbenzene
For Capacity
Total a demineralizer to3 perform properly, certain hydraulic requirements
- 28.3 kgr/ft
must be met.Operating
Standard The ion Conditions
exchange process is not only governed by chemistry but
alsoOperation
by kinetics and maximum utilization
Rate Solution of the ion exchange
Minutessites. Figures
Amount5-10,
5-19, 5-25,
Service and 5-29 illustrated
8–40 BV/h the manufacturer’
Decationized water s recommended flow rates for
the Backwash
exchange resins. 5–7Table
m/h 5-3 is a supplement
Decationized waterto this data,
5–20 and provides
1.5–4 BVgener-
al hydraulic guidelines.
(39–77˚F)These guidelines are integral to the following calcula-
tions.
Regeneration 4 BV/h 2–4% NaOH 30 2–6 lbs/ft3
From the
Slow rinse calculations
4 BV/h performed above, the
Decationized water capacity
20 of the SAC
1–5 resin
BV was
determined
Fast rinse to be 1615.8
BV/hkgr/ft3. Decationized
The influentwater
water quality
15 contains4200 BV ppm of
cations
BV =as CaCO3,
bed volumesproduction is 100 GPM, and the service run time is 16 hours.
OneBackwash
grain per gallon is
Expansion equivalent to 17.1 ppm of ions as CaCO3. Thus, the
35–50%
quantity of ions
*Data courtesy to Purolite
of The be removed
Company,is:
Bala Cynwyd, Pennsylvania.

05.qxd 8/8/06 1:56 PM Page 127
Figure 5-30
lbs./cu.ft.
g/l
WBA Resin Base Operating Capacity. Courtesy of the Purolite Company, Bala Cynwyd, PA.
200 ppm ÷ 17.1 ppm per gr/gal x 100 GPM x 960 minutes = 1,122,807
grains (5.10)
3
The volume of SAC resin (capacity 15.8 kgr/ft ) needed to treat this quanti-
ty of water is 71 ft3. Most resin suppliers recommend that a safety factor of 0.85
to 0.9 be incorporated into resin volume design. Using a 0.9 factor, the final
design volume becomes:
71 ft3 ÷ 0.9 = 79 ft3 (5.11)
A 2 GPM/ft3 volumetric flow rate as suggested in Table 5-3 gives a bed depth
of 48 inches with a tank diameter of 5 feet. This is within good design proce-
dures.
Calculations for the SBA resin are similar. The capacity for the resin was
determined to be 13.8 kgr/ft3. (Anion resins usually have less capacity than

05.qxd 8/8/06 1:56 PM Page 128
Table 5-3
Guidelines for Good Hydraulic

Performance of Ion Exchangers
Parameter Range of Value
Volumetric flow rate (gpm/ft3) 0.25 to 5
Cross-sectional flow rate (gpm/ft3) 4 to 12
Pressure drop per foot of resin (psi/ft.)* 0.1 to 1.5
Minimum bed depth (inches)** 30
* Pressure drop is determined by cross-sectional flow rate and water

temperature. Values shown are for 4–10 gpm per sq. ft. flow rates and
a temperature range of 40˚F to 100˚F.
** The book Practical Principles of Ion Exchange Water Treatment (see
Bibliography) recommends 9 feet as a maximum practical bed depth.
cation resins). Including silica, the water contains 210 ppm of anions as CaCO3.
The quantity of anion resin required is 94 ft3 including the 0.9 design factor. At
this point, a choice is possible. A 48-inch bed depth could again be used, which
would increase the diameter of the vessel to 5 feet. Or, the diameter of vessel
could be set at 5 feet to correspond with that of the cation vessel, and the bed
depth could be increased to 58 inches. Both would be satisfactory.
Similar calculations can be performed for WAC and WBA resins. Once the
resin capacity is determined, the volume can be calculated based upon the quan-
tity of ions to be removed.
Backwash Requirements
When resin exhausts it is backwashed before regeneration to remove resin
fines and particles that were filtered by the bed. Backwash water is introduced at
the bottom of the bed to lift and mix the resin. For backwashing to be effective,
the resin must be thoroughly agitated and expanded. This requires extra vessel
height, known as freeboard. Figure 5-10 shows that the manufacturer recom-
mends a bed expansion of 50 to 75% for the SAC resin. The bed depth has
already been calculated at 48 inches, so a 75% volume addition increases the
straight side-height of the vessel to 84 inches.
Figure 5-19 also recommends 50 to 75% bed expansion for the SBA resin.
Since the bed height in this vessel was calculated to be 58 inches, a 75% bed
expansion increases the straight-side height to 101 inches.
The efficiency of bed expansion is very dependent on the temperature of the
water. Colder water, being more dense, expands the bed more easily. Water flow

05.qxd 8/8/06 1:56 PM Page 129
rates must be carefully regulated to prevent resin from being discharged through
the service water influent distributor. Resin manufacturers will supply charts and
equations for calculation of the backwash flow rate based on temperature.
Backwash times typically range from 5 to 20 minutes, with 15 minutes being
an average. This allows 1 to 3 bed volumes to pass through the vessel.
Regeneration Requirements
Many aspects of regeneration have already been discussed. Regeneration is
carried out at a much slower flow rate (0.2 to 0.8 GPM/ft3) than either the ser-
vice or backwash processes in order to give the regenerant time to be effective.
Regeneration of cation exchangers with sulfuric acid is often performed in a step-
wise process to prevent formation of calcium sulfate deposits within the resin.
The regenerant is initiated at a 2% concentration and is raised to 4% and perhaps
even 6% later on in the process. SBA regenerant strength is usually 4% through-
out the process. Regeneration proceeds more quickly in a cation bed than an
anion bed. Cation regeneration may be complete in only half an hour, whereas
anion regeneration may take an hour or more.
Regenerant concentrations for weak acid and weak base exchangers are sig-
nificantly lower due to each resin’s high efficiency of regeneration. In systems
where WAC and WBA exchangers are used in combination with SAC and SBA
exchangers, the waste regenerant from the strong exchangers may be used to
regenerate the weak resin exchangers. Recommended regenerant dosages for
SAC and SBA resins are 4 to 10 lbs/ft3 of acid and caustic, respectively, but for
weak acid and weak base resins, they may be as low as 2 lbs/ft3.
All resins undergo swelling as they convert from one ionic form to another.
SAC and SBA resins both swell during regeneration, but the change is usually
minor (5 to 8% for SAC resins and 10 to 20% for SBA resins). WAC and WBA
resins swell during the service run, and these changes are much more dramatic,
with swelling up to 100% being possible. In fact, vessel design for weak acid and
weak base resins may be dictated more by swelling than for backwash volume.
Rinsing
Upon completion of regeneration, the exchange resins are rinsed to remove
the regenerant. The first step is a slow rinse at the same flow rate as the regener-
ant. This allows the remaining regenerant to remove ions as it is being rinsed
from the bed. The slow rinse may take anywhere from 30 to 60 minutes. At the
end of the slow rinse, the water flow rate is increased, often to the service rate,
for the fast rinse. The fast rinse duration may also be 30 to 60 minutes. The fast
rinse puts several bed volumes of water through the resin.

05.qxd 8/8/06 1:56 PM Page 130
Where possible, previously demineralized water should be used for rinse

and regenerant makeup. This prevents contaminants from being reintroduced to
the exchange beds during the regeneration process.
Mixed-Bed Polishing
Mixed-bed (MB) exchangers contain both SAC and SBA resins intimately
intermixed. The primary purpose of a MB exchanger is to polish the already puri-
fied water. MB quality water is needed for high-pressure boiler makeup and at
some manufacturing facilities, particularly in the electronics and semiconductor
industries.
Because a MB exchanger has such a light ionic load, flow rates can be
increased. A typical cross-sectional flow rate is 20 GPM/ft2. The mixed bed is
sized to provide very long term operation with infrequent regenerations. This can
be done without excessive quantities of resin because the ion loading is so low.
Consider a SAC/SBA demineralizer with an effluent that contains 1 ppm of
cations as CaCO3. If the cation exchange resin from the earlier example is used
in the MB exchanger, 50 cubic feet could theoretically process 13,500,000 gal-
lons of water. At 100 GPM, the MB cation resin would not exhaust for three
Figure 5-31
CONDUCTIVITY (micromho)
MAXIMUM MINERAL ACIDITY (ppm)

ACTUAL MINERAL ACIDITY (ppm)
SODIUM (ppm)
>
HARDNESS (ppm)
0 <
SERVICE CYCLE >
Effluent Characteristics—SAC Exchanger

05.qxd 8/8/06 1:56 PM Page 131
months. (In actuality, the higher flow rate can reduce the operating capacity
somewhat. Even so, very long run times are still normal.)
Although regeneration of mixed-beds is performed infrequently, it is a very
important process and must be handled with care, since two resins are involved.
Regeneration is usually performed in the exchange vessel. Anion resin is lighter
than cation resin, and when the MB resin is backwashed, the process causes the
two resins to settle into distinct layers. By design, a collector system is installed
at the cation/anion resin interface. During regeneration, acid is introduced below
the cation resin and flows upwards. Caustic is introduced above the anion resin
and flows downwards. The waste regenerant from each is collected at the inter-
face. It is very important that the resins settle properly and that the division
between them occurs at the central collector. If cation resin remains in the anion
zone, it will absorb sodium from the caustic regenerant. Conversely, anion resin
will pick up sulfate from the acid regenerant. A technique for providing good
separation is to design the system for and include a small amount (10% or so) of
inert resin that has a density between that of the cation and anion resin. The inert
resin will settle at the interface and provide separation between the active resins.
The mixed-bed regeneration process is similar to that for cation and anion
resins with regard to acid and caustic concentrations. Once the resin has been
rinsed, it is remixed with air. The air remix process is very important to re-estab-
lish the heterogeneity of the cation and anion resins. Typical guidelines suggest
an air flow rate of 7 to 10 SCFM for 15 minutes.
Monitoring Performance of Ion Exchanger Vessels

Figure 5-31 illustrates the typical effluent water quality from a cation bed
during a normal service cycle. At the very beginning of the service cycle, the
effluent is dumped to waste because it contains trace amounts of acid. The acid
residual usually disappears quickly. The effluent water is routed into the system
just as this process reaches completion. During the service cycle, the cation efflu-
ent contains free mineral acids (FMA), e.g., H2SO4, HCl, and also H2CO3 and a
small amount of sodium. Once the bed has reached exhaustion, sodium begins
to break through. The FMA and conductivity both begin to decline because the
excess sodium combines with anions to form neutral salts. Of these changes in
effluent quality, the increase in sodium is most easily detected. On-line sodium
monitoring can be very effective in detecting resin exhaustion.
The quality of water during an anion service cycle is outlined in Figure 5-
32. A slight conductance is always present in the effluent because the sodium
ions that leak from the cation exchanger form sodium hydroxide in the anion
bed. The conductivity usually remains constant throughout the service run.
Likewise, the effluent contains a constant level of silica due to leakage of this

05.qxd 8/8/06 1:56 PM Page 132
Figure 5-32 weakly held constituent.

When the anion exchanger
exhausts, silica levels almost
immediately begin to rise.
On-line silica monitoring
CONDUCTIVITY (micromho) can be a very effective tool to
detect resin exhaustion.
Another phenomenon also
pH occurs that can be used as a
tool. When the bed just
begins to exhaust, the first
bit of silica that comes off
combines with sodium
hydroxide to form sodium
SILICA (ppm) silicate. Sodium silicate has a
>
lower conductivity than

0 < SERVICE CYCLE > NaOH. With proper observa-
Effluent Characteristics—SBA Exchanger
tion, the small dip in con-
ductivity can be used to indi-
cate bed exhaustion. Because this measurement requires a trained eye, on-line
silica monitoring is a better choice for most facilities.
For systems with a mixed-bed exchanger, the mixed-bed acts as a buffer, at
least for a while in the event of a cation or anion exchanger overrun. However, if
the mixed-bed fails or becomes exhausted, nothing prevents the contaminants
from traveling to the steam generating system. On-line sodium and silica moni-
toring of the mixed-bed effluent is highly recommended.
Many plant operators like to run the demineralizers to a certain setpoint and
then shut the system down for regeneration. The throughput volume is chosen
to be less than the known level of bed exhaustion. This method has merits
because it can help prevent demineralizer overrun and system upsets. The pri-
mary drawback is that more regenerant chemical is needed, and more frequent
regenerations are harder on the resin. Even when control of the demineralizer is
based on throughput, on-line monitoring of the effluent it still recommended.
Demineralizer Component Fundamentals

The various components that go into a demineralizer, i.e., vessels, distribu-
tors, valves, and so forth (Fig. 5-33), must all be designed using good engineer-
ing practice for the demineralizer to operate properly. The following subsections
provide some general ideas regarding these items.

05.qxd 8/8/06 1:56 PM Page 133
Figure 5-33
Skid-Mounted Demineralizer System. Photo courtesy of U.S. Filter.
Distributors
The flow rate of water or regenerant across the resin should be uniform to
prevent channeling or areas of high and low flow. Distributor design is most
important. Distributors are usually supplied in a lateral type arrangement so that
the spray uniformly covers the entire surface area of the resin. The spray nozzles
and piping should be designed to provide the same quantity of water from each
opening.
Vessels
Ion exchanger vessels are rubber lined to prevent corrosion. The rubber lining
must be installed properly and spark tested to eliminate any holidays in the liner. If
liquid gets behind the liner, where it can attack a localized area of the vessel shell,
shell failure may result. Vessels must also be fabricated in accordance with pres-
sure-vessel design codes. This becomes particularly important for on-line conden-
sate polishers, which may operate at several hundred pounds of pressure.
Valves
The valving arrangement on a demineralizer can be fairly complicated.
Valves automatically open and close during changes between service, regenera-

05.qxd 8/8/06 1:56 PM Page 134
tion, and rinse. Positive shutoff valves such as block valves are recommended so
that process fluids do not infiltrate unwanted areas.
Materials
Materials for piping, valves, and tanks must be corrosion resistant. PVC pip-
ing has been used on some demineralizers, although it can break if it is acciden-
tally hit with a piece of equipment. Stainless steel (316) is much more durable
and can handle most chemical solutions. (Stainless steel is not recommended if
hydrochloric acid is used as the cation regenerant.) Alloy 20 provides superior
resistance to dilute sulfuric acid solutions, although it is very expensive. Both
lined and unlined tanks have been used for concentrated acid and caustic stor-
age. A liner will help prevent corrosion.
Packed-Bed Demineralizers
The latest generation of demineralizers include systems based on packed-
bed technology. In these demineralizers, the vessel is almost completely filled
with resin, with perhaps only 10% freeboard. Two processes are available, both
based on countercurrent operation. In one, the process water flows from top to
bottom of the vessel and the regenerant from bottom to top. Unlike convention-
Figure 5-34
Short-Bed Ion Exchange Unit. Photo courtesy of Eco-Tec, Inc., Pickering, Ontario, Canada.

05.qxd 8/8/06 1:56 PM Page 135
al countercurrent regeneration, however, the regenerant is introduced at enough

pressure to lift the entire bed as a plug to the top of the vessel. The other type of
packed-bed system essentially reverses these functions. The process water raises
the bed, and the resin is regenerated from top to bottom. These processes require
different distributor and collection systems than conventional units, whose dis-
tributor laterals would be subjected to mechanical stress by the moving plug of
resin. Distributor and collector plates with screened openings are located at the
bottom and top of the packed-bed vessel, and serve as the restraining barrier for
the resin.
Manufacturers of these systems claim several advantages versus regular
countercurrent demineralizers. One advantage is that a smaller vessel will hold
the same amount of resin as a standard ion exchange unit. Better regeneration
efficiency is another reported advantage. This is due to the more intimate contact
of regenerant with the tightly compacted bed than occurs in a standard counter-
current unit where the bed may be somewhat fluidized.
One area where packed-bed technology has been applied is retrofitting of
existing co-current demineralizers. Water quality can be significantly improved
by this change.
Many of the major water treatment firms such as U.S. Filter, Glegg, Indeck,
and Ecodyne manufacture these systems or have licensing agreements with resin
manufacturers including DOW, Rohm & Haas, Purolite, and Bayer Corporation.
An even more novel approach to packed-bed technology is the short-bed
system (Fig. 5-34). In these systems, the resin is packed into minimum-depth
vessels and no freeboard exists. The main features of short-bed systems include:
• The bed height may be as short as 6 inches
• Very fine resins are used for improved reaction kinetics
• Surface area flow rates may be as high as 30 GPM/ft2 as compared to stan-
dard design values of 8 to 12 GPM/ft2
• Run times range from 5 to 30 minutes with a 20 to 40 second regenera-
tion and a 1 minute rinse
• After regeneration the effluent is recirculated to the influent until water
quality reaches desired values. This may take a couple of minutes.
The obvious advantage of this system is its small size. For facilities with lim-
ited space, or where a demineralizer must be retrofitted into a confined area,
these systems can be very advantageous. Water quality and regeneration efficien-
cy are also reported to be very good. One drawback to these systems is that with
zero freeboard, the resins cannot be backwashed. The influent must be well fil-
tered to remove suspended solids.

05.qxd 8/8/06 1:56 PM Page 136
Other Makeup Technologies

Ion exchange does not have the same dominance it once had on the high-puri-
ty treatment market. This is principally due to the development of reverse osmosis
technology. The hybrid ion exchange/membrane methods of electrodialysis and
electrodeionization are also making inroads into the makeup water market.
Reverse Osmosis
Reverse osmosis technology has developed substantially. Systems have
become more reliable, and membrane performance has improved to a level where
RO systems can routinely remove 99% or greater of the influent dissolved solids.
Reverse osmosis has become particu-
Figure 5-35
FLOW
larly popular as a retrofit ahead of an
existing demineralizer, or as part of a
combined system, i.e., RO plus mixed
bed, for new installations. The economics
are particularly favorable when reverse
osmosis is used to pretreat high TDS
waters. The RO greatly reduces ion load-
ing on the demineralizer, which lowers
Conventional Filtration. the associated regeneration frequency and
chemical costs. Although reverse osmosis is considered to be a filtration tech-
nology, the process is more complex than conventional filtration. In an ordinary
depth or weave filter, water flows perpendicularly to the filter (Fig. 5-35).
Particles are removed throughout the depth of the filter. When the differential
pressure between the inlet and outlet becomes too large, the filter is replaced. In
an RO system, water flows parallel to the membrane surface. Applied pressure at
the influent forces a portion of the water through the membrane as it passes from
one end to the other. Solids are swept along with the water that does not pass
through the membrane. This water, which becomes increasingly concentrated,
flows to the end of the pressure vessel and is discharged (Fig. 5-36). This is
known as crossflow filtration.
Figure 5-36
FEED REJECT
PERMEATE
Cross-Flow Filtration.

05.qxd 8/8/06 1:56 PM Page 137
The feature that makes RO unique as compared to other filtration methods is

its ability to removed dissolved solids down to the smallest ions. RO membranes
have pore structures that range from 1Å to 10Å in diameter. Because the pore
diameters are so small, a molecular layer of water builds up on the membrane sur-
face and inhibits ions from passing through the pores. The ions that most readily
penetrate an RO membrane are monovalent, such as sodium and chloride.
RO Membrane Design
Two types of RO membrane design are most common today: hollow-fiber and
spiral-wound. In hollow-fiber configurations, a membrane element consists of
many fibers bundled together. Each fiber individually treats a portion of the feed-
water. The purified water (permeate) and concentrated water (reject) are collected
at the end of the vessel. Hollow-fiber systems have not proven popular for brack-
ish water systems and have not been extensively used for most of the industries
addressed by this book. By far, spiral-wound membrane configurations are more
common. The remainder of this discussion is based on spiral-wound systems.
Spiral-wound membranes are manufactured in flat sheets, which are wound
around a central core to produce a membrane element (Fig. 5-37). Several elements
are placed in series and are sealed in a pressure vessel (Fig. 5-38). Feedwater is
applied to the end of each element, and it flows along the spacers to the opposite
end. Permeate passes to the central core of the element, while the reject is collect-
ed and discharged at the element end cap.
RO Membrane Material
Two types of material are common for RO membrane configuration: cellu-
lose acetate/triacetate (CA) and polyamide. The latter material is typically layered
with other membranes for support. These layered polyamide membranes are
known as thin-film composites (TFC).
CA membranes were the first to be developed. The material provides advan-
tages and disadvantages, as compared to its polyamide TFC counterpart. The
important advantages of CA material include:
• The membrane is chlorine tolerant up to about 1.0 ppm and works well
in systems where microbiological fouling must be controlled.
Figure 5-38
FEEDWATER PRESSURE VESSEL PERMEATE
INLET REJECT
MEMBRANE MEMBRANE MEMBRANE >
> ELEMENT ELEMENT ELEMENT >
Membrane Elements in a Pressure Vessel.

05.qxd 8/8/06 1:56 PM Page 138
Figure 5-37
Spiral-Wound Membrane Structure. Illustration reproduced

with permission from Osmonics, Inc., Minnetonka, Mn.
• The membrane is nonionic and has less tendency to attract ionically

charged substances such as coagulants and water treatment polymers.
Disadvantages include:
• Feed pressure is higher than for a TFC membrane. Typical feed pressures
of CA systems for the applications described in this book range from 200
to 400 psig.

05.qxd 8/8/06 1:56 PM Page 139
• In the absence of a microbiocide, CA membranes will be attacked and

degraded by microbes.
• CA membranes are only stable within a narrow pH range of 4 to 6.
Excursions outside of this range will cause membrane degradation. Since
many waters are neutral to alkaline, this requires acid feed ahead of the
system.
TFC membranes also have advantages and disadvantages. The advantages

include:
• Operating pressures are lower. Membranes have been developed that

operate very well with feed pressures as low as 150 psig. Some reports
indicate that feed pressures may go as low as 100 psig.
• TFC membranes are not attacked by microbes. However, they can be
fouled with microbiological deposits.
• TFC membranes are stable over a pH range of 2 to 12.
Disadvantages include:
• The membranes are attacked by oxidants. This may require an antioxidant

feed ahead of the system.
• The membrane surface exhibits a negative charge, which will attract coag-
ulants and cationic polymers.
RO Pretreatment
Reverse osmosis units are very fine filtration systems, which concentrate dis-
solved solids. This makes them subject to scaling, fouling, or chemical attack.
The effects of oxidants and pH have been mentioned above. Of equal concern is
fouling. Suspended solids are not compatible with RO membranes, although
reverse osmosis can be an excellent method for removing contaminants such as
colloidal silica, which passes through a demineralizer. The makeup water pre-
treatment methods mentioned at the beginning of this chapter will often produce
water suitable enough for feed to a RO. A rule of thumb says that waters with a
turbidity of less than 1 NTU will be suitable. This is only a general guideline.
Another measurement for determining the fouling potential of water is known as
the Silt Density Index (SDI). SDI is a measure of the effect suspended solids have
on water flowing through a filter. A SDI value of less than 5 usually indicates that
fouling will not be a problem. Supplement 5-5 outlines the SDI procedure and
calculations.

05.qxd 8/8/06 1:56 PM Page 140
Even if suspended solids are low in concentration, other constituents can

cause problems in RO systems. When water flows through a reverse osmosis
pressure vessel, the reject continually accumulates dissolved solids. Calcium car-
bonate, sulfate, or other compounds can build up to a point where precipitation
begins to occur. Besides calcium carbonate and sulfate other potential scales
include silica and alkaline metal silicates, strontium sulfate, barium sulfate, and
calcium fluoride. Calcium, alkalinity, and some silica can be reduced by
upstream softening and/or acid feed. Barium and strontium sulfate are more dif-
ficult to deal with and may require reduced output from the RO. The reputable
membrane manufacturers have developed programs that will calculate the solu-
bility limits for these salts in a particular application. The program will warn the
user if any solubility limit is exceeded.
Where scaling is a problem, antiscalants can help. Common antiscalants
include polyacrylates and phosponates. The correct antiscalant or blend can con-
trol calcium sulfate at 230% above the saturation limit, strontium sulfate 800%
above the saturation limit, and barium sulfate 6000% above the saturation limit.
Water treatment chemicals can also affect membrane performance.
Coagulating agents of the cationic variety are particularly troublesome to RO
membranes, especially to TFC membranes whose surface is negatively charged.
This has been an overlooked item. If these agents are present, methods to remove
them must be considered.
The importance of obtaining accurate influent water quality data cannot be
overemphasized. Ideally, historical data will be available. If not, then periodic
analyses should be collected as far in advance as possible of the decision to pur-
chase a reverse osmosis unit. (I made the mistake of using a “snapshot” SDI
analysis to affirm that a reverse osmosis unit could be placed ahead of a dem-
ineralizer. The analysis, conducted by the eventual RO supplier, showed an SDI
of less than one. After the RO was purchased, plant personnel rechecked the SDI
and obtained readings in the 30s. The company had to go to some of trouble to
lower the SDI to an acceptable value.)
RO Design
A reverse osmosis unit has been called nothing more than a high-pressure
pump, some pressure vessels, and pipe. In truth, the operation is more compli-
cated than this. Spiral-wound membrane elements can come in different sizes.
The most popular size is 8 inches in diameter by 40 inches in length. These are
loaded in series into a pressure vessel (refer again to Fig. 5-38), with four, five,
or six elements per vessel being most common. Each element can pass a certain
amount of water, and this volume is usually measured in gallons per day (GPD).
Common values for 8 in. x 40 in. elements range from 4000 to over 13,000 GPD.

05.qxd 8/8/06 1:56 PM Page 141
The rate at which water passes through the membrane is known as the flux and
is measured in gallons per square foot per day (GFD). Common flux rates for
water sources have been reported as follows:
• Surface water - 8 to 14 GFD

• Well water - 14 to 18 GFD
For normal surface and groundwaters, each pressure vessel will produce
about 50% purified water (permeate) and 50% concentrated water (reject or con-
centrate). This does not seem very efficient. However, the concentrate is often
still pure enough to be treated again at another 50/50 split to produce 75%
capacity. Sometimes even the second concentrate can be treated to give an over-
all RO output of 87.5%.
With this background information, we will look at a simple reverse osmosis
system designed to produce 300 GPM water with 75% recovery of the influent.
Conditions are as follows:
• Surface water
• SDI < 5
• Five membranes per pressure vessel
• Membrane capacity - 10,000 GPD
The system will be designed in a two-stage process, whereby the concentrate

from the first stage is sent to a second stage for further treatment. Since 300 GPM
of purified water is to be produced, and the overall efficiency is 75%, the influ-
ent quantity must be 400 GPM. This is 576,000 GPD. At a capacity of 10,000
GPD per element, 58 elements are required. Rounding off gives 12 pressure ves-
sels with 5 elements apiece. At a 50/50 split between permeate and reject, the
concentrate flow to the second stage is 288,000 GPD. Treatment of this flow
requires 29 pressure vessels. Rounding off again gives 6 pressure vessels with 5
elements apiece.
This is, of course, a very simplified examination of RO sizing, but it does
provide an example of the flexibility that is available. The number of elements
can be varied per pressure vessel, and the quantity of pressure vessels themselves
can be adjusted to obtain the desired flow rate.
RO elements have been improved such that 99% or greater salt rejection is
available. Over time, however, membranes will degrade until after two or three
years of operation, salt rejection may be only 95 to 97%. Even so, the RO still
performs a valuable service. Consider the water quality that was used for the
demineralizer calculations outlined earlier in this chapter. Total dissolved solids
were 410 ppm as CaCO3. At 97% salt rejection, the RO effluent would contain

05.qxd 8/8/06 1:56 PM Page 142
12 ppm TDS. If this RO were retrofitted ahead of the hypothetical SAC/SBA dem-
ineralizer calculated earlier, the resin run lengths could be greatly extended (a
general rule of thumb is 20 times). Retrofitting is in fact what many plant man-
agers and engineers are now doing. The driving factor is economics. It is not
impossible for regenerant chemical costs to exceed $100,000 per year for even a
moderately sized (200 GPM) demineralizer. Operating costs for the RO pump
(power), perhaps a small amount of antiscalant feed, periodic chemical cleaning
of the RO membranes, and (reduced) regenerant chemical feed to the deminer-
alizer may be less than a fourth of the normal chemical regenerant cost. Payback
time for the RO might be as short as two years. One factor not included is the
cost for membrane replacement. Membranes typically last from three to seven
years, although longer lives have been recorded. Membrane replacement costs
may be a third of the original price of the RO.
A technique that is gaining acceptance for new high-purity water installa-
tions is reverse osmosis followed by mixed-bed polishing. This arrangement per-
forms very similarly to an SAC/SBA/MB demineralizer.
Various articles and technical papers have been presented over the last few
years that attempt to give a rule of thumb guideline for the water quality at which
RO is favored over SAC/SBA, when installed ahead of a mixed-bed. Although the
topic is debatable, it is evident that the economics have improved for RO.
Around 1990, it was suggested that an influent TDS of approximately 350 ppm
or higher made RO economical. Reports later in the decade indicated that this
level had dropped to as low as 150 ppm. The economics must be examined on
a case-by-case basis.
A technique for producing relatively high-purity water without the use of
any demineralization is double-pass RO. (Pass and stage should not be con-
fused.) In this system, the permeate from the first stage is sent to another set of
membranes for further purification. This process is capable of reducing solids
concentrations to less than 1 ppm. The concentrate from the second pass is recir-
culated to the influent, so the overall efficiency of the reverse osmosis system is
not diminished by adding this second pass. Two-pass RO permeate is potential-
ly suitable for feed to low- and medium-pressure boilers.
RO Components
Various components besides the membranes make up a reverse osmosis sys-
tem. They include the pump, pressure vessels, piping, and instruments (Fig. 5-
39). The following sections outline some of the most important aspects of this
equipment.
Pumps. Centrifugal, multistage pumps have proven best for RO applica-
tions. They can easily generate the pressures needed, although pressures have

05.qxd 8/8/06 1:56 PM Page 143
Figure 5-39
Skid-Mounted Reverse Osmosis Unit. Photo courtesy of U.S. Filter.
become lower as membrane quality has improved. Stainless steel is recommend-

ed as the pump material to give the pump long life and durability. RO feed
pumps tend to be noisy, but they can be installed submersibly to minimize noise.
The submersible housing resembles a pressure vessel and is installed in a similar
fashion on the RO skid.
Pressure vessels. Pressure vessels are typically fabricated from stainless
steel or fiberglass. Both have proven effective. At one time, stainless steel hous-
ings were more flexible from a design standpoint because feedwater lines could
be tapped into the side of the vessel. This arrangement allows a technician to
remove membranes without having to disconnect any plumbing. Fiberglass did
not have the structural strength to support side entry lines. This has changed,
and now fiberglass vessels are being constructed with side-entry ports.
Piping. For high-pressure RO piping, stainless steel is the popular choice.
Often, however, one will see high-pressure flexible hose connecting one pressure
vessel to another.
RO Flow Control and Monitoring

The three most important measurements for monitoring RO operation are
flow, pressure, and conductivity. Other important measurements include tem-
perature, pH (especially for CA membranes), and oxidation reduction potential
(ORP, especially for TFC membranes). Flow is extremely important because it

05.qxd 8/8/06 1:56 PM Page 144
will change if membranes foul, tear, or degrade due to chemical or microbiolog-

ical attack. Monitoring flow rate in a reverse osmosis unit is not a simple task, as
temperature directly affects the amount of water that will pass through the mem-
branes. When water cools it becomes more dense, and its passage is restricted
through the membranes. At the same pump pressure, flow may decrease by
almost 50% when the influent temperature drops from 77˚F to 50˚F. Conversely,
the flow may increase by 10% with a 10 degree rise in temperature above 77˚F.
Membrane manufacturers will supply data that shows how their membranes per-
form with temperature. The optimum rating is assumed to be at 77˚F, and con-
version factors are provided for all other reasonable temperatures.
The conversion factor must be included in any flow calculations at temper-
atures other than 77˚F. This is called “normalizing” the flow rate. Without this
correction, it would be difficult to accurately determine whether a change in flow
rate was due to temperature effects or a change in membrane performance.
The change in flow due to temperature brings up an interesting problem
when designing a reverse osmosis system. How should the feed pump and mem-
brane quantity be determined? Several methods are possible. The system could
be sized to produce the required capacity at the coldest inlet water temperature.
The system might then be operated less frequently in the spring and summer due
to the increased capacity. Or, if an absolute flow rate was not required, the sys-
tem could be sized for an average water temperature, with the knowledge that
output would be greater or smaller depending on the time of year. A third option
is to use a variable speed feed pump with controller. Capacity could be regulat-
ed as the operator saw fit. A fourth possibility is to install a heat exchanger on
the RO influent and regulate the influent temperature. This is a very viable
option at a facility where auxiliary steam is available.
Conductivity monitoring of the permeate is another important item.
Conductivity is usually measured on-line. As membranes age, they lose some of
their salt removal capacity. This shows up as an increase in conductivity and nor-
mally is not of great concern. Conductivity is much more important for detect-
ing a major problem such as a tear in a membrane or degradation of membranes
by chemicals in the water or microbiological fouling.
Where CA membranes are used, continuous measurement of pH is critical
to detect overfeed, underfeed, or failure of the acid treatment pump. With TFC
membranes, continuous ORP measurement may be needed to monitor perfor-
mance of a dehalogenating chemical feed system.
The flow rate of the reject or concentrate stream of a reverse osmosis system
is always controlled. As water flows along the membranes it carries solids with it.
This prevents the excessive buildup of dissolved solids along the membrane sur-
face. The reject flow rate must not be allowed to drop too low, as otherwise mem-
brane performance could be affected. (This concentration of solids is known as

05.qxd 8/8/06 1:56 PM Page 145
the Beta factor, and is calculated by membrane projection programs supplied by

various manufacturers. It will notify the user if the design criteria selected give a
low reject flow rate.) A value commonly used for determining the reject flow rate
is known as the concentrate-to-permeate ratio. It applies to the individual ele-
ments. Concentrate-to-permeate ratios generally range from 4 to 1 to 9 to 1
depending on the quality of the influent.
RO Alarms
A number of circumstances might occur that could cause damage to a
reverse osmosis system, particularly the membranes. Therefore, RO systems are
typically equipped with a number of alarms or automatic shutdown devices to
protect the equipment. Some of these have been hinted at in the previous sec-
tion. A discussion of the most important ones follows.
High permeate pressure. Membranes are designed to function with flow in
one direction. Backpressure applied to the membranes from the permeate dis-
charge could tear the leaves of the membranes from their support. The system
can be designed to shut down if the permeate pressure reaches a certain limit.
The line can also be equipped with a relief valve for additional protection.
High membrane pressure. Backpressure may be exerted on the membrane
if the reject valve should somehow be closed. This condition must be corrected
quickly. An alternate alarm that would serve this and another purpose is low
reject flow, which would obviously occur if the valve was shut. Low reject flow
is serious because a decrease in flow rate reduces the ability of the concentrate to
sweep solids from the membrane surface. Scaling may result.
High/low influent pH or high ORP. These alarms indicate a problem with
the chemical feed systems that were installed to protect the membranes.
High permeate conductivity. A sudden increase in conductivity indicates
performance problems with one or more of the membranes. The increased salt con-
centration in the permeate could have serious effects on downstream equipment.
Low/high pressure RO feed pump pressure. Low pressure indicates a
problem with the pump. High pressure indicates a possible obstruction or flow
problem in the RO system.
Size of a Reverse Osmosis System

The most common membrane elements are 8 inches in diameter by 40 inch-
es long. Four, five, or six of these mounted in series in a pressure vessel can
extend the vessel length to over 20 feet. It is pressure vessel length that most
determines the size of the system. The cylindrical vessels can be placed in a vari-
ety of parallel configurations, depending on the space requirement. They can be
stacked or placed side by side.

05.qxd 8/8/06 1:56 PM Page 146
When installing a reverse osmosis system in a building, space should be

made available at one or both ends of the skid so that the membranes can be
pulled for maintenance or replacement.
RO Cleaning
Periodically, membrane performance may decline enough to require a chem-
ical cleaning. Because the fouling is on the concentrate side of the membrane, the
chemical cleaning solution is injected through the reject line. The chemical is
mixed up in an external tank and is circulated with a low-pressure pump over
the membranes. Usually, the cleaning system piping network is set up so that
only a portion of the pressure vessels are cleaned at one time. The recommend-
ed flow rate for an 8-inch diameter pressure vessel is 40 GPM.
Common cleaning chemicals include citric acid for calcium carbonate
removal, a carbonate/EDTA chelant for calcium sulfate, and alkaline phos-
phate/EDTA solutions for organically fouled resins. The strength of solution is
usually between 1 and 2%. The solutions work better when warm, so the mix-
ing tank is equipped with a heater to raise the temperature of the cleaning agent
(105˚F is recommended).
For clean up of microbiological foulants, recommended treatments include
a 10 ppm sodium hypochlorite solution for CA membranes and a 400 ppm per-
acetic acid solution for TFC membranes.
Electrodialysis and Electrodialysis Reversal

Electrodialysis (ED) and electrodialysis reversal (EDR) are based upon mem-
brane technology and ionic attraction to an electric field. A simple electrodialy-
sis schematic with the power off is shown in Figure 5-40. Water passes through
cells that are separated by either cationic or anionic membranes. When a poten-
tial is applied across the membranes, cations are attracted to the cathode, and
anions to the anode. Ionic flow is controlled by the membranes. Cations pass
through the anion-impermeable membranes and anions pass through the cation-
impermeable membranes. This generates either purified or concentrated water
within the compartments (Fig. 5-41). In the figure, compartments #2 and #4
contain purified water and compartments #1, #3, and #5 contain concentrate.
The purified and concentrated streams are collected at the discharge of the com-
partments and sent to process and waste, respectively.
Flexibility in system design is possible by varying the number of compart-
ments. Many compartments placed together comprise a stack. The number of
compartments per stack and number of stacks can be selected to produce the
required quantity of water.

05.qxd 8/8/06 1:56 PM Page 147
Figure 5-40
FEED
>
#1 #2 #3 #4 #5 #6
CAT. CAT. CAT. CAT. CAT. CAT.
CATHODE ANODE
AN. AN. AN. AN. AN. AN.
CATIONIC ANIONIC CATIONIC ANIONIC CATIONIC

MEMBRANE MEMBRANE MEMBRANE MEMBRANE MEMBRANE
Schematic of Electrodialysis Cells with Power off.
The electrodialysis process has been modified into electrodialysis reversal, in

which the polarity of the field is regularly alternated. This reduces the formation
of deposits on the membranes.
EDR systems are capable of removing up to 85% of influent dissolved solids
with recoveries of 75 to 85%. The process is not good at removing silica, as sili-
ca is only very slightly ionized and is not strongly attracted to the electrodes. EDR
offers several advantages, most notably:
• Water is not forced through the membranes as it is in RO, so the mem-

branes are less susceptible to fouling.
• EDR membranes are usually constructed of durable material such as poly-
sulfone. This material is resistant to oxidizing biocides, and in fact, main-
taining an oxidant residual in the EDR helps keep the membranes free of
deposits.
The cost of an EDR system is roughly 30 to 50% higher than that of a reverse
osmosis unit. EDR is not extensively used for boiler makeup at this time, being
more popular in the potable water industry. EDR has also shown promise as an
ion concentrator for zero discharge systems.
Electrodeionization
The electrodialysis process has been carried a step further with the develop-

05.qxd 8/8/06 1:56 PM Page 148
Figure 5-41
FEED
>
#1 #2 #3 #4 #5 #6
CAT. CAT. CAT. CAT.

CATHODE ANODE
AN. AN. AN. AN.
CATIONIC N ON C CATIONIC
C O C ANIONIC C O C N ON C CATIONIC
ANIONIC C O C
MEMBRANE MEMBRANE MEMBRANE
MEMBRANE MEMBRANE MEMBRANE MEMBRANE
MEMBRANE MEMBRANE MEMBRANE
Schematic of Electrodialysis Cells with Power On.
ment of electrodeionization (EDI). Essentially, these are ED systems to which

mixed cation and anion exchange resins have been added. EDI systems offer sev-
eral advantages, but two stand out:
• The addition of ion exchange resin greatly improves the process with
regard to silica removal.
• The electric field generates H+ and OH-, which regenerate the resin on-
line. This eliminates acid and caustic regeneration.
An EDI system can produce water of a mixed-bed quality. The units are pri-
marily used as polishers, and a system composed of a reverse osmosis unit fol-
lowed by an EDI unit might be ideal for generating high-purity water with no
chemical regeneration. EDI systems must be protected against the intrusion of
suspended solids because the resin cannot be backwashed as is possible in con-
ventional ion exchange units.
Conclusion
Production of makeup water suitable for feed to steam generating units
requires many different processes. A number of options are available to produce
high-purity water, and some treatment method can almost always be developed.
As always, proper design and operation are essential for good operation.

05.qxd 8/8/06 1:56 PM Page 149
Supplement 5-1
UV-Light Disinfection
Figure 5-42 shows the makeup treatment process at a manufacturing facili-

ty. Flow rate through the system is 150 GPM. Although the steam generating
boilers operate at low pressure, the manufacturing process itself requires very
high purity water, free of microorganisms. Four ultraviolet light contact cham-
bers are located at various points in the makeup process. The UV wavelength in
each chamber is 254 nanometers, which is optimal for biocidal action. These UV
light chambers provide the disinfection needed. In the event of a microbiological
upset, the plant operators can inject iodine and a nonoxidizing biocide to the
makeup water.

05.qxd 8/8/06 1:56 PM Page 150
Figure 5-42
Makeup To RO Reject To
Treatment System Reclaim System
V
RO
Media V Storage Heat UV (3 Units) UV Storage
V
V V V V
Filter Tank Exchanger 254 nm Two-Stage 254 nm Tanks
Sulfuric Iodine* Non-oxidizing

Acid Biocide*
Bypass Line
UV Mixed-Bed Storage 0.3 Micron UV

V
V
254 nm Polisher Demineralizer Tank Filter 254 nm
V
V
Purified Water RO Water
To Plant To Plant
* Supplemental Feed
Flow Diagram of an Actual Makeup Water System Showing UV Treatment Locations
Supplement 5-2
Multimedia Filtration
Media in a filter is sized to provide maximum filtration with minimum plug-

gage. This requires larger, lighter media granules at the top followed by smaller,
denser media in one or two more stages. Lighter media is typically anthracite,
while heavier media may include sand, garnet, quartz, or a combination of these.
A heavy layer of large material such as gravel may be placed at the bottom of the
filter to support the filtration media.
Media densities are selected to allow good re-stratification of the layers upon
backwashing. The following is an actual example of a proposed multimedia fil-
ter for a utility. The particle size of the media reflect typical standards.
Media Particle size Depth

Anthracite 0.85 to 0.95 mm 18 in.
Quartz sand No. 20 8 in.
Garnet 30 to 40 mesh 4 in.
Quartz support bed Graded* 24 in.
*Support bed specified to contain quartz gravel of 1/2 X 3/4 in., 1/4 X 1/2 in.,
and 1/8 X 1/4 in. size.

05.qxd 8/8/06 1:56 PM Page 151
Supplement 5-3
Activated Carbon
Activated carbon is produced by reacting ground-up carbonaceous material

with steam at high temperatures to produce a highly porous, carbon compound.
Activated carbon may be manufactured from several different materials including
coal, wood, and coconut shells. The porous nature of activated carbon granules
give the material great adsorptive properties, especially towards large, natural
organics and oxidizing biocides. Chlorine or other oxidants will be removed in
the first few inches of an activated carbon bed.
Materials that escape from an activated carbon bed can affect downstream
operation. Carbon fines will foul RO membranes, and mineral leakage from
excessive ash deposits will increase loading on a demineralizer. Activated carbon
filter specifications should, therefore, be somewhat detailed so that the facility
receives a product that performs well. Listed below are the major criteria for
selecting an activated carbon.
Material. Coconut shells are normally considered best. Wood is too soft
and coal usually contains too much ash.
Size. Granules in a 12 x 40 mesh size are reported to work well. At one
facility, the specification calls for a 16 x 50 mesh size, with a mean particle diam-
eter of 0.5 to 0.8 mm.
Hardness Number. This number reflects the hardness of the material and
is calculated by testing the material in a ball mill grinder. A hardness number of
95 or greater is recommended.
Iodine Number. This is another laboratory-generated value. It is the mil-
ligrams of 0.02N iodine that can be absorbed by 1 gram of activated carbon. An
iodine number of 900 or greater is recommended.
Ash. An ash content of 0.5% or less is recommended.
Supplement 5-4
Sulfuric Acid and Caustic Specifications
Listed below are maximum contaminant guidelines for sulfuric acid and
caustic purchased for demineralizer regeneration.
Sulfuric Acid
Iron - 20 ppm

05.qxd 8/8/06 1:56 PM Page 152
Copper - 0.5 ppm

Manganese - 0.2 ppm
Chloride - 10 ppm
Sodium Hydroxide (Rayon Grade)
Sodium Carbonate - 0.5%
Sodium Chloride - 0.5%
Sodium Sulfate - 0.5%
Silica - 10 ppm
Iron - 10 ppm
Chlorates - 100 ppm
Supplement 5-5
Silt Density Index
The silt density index (SDI) is a measure of the fouling potential of the water.
A simple test is available to calculate SDIs. In the test, water is filtered through a
0.45 micron membrane at a pressure of 30 psig. Measurement is taken of the
time for 500 ml of water to pass through the filter. The stream is allowed to con-
tinue flowing through the filter for a set period of time (5, 10, or 15 minutes),
after which the filtration rate is again measured. The SDI is calculated by the fol-
lowing equation:
SDI = 100 ( (1 ( ti/tf)/ T, where

ti = Initial 500 ml flow measurement
tf = Final 500 ml flow measurement
T = Time between measurements (15 minutes is standard)
As an example, consider the following readings that were actually taken by

utility personnel evaluating the feasibility of installing a reverse osmosis unit
ahead of a demineralizer.
ti = 34 seconds
tf = 66 seconds
T = 15 minutes
SDI = 100 x (1-34/66)/15 = 3.2
Listed below is a simple BASIC program that will automatically calculate the
SDI when the user inputs values for ti, tf, and T. This program can very useful if
the plant chemist must routinely monitor silt density indices.

05.qxd 8/8/06 1:56 PM Page 153
SDI.BAS
10 CLS:LOCATE 5,1
20 PRINT “SILT DENSITY INDEX CALCULATION”
30 PRINT
40 PRINT “ENTER THE TIME IN SECONDS TO COLLECT 500 ML OF SAMPLE.”
50 INPUT TI
60 PRINT
70 PRINT “ENTER THE TIME N MINUTES BETWEEN READINGS. 15 MINUTES IS STAN-
DARD.”
80 INPUT T
90 PRINT
100 PRINT “ENTER THE TIME IN SECONDS TO COLLECT 500 ML DURING THE FINAL
SAMPLING.”
110 INPUT TF
120 PRINT
130 SDI=100*(1-(T1/TF))/T
140 PRINT “THE SDI AT “;:PRINT USING “##”;T;:PRINT “ MINUTES = “;:PRINT
USING “##.#”;SDI

05.qxd 8/8/06 1:56 PM Page 154

06.qxd 8/8/06 1:57 PM Page 155
Chapter 6
Cooling Water
Chemistry
Introduction
Cooling of fluids is an essential process at power-generation and industrial
plants. Control of cooling water chemistry is very critical in preventing corrosion
and fouling, and ensuring equipment reliability.
The most important uses for cooling water include:
• Condensing turbine exhaust steam

• Cooling process fluids
• Protecting high pressure pump bearings
The volume of cooling water required for these processes, especially steam
condensation, is often quite large. (Sufficient cooling water flow to steam con-
densers allows the condensed steam to reach its lowest temperature, which is
important with regard to the thermodynamic concepts outlined in Supplement
2-1.) High cooling water flow rates usually prohibit the mechanical purifying

06.qxd 8/8/06 1:57 PM Page 156
techniques (demineralization, RO) that are used for smaller purposes such as
boiler makeup production. Yet, some form of treatment is required to protect
equipment. The following sections discuss types of cooling systems, water qual-
ity variations between different sources of supply, and treatment methods.
Changing environmental regulations and improvements in cooling water treat-
ment chemicals have added a complexity to treatment selection.
Cooling Systems
The type of cooling system most suitable for a process depends upon
process operation, flow requirements, availability and quality of water, and envi-
ronmental requirements regarding discharge. Most cooling systems belong to one
of the following categories:
• Once-through
• Open recirculating (evaporative cooling towers)
• Closed recirculating
Dry cooling systems for steam condensation are becoming more popular,
especially at facilities where makeup water is scarce or where liquid discharges
are restricted or banned. A detailed discussion of dry cooling systems is outside
the scope of this book.
Once-Through Systems
Once-through systems are most often used to cool turbine exhaust in a con-
denser and provide service water to the plant. Flow rates range from tens of thou-
sands to hundreds of thousands of gallons per minute. The source of supply for
once-through systems is usually a lake, river, or ocean due to the large quantities
of water needed.
The greatest advantage of once-through cooling is that the water does not
become concentrated as it does in a cooling tower, and thus has a much lower
scale-forming or corrosion potential. (Sometimes, as Case History 6-1 illustrates,
unusual events can upset the somewhat steady-state conditions of a once-
through source.) Another advantage is that the open body of water which serves
as the supply will usually be cooler than water from a corresponding cooling
tower. This improves the thermodynamic efficiency of the condenser.
The primary disadvantage of once-through cooling is that the water is
returned to the source at a higher temperature. Increasingly stringent regulations
on thermal discharge and the effects on aquatic organisms are restricting new
construction of once-through systems, and it may take a variance from environ-

06.qxd 8/8/06 1:57 PM Page 157
Cooling Water Chemistry 157
mental permitting agencies to install a once-through system.
Open Recirculating Systems

In an open recirculating cooling system, water circulates through the con-
denser or heat exchanger to a cooling tower and then is returned to the exchang-
er (Fig. 6-1). An open cooling system can provide the same high-volume flow
rate as a once-through system, but with much less water discharge.
Cooling Towers
Figure 6-1
EVAPORATION & DRIFT
WARM WATER RETURN

<
COOLING
MAKEUP > TOWER TURBINE
>
>
>
COLD WATER SUPPLY

BLOWDOWN
>
Basic Recirculating Water Loop CONDENSER
Figure 6-2
Cooling Tower. Photo courtesy of the Marley Cooling Tower Company, Overland Park, KS.

06.qxd 8/8/06 1:57 PM Page 158
Figure 6-3
>
AIR
OUT
CIRCULATING
WATER IN CIRCULATING
WATER IN
FILL
<
AIR AIR
> <
IN IN
>
TO
BASIN CONDENSER
Dual-Entry, Crossflow Cooling Tower Arrangement.
The heart of an open system is the cooling tower (Fig. 6-2), which cools
water by cascading it through air. Modern cooling towers can basically be divid-
ed into two categories, mechanical draft (air forced through the tower) and nat-
ural draft (air flows naturally through the tower). Mechanical draft towers can in
turn be divided into two other categories, crossflow and counterflow (single or
dual entry), in which the air is either pushed (forced-draft) or pulled (induced-
draft) through the tower. Figures 6-3 and 6-4 illustrate the basic outline of dual-
entry crossflow and counterflow towers, respectively. The principle of operation
is relatively straightforward. Warm cooling water is sprayed into flowing air,
which absorbs heat. The warmed air is ejected from the tower, while the cooled
water falls to a basin where it is recirculated. The falling water is broken up by
fill material in the tower, which enhances air/liquid contact and heat transfer.
Both the crossflow and counterflow design have advantages and disadvan-
tages. Air pressure drop and water pump head requirements may be greater in a
counterflow tower, but cooling efficiency is normally a bit higher as well.
Maintenance on crossflow towers is often easier, but these towers, being shorter
in height, are more susceptible to recirculation of discharged air to the air intake.
Fan location also has an effect on tower performance, particularly on power
usage and durability of the fan materials. In an induced-draft tower, where a fan
is located at the air outlet of each cell, the fan blades, shafts, and other compo-
nents are constantly subjected to very humid air flows containing trace amounts
of suspended solids. The tower structure must be strong to handle the fan vibra-
tion, and the elevated location of the fans can hamper maintenance. Forced-draft
fans, located at the air inlets, are not subjected to such a humid environment, but
they cannot produce the same air exit velocity. Lower exit velocities increase the
possibility of air recirculation to the cooling tower inlet. Both forced and induced
draft towers are equipped with louvers on the air inlets to regulate air flow. This

06.qxd 8/8/06 1:57 PM Page 159
Figure 6-4
>
AIR
OUT
< MIST ELIMINATORS
CIRCULATING WATER IN
< FILL
AIR > < AIR

IN IN
>
TO
BASIN CONDENSER
Dual-Entry, Counterflow Cooling Tower Arrangement.
can be of benefit in winter to help minimize icing.

For very large cooling water systems, natural-draft, counterflow towers are
common. These are typified by the several-hundred-feet tall hyperbolic towers
that one sees at a large utility, and that many citizens often mistake for nuclear
reactor buildings. Hyperbolic towers (Fig. 6-5) generate a chimney effect to
induce air flow. As the air warms during its contact with the sprays, it rises and
flows out of the tower, thus pulling in cooler air from below. As Figure 6-5 illus-
trates, cooling takes place in the lower portion of the tower. The remaining
height of the tower generates the natural draft. Hyperbolic towers are designed
in this shape because the hyperbolic shape imparts a great deal of natural
strength.
Natural draft towers trade savings in fan horsepower for increased pumping
head and higher capital cost. They perform best in humid locations and for win-
ter peaking applications. However, they can be more seriously affected by ice for-
mation in cold climates, since there is no way to reduce air flow.
One of the most important components of any cooling tower is fill. Fill

06.qxd 8/8/06 1:57 PM Page 160
Figure 6-5
>
AIR
OUT
MIST ELIMINATORS
<
RETURN WATER
< DISTRIBUTION SYSTEM
FILL
<
WARM WATER >

< AIR
RETURN
IN
>
TO
BASIN CONDENSER
Hyperbolic Cooling Tower Arrangement.
breaks up the water flow to enhance air/water contact, and it comes in many
shapes and patterns. Wooden splash bars were once the primary material, but
these have mostly been supplanted by more sophisticated designs and materials,
as is illustrated in Figures 6-6 and 6-7. These include vertically aligned sheets of
corrugated plastic known as film fill, which is used in counterflow towers. Film
fill, while being very efficient, can be problematic if microbiological fouling or
scaling is not carefully controlled, as the close spacing of film fill layers makes the
material susceptible to plugging. In fact, the increasing use of film fill has been a
driving factor for improvements in treatment programs.
Cooling Tower Calculations

A cooling tower ís primary purpose is to remove heat while minimizing
water usage. Heat is transferred by two mechanisms. A portion of the cooling
water, generally 1 to 3%, actually evaporates as it mixes with the air. Latent heat
is given up in this phase change. Sensible heat transfer, in which heat is
exchanged without a phase change, makes up the balance. A simple formula
exists for calculating the approximate amount of evaporation that occurs in a

06.qxd 8/8/06 1:57 PM Page 161
Figure 6-6
Plastic Splash Bars. Photo courtesy of the Marley Cooling Tower Company, Overland Park, KS.
cooling tower:
E = (ƒ x R x ∆T) / 1000, (6.1)

where
E = Evaporation in gallons per minute (GPM)
R = Recirculation rate of the cooling water (GPM)
∆T = Temperature difference (range) between the hot and cool circulating
water (˚F)
ƒ = A correction factor for evaporation,
where
ƒ (average) = 0.75
ƒ (in summer) = 0.85
ƒ (in winter) = 0.65
1000 is the approximate latent heat of vaporization
The ƒ factor is the ratio of the amount of heat exchanged by evaporation ver-
sus sensible heat transfer, and is a function of wet bulb temperature and relative
humidity.
The number of times that water is concentrated in the cooling tower is
known as the cycles of concentration (C). C is usually determined by a compar-
ison of some very soluble ion, such as chloride or magnesium, in the makeup to

06.qxd 8/8/06 1:57 PM Page 162
Figure 6-7
Film Fill. Photo courtesy of the Marley Cooling Tower Company, Overland Park, KS.
the recirculating water. Specific conductivity is also used for this determination,
although it is not completely linear with increasing dissolved solids.
Conductivity is also influenced by treatment chemicals added to the water.
The maximum cycles of concentration depend on the effectiveness of corro-
sion and scale inhibitor programs, and on the water quality of the makeup to the
tower. Soft water or softening of hard water may allow for increased cycles of
concentration due to the reduced scaling potential, although softening may
increase the corrosion potential. Cycles must be kept low when the makeup sup-
ply is scale forming or highly ionic. Seawater is a prime example of the latter.
Cycles of concentration in a seawater-cooled tower may be limited to less than
two due to the high ionic concentration of the recirculating water, and its poten-
tial for corrosion.
Regardless of the quality of the makeup water or effectiveness of a corrosion
or scale inhibitor program, cycles of concentration have a limit in any system.
Some water must be continually removed to prevent excessive buildups of dis-
solved solids. This is known as blowdown (BD). An additional, very small blow-
down occurs due to entrainment of water droplets in the air exiting the tower.
This is known as drift (D). Drift typically ranges from about 0.3 to 0.05% of the
recirculation rate, depending on the type and efficiency of the cooling tower.
Some of the more modern towers may have drift values below 0.01%.
Evaporation, blowdown, drift, and makeup comprise the water balance
around a cooling tower (Fig. 6-8), and a set of simple equations has been derived
to calculate this balance. Evaporation has already been given. The others are:

06.qxd 8/8/06 1:57 PM Page 163
Figure 6-8
EVAPORATION
&
DRIFT
<
CIRCULATING LOOP
MAKEUP TO CONDENSER AND BACK
> >
>
BLOWDOWN
Cooling Tower Water Balance.
MU (flow) ≈ BD Cl- or Mg++ conc. (6.2)

C=
BD (flow) MU Cl- or Mg++ conc.
BD + D = E / (C – 1) (6.3)
MU = E + BD + D (6.4)
Supplement 6-1 provides a BASIC program that calculates the evaporation,

blowdown, and makeup flow rates of a cooling tower based on user-inputted val-
ues for recirculation rate, ∆T, and cycles of concentration.
Drift is a form of blowdown, and it can have a significant impact on blow-
down calculations. For the previous example (R = 100,000 GPM, ∆T = 20˚F, and
ƒ = 0.75), blowdown at five cycles of concentration is 365 GPM. Drift at a 0.01
percentage rate of recirculation is 10 GPM, or 2.6% of the total blowdown. If the
tower is operated at 15 cycles of concentration, blowdown is lowered to 97 GPM,
but drift is still 10 GPM. The drift now accounts for almost 10% of the total
blowdown.
The most critical value in determining cooling tower efficiency and size is
the wet-bulb temperature of the entering air. Wet-bulb temperature is a measure
of the maximum cooling capabilities of air, and is a function of the actual (dry-
bulb) temperature and moisture content (relative humidity) of the air. Wet-bulb
temperatures are often determined by sling psychometry in which a wetted fab-
ric is placed upon the bulb of a thermometer, and then the thermometer is spun
in a circular motion to induce evaporation. The minimum temperature reached
in this process is the wet-bulb temperature, and it represents the theoretical min-
imum temperature that could be attained in a cooling tower. Because the wet-
bulb temperature is dependent on both the dry-bulb temperature and the mois-
ture content of the air, a virtually infinite amount of conditions are possible for
any wet-bulb reading. For instance, in the cooling tower heat transfer example

06.qxd 8/8/06 1:57 PM Page 164
Figure 6-9
Relative Cooling Tower Size versus Approach Temperature. Reprinted with permission from Cooling
Tower Fundamentals by the Marley Cooling Tower Company, Overland Park, KS.
listed earlier, 75% of the heat exchange was due to evaporation and 25% to sen-
sible transfer. A 100% humid air could provide the same cooling if the wet-bulb
temperature were the same. For this to happen, the dry-bulb temperature would
have to be much lower. All heat exchange would be by sensible transfer. This is,
of course, an absurdity when designing cooling towers, but it does illustrate the
factors that must be taken into account.
Cooling towers can never reach 100% efficiency no matter where they are
located. The difference between the temperature of the cooled water and the wet-
bulb temperature is known as the approach. According to the Marley Cooling
Tower Company, the closest approach that cooling tower manufacturers are usu-
ally willing to guarantee is 5˚F. Figure 6-9 shows the tower size factor versus the
approach temperature. As can be seen, the typical design approach temperature
is 15˚F. An approach temperature of half that amount would double the size of
the cooling tower, and the minimum guaranteed approach of 5˚F would require
a tower over 2º times larger. The curve is asymptotic, which essentially precludes
closer approach temperatures than 5˚F due to economic and size constraints of
the tower. Figures 6-10 and 6-11 show an example of daily and seasonal varia-
tions of wet-bulb temperatures. These may play a very important part in the
design of the cooling tower. For example, seasonal operation of a cooling tower
might require an evaluation of several wet-bulb temperatures. Additional factors,
such as the proximity of the tower to sources that might affect the entering wet-
bulb temperature, i.e., other cooling towers, also may require adjustment of the
entering wet-bulb temperature. A reputable cooling tower firm will take all of
these factors into account when designing the tower. The performance of the
tower can change over time due to degradation if it is not properly maintained.

06.qxd 8/8/06 1:57 PM Page 165
Figure 6-10
Illustration of Daily Wet-Bulb Temperature Variations. Reprinted with permission from Cooling Tower
Fundamentals by the Marley Cooling Tower Company, Overland Park, KS.
Cooling towers have been constructed from a number of different materials.

For many years, the principal material for mechanical draft cooling towers was
redwood or pressure-treated Douglas fir. Surprisingly, given all the water that
flows through a tower, wood towers are susceptible to fire, especially during unit
outages, and a number of them have burned down over the years. Other materi-
als have, to some extent, supplanted wood-frame cooling towers. These materi-
als include concrete, fiberglass-reinforced plastics, and even ceramics. Concrete
is the material of choice for hyperbolic cooling towers because of the tower size.
Concrete has also become a more popular material for mechanical-draft cooling
towers. Galvanized steel is often used for smaller, prefabricated commercial tow-
ers, but it is very susceptible to corrosion, particularly because a cooling tower
constantly replenishes the oxygen content of the circulating water. Stainless steel
provides much better protection, but at a much higher material cost. Its use is
limited to small, commercial towers.
Closed Cooling Systems

Closed cooling systems require very little makeup and are used for relative-
ly small plant heat exchange systems such as pump bearing coolers and building
heat. Because water losses in a closed system are usually minor, chemical treat-
ment of these systems is uncomplicated. Exceptions occur if the system is sub-
ject to major water leaks or air intrusion. More information about closed cooling
water systems appears in the following sections.

06.qxd 8/8/06 1:57 PM Page 166
Figure 6-11
MONTH
Common Seasonal Wet-Bulb Temperature Variations. Reprinted with permission from Cooling Tower
Fundamentals by the Marley Cooling Tower Company, Overland Park, KS.
Cooling Water Corrosion, Scale, and Deposit

Mechanisms
Due to water’s ability to dissolve most substances to some extent, and its
ability to support microbiological life, every cooling water system is subject to
potential operational problems. These include:
• Corrosion
• Scaling
• Fouling
• Microbiological fouling
These phenomena and their effects on cooling systems are often interrelat-
ed. For example, microbiological deposits restrict heat transfer similar to scales
and also generate under-deposit corrosion cells. Slime produced by microorgan-
isms will trap silt and increase fouling. Corrosion products may dislodge and
form deposits in heat exchangers. The source of water and the type of system
(once-through, recirculating, and closed) have a great impact on these phenom-
ena. Environmental considerations are playing a major role in control methods.
Corrosion
Corrosion control requires careful evaluation because of the many and var-
ied corrosion mechanisms that can occur. Cooling water systems may contain
several different metals, including carbon steel water lines, copper-alloy heat
exchanger tubes, galvanized structural supports, bolts, nuts, etc. These may all
suffer corrosion, sometimes from chemicals added for scale or foulant control.
Even concrete and wood components are subject to attack and degradation.
Corrosion of metals is an electrochemical reaction in which an electron
transfer takes place between the corroding material and the corrosive medium.
The driving force for corrosion reactions is the potential between the electron

06.qxd 8/8/06 1:57 PM Page 167
acceptor (the corrosive medium) and the electron donor (the metal). Every metal
exhibits a different tendency to release or accept electrons in a corrosion cell.
This tendency is measured against the standard hydrogen ion half cell, in which
a one molar solution of hydrogen ions is assigned a half cell potential of zero.
2H+ + 2e- H2 E0 = 0.00V (6.5)
Elements such as zinc and iron, release electrons to generate hydrogen.
Fe Fe+2 + 2e- E0 = -0.44V (6.6)

2H+ + 2e- H2 E0 = 0.00V (6.5)
——————————-
Fe + 2H+ Fe+2 + H2 E = -0.44V (6.7)
Table 6-1 These elements are said to be

reactive. Elements that do not release
Standard Oxidation-Reduction
hydrogen in acid solutions are said to
Potentials For Some Elements at 25˚C
be passive or noble. Table 6-1 illus-
trates the electrochemical potentials
Metal Standard Potential
for several common cooling water
Volts
system construction materials.
Aluminum -1.67
A number of factors influence
Zinc -0.76
the rate of reaction of metals in acids,
Iron -0.44
but the most important is the acid
Hydrogen 0.00
concentration. The higher the hydro-
Copper +0.34
gen ion concentration, the faster the
reaction proceeds. This is the primary
reason that the pH of many iron-based piping systems, such as boiler tubes and
feedwater networks, is maintained within an alkaline range. At a pH range of 9
to 10, commonly maintained in most boilers, hydrogen ion concentrations are
only 10-9 to 10-10 grams per liter. This is much too low to cause hydrogen attack.
If acids were the only corrodent, corrosion would be easy to control.
However, many other corrodents exist, of which the prime culprit is oxygen. One
of the most common corrosion reactions is illustrated in Figure 6-12. The fol-
lowing processes make up the corrosion cell:
• Iron releases electrons, which travel through the metal to another site
where they reduce oxygen and water to hydroxide ions. The site of reduc-
tion is known as the cathode.
• Oxidized iron atoms leave the metal substrate and enter the solution. This
site is known as the anode.

06.qxd 8/8/06 1:57 PM Page 168
Figure 6-12
CATHODE ANODE
+2 +2
Fe Fe
+2
O2 + H2 O > O H- Fe +2
v
Fe
<
e-
PIPE WALL
Steel Corrosion Cell.
• The iron and hydroxide ions combine to form a precipitate. Under the con-
ditions shown, the precipitate eventually converts to rust (Fe2O3 • xH2O).
The corrosion cell is written:

2Fe 2Fe+2 + 4e- (Anodic reaction) (6.8)
O2 + 2H2O + 4e- 4OH- (Cathodic reaction) (6.9)

_________________________
2Fe + O2 +2H2O 2Fe(OH)2 (6.10)
The Fe(OH)2 produced in this reaction is eventually converted to rust

(Fe2O3 • xH2O) by further reaction with dissolved oxygen.
Equation 6.9 is the principle cathodic reaction in neutral or alkaline solu-
tions. Other prominent cathodic reactions include:
O2 + 4H+ + 4e- 2H2O (6.11)

2H2O + 2e- H2 + 2OH- (6.12)
Several important aspects of corrosion and these reactions should be men-

tioned. First, as Equations 6.8 and 6.9 illustrate, electron flow is balanced
between the anodic and cathodic half-cell reactions. Severe corrosion results
when limited anodic sites are present in a large cathodic environment. This
occurs underneath deposits. The liquid underneath the deposit becomes oxygen
depleted, causing the metal at this site to become anodic to bare metal locations
where oxygen is in good supply. As the corrosion proceeds, the pit continues to
increase in depth and may eventually penetrate the tube wall. Pitting is a very
insidious mechanism, and tube failure can occur with relatively small metal loss.
Second, corrosion via the mechanisms shown in Equations 6.5 and 6.11 has
been common in systems treated with acid for scale control. Overfeed of acid
increases the potential of these reactions, as they are dependent upon the hydro-
gen ion concentration.
As was illustrated in chapter 2, copper alloys form a layer of cupric oxide
when placed in service. This layer tends to protect the base metal from corrosion.

06.qxd 8/8/06 1:57 PM Page 169
Also, copper is a noble metal and does not undergo the simple acid attack shown
in Equations 6.5 and 6.6. Copper alloys can corrode, however, in the presence
of compounds that complex or bond with copper. Ammonia, of course, is the
primary copper complexor, however its concentration in natural waters is usual-
ly very small. Ammonia can become a problem where treated sewage water is
used for cooling water makeup.
Sulfide is another problematic corrodent, and copper alloys should not be
used to handle waters that contain sulfides. Case History 6-2 discusses corrosion
of copper-alloy condenser tubes by sulfides from an unexpected source.
Reports suggest that the most common water-side corrosion problems are
due to dezinkification or denickelification of Admiralty metal and copper-nickel
alloys, respectively. These corrosion mechanisms are influenced by excessive
chlorination, low pH, and high chlorides. Some copper alloys, particularly
Admiralty, are soft and may suffer erosion from the flowing cooling water. This
condition typically occurs at the inlet end of the tubes.
Oxygenated acid solutions are also detrimental. Corrosion of copper-alloy
condenser tubes by these types of solutions has occurred in cooling systems
treated with sulfuric acid when the acid feed system malfunctioned and lowered
the pH of the oxygen-containing, cooling water.
Corrosion Influencing Mechanisms

Corrosion is influenced by mechanical factors. Anodes and cathodes will
develop in a metal just due to surface irregularities. Temperature affects corrosion
rates in more than one way. Increasing temperature usually increases the rate of
reactions, just as it does for most chemical reactions. Corrosion rates of steel are
generally much higher in summer than winter. Conversely, colder water increas-
es the solubility of some substances, i.e., oxygen, and can introduce more corro-
dent to the metal surface.
Galvanic corrosion occurs when two dissimilar metals are physically con-
nected in an aqueous environment. The corrosion may become quite serious if a
reactive metal with a small area is physically connected to a more noble metal
with a larger area. The nonreactive metal acts as a cathodic site, and the reactive
metal as an anodic site. Since the ratio of cathode to anode is large, the anodic
metal may fail in a short period of time. A classic example is the use of steel bolts
for support of copper-alloy tube sheets in cooling water heat exchangers. The
bolts fail in a remarkably short period of time. The reverse situation, copper bolts
attached to steel sheets, is much better because a very large anode is coupled to
a very small cathode. Although galvanic corrosion still occurs in this situation,
the effect is spread across the large anode, thus minimizing concentrated attack.
Galvanic effects must be taken into consideration for any system containing dis-

06.qxd 8/8/06 1:57 PM Page 170
similar metals. Methods to minimize this problem include limiting the

anode/cathode ratio, using impressed-current cathodic protection or sacrificial
anodes, coating the metal surfaces to prevent their contact with the cooling
water, or electrically insulating the two dissimilar metals. Dielectric unions con-
necting copper and iron pipes are a common example of this latter method.
Nonmetallic Corrosion
Other cooling water system materials can degrade as well, albeit by different
mechanisms than those shown above. Wood may be subject to degradation (rot)
by various types of fungi, which attack either the cellulose or lignin structure of
the material. Concrete degradation may be caused by too many or too few ions
in the water. Both chloride and sulfate can attack concrete, although special types
of concrete are available to resist these ions. Sulfate attack is prevalent in many
cooling towers operated at high cycles of concentration using sulfuric acid to
control pH. Chlorides at high concentration will attack the reinforcing bars and
cable in concrete.
Conversely, concrete may corrode if the water is too soft because calcium
ions will leave the concrete to establish equilibrium with the water. Supplemental
calcium additions to the water may be needed to minimize this reaction.
Protective coatings that serve as moisture barriers can also be beneficial.
Corrosion Inhibitors
Corrosion inhibitors work by protecting the material surface. Corrosion is
influenced by a wide variety of conditions, and it is often difficult to precisely
predict when and where it will occur. A corrosion inhibitor program should be
designed to provide total protection.
Inhibitors effectively depolarize (reduce or stop the electrical flow of) the
corrosion reaction. Individual corrosion inhibitors provide anodic or cathodic
protection, or both when blended. In general for cooling applications, cathodic
inhibitors precipitate at the locally high pH cathodic site to form a barrier that
limits the rate of oxygen reduction. Anodic inhibitors generally promote the for-
mation of a stable metal oxide at the anodic surface. This limits metal dissolu-
tion. The most common mild steel corrosion inhibitors include:
Anodic Cathodic
Molybdate Zinc
Orthophosphate Polyphosphate
Nitrite Phosphonate
Silicate

06.qxd 8/8/06 1:57 PM Page 171
The mechanisms by which the anodic inhibitors perform is revealing.

Molybdate combines with iron ions formed at anodes to produce a molecular fer-
ric-molybdate film at the metal surface. If molybdate residuals are properly main-
tained, the film will eventually cover the entire metal surface. Orthophosphate
actually serves as both an anodic and cathodic inhibitor. It forms phosphate com-
plexes with metal ions at the anode, but when calcium is present in the water will
form a calcium phosphate precipitate at the cathode. Nitrites, which are general-
ly restricted to closed cooling water treatment, actually passivate metal and cause
it to form its own protective layer. Silicates react with metal ions at the anode to
form a protective film.
Anodic corrosion inhibitors can cause problems if not properly applied.
Molybdate films are not exceptionally strong when used alone and may require a
high molybdate residual (50 ppm or more) to maintain the protective layer. This
can be expensive. Orthophosphate may form calcium and iron phosphate sludges
that retard heat transfer in heat exchanger tubes. Nitrites are a source of nutrition
to microbes, which almost totally eliminates their use in open cooling systems.
(Even in closed cooling systems, microbiological fouling may be a problem. See
Case History 6-3). Silicates, which are weak inhibitors anyway, may be inappro-
priate for cooling towers because silica alone or combined with magnesium may
begin to form tenacious deposits when silica concentrations reach 150 ppm.
Anodic inhibitors also present another problem. If the system is under-
dosed, a few anodic sites may remain in a large cathodic environment. This is dif-
ferential area affect can cause severe pitting at the anodic sites. If a proper resid-
ual cannot be maintained, it is often better to suspend the treatment program.
Cathodic inhibitors perform similarly to anodic inhibitors in many respects.
Zinc forms a hydroxide precipitate at the cathode. Polyphosphates, like
orthophosphate, combine with calcium to form a precipitate. Phosphonates
(organo-phosphorous compounds) behave similarly and form a deposit on the
cathodic surfaces.
Cathodic inhibitors can cause difficulties under certain conditions. Zinc salts
are not particularly tenacious and may wash away. Zinc is also toxic to some
aquatic organisms, and its use has been curtailed due to this factor.
Polyphosphates break down to form orthophosphate, which can increase scale
and sludge formation in the system. Some reports claim that calcium phosphate
has become the most common scale in cooling water systems. Phosphonates are
degraded by oxidizing biocides, which reduces the effectiveness of the inhibitor
and generates orthophosphate. Because a single inhibitor is often inadequate, a
combination of inhibitors is frequently used. For example, molybdate and zinc

06.qxd 8/8/06 1:57 PM Page 172
Table 6.2 will provide anodic and

cathodic protection.
Common Corrosion Inhibitors
Molybdate and orthophos-
and Dosage Levels
phate provide combined
anodic protection, and
Inhibitor Dosage Range (ppm)
zinc/phosphonate provide
Molybdate 25–50
combined cathodic protec-
Orthophosphate 5–15
tion. Table 6-2 illustrates rec-
Polyphosphate 10–30
Phosphonates 5–10 ommended dosage ranges for
Molybdate/Phosphate 5–10/5–10 various combined inhibitor
Molybdate/Phosphonates 5–10/5–10 programs.
Zinc/Phosphate 1–3/5–10 Two factors have had a
Zinc/Phosphonates 1–3/5–10 great impact on the use of
corrosion inhibitors. First,
environmental regulations on
cooling water discharges are becoming more and more strict. Specifically, metals
concentrations in cooling tower blowdown directly discharged to a receiving
body of water are being severely curtailed. This is primarily directed towards
zinc. Molybdenum is not very toxic to aquatic life. Second, many cooling tower
chemists and operators have switched to alkaline treatment programs, in which
corrosion is minimized naturally. Where corrosion inhibitors are still used, the
trend is phosphate/phosphonate treatments with perhaps a supplemental poly-
mer addition to inhibit calcium phosphate scale formation.
A point should be made here about chromate. Chromate is probably the
most effective corrosion inhibitor because it establishes an iron-chromate surface
layer that causes mild steel to behave as if it were stainless. Mild steel corrosion
rates of less than 1 mil per year (mpy) were common.
Figure 6-13 Chromate was the inhibitor of choice for many cool-
H
ing water systems, especially open recirculating sys-
CH 3 N tems. Unfortunately, chromium in the +6 oxidation
N state has proven to be harmful to aquatic life, and its
N aerosols have been declared carcinogenic. Chromate
use has virtually disappeared in the United States in
Tolyltriazole. systems that discharge to open bodies of water or
where the water is aerated across a cooling tower. The
Figure 6-14
alternative treatments mentioned above have
H
replaced chromate primarily because of the environ-
C 4 H9 N mental issue. While not as effective as chromate,
N these programs can still reduce mild steel corrosion
N below 5 mpy, and sometimes 1 mpy.
Copper corrosion prevention programs are com-
Butylbenzyltriazole.
monly referred to as yellow-metal treatments. Azoles

06.qxd 8/8/06 1:57 PM Page 173
such as tolyltriazole (TTA, Fig. 6-13) or butyl-benzotriazole (Fig. 6-14) are most
common. The azoles function by bonding to the cuprous oxide layer through
nitrogen atoms on the molecule. The azole not only stabilizes the cuprous oxide
from further oxidation, but the planar azole molecules form a barrier on the
metal surface. Azole dosages of 1 to 2 ppm are common. The azoles, which dis-
solve much more readily in basic solutions, have become popular in blended for-
mulations of alkaline scale inhibitors.
Scale
Scale and deposit control are very important in cooling systems because of
the effects that deposits have on heat transfer, corrosion of metals, and fouling of
cooling tower film fill. The following sections outline methods for monitoring
and treatment of scale in cooling water systems.
Calcium Carbonate
The potential for scale formation is much Table 6-3
higher in open, recirculating cooling systems,
Most Common Cooling
where the solids concentration may be 5, 10, or
Water Scales
perhaps 20 times greater than that in the make-
up. The common cooling water scales are listed Calcium Phosphate
in Table 6-3. Most of these compounds are cal- Calcium Carbonate
cium based, with the two most common being Silica
calcium phosphate and calcium carbonate. Calcium Sulfate
Calcium carbonate scale forms when calci- Calcium Fluoride
um and alkalinity exceed saturation values and
begin to drop out as calcium carbonate (CaCO3). A number of programs have
been developed to calculate scaling potentials. In the 1930s, Langelier developed
the first set of calculations for this purpose, and the value derived from the cal-
culations is known as the Langelier Saturation Index (LSI). Calcium carbonate
precipitation is dependent not only on the calcium and alkalinity in solution but
also on the pH. Langelier’s equations take this into account:
pHs = (pK2 – pKs) + pCa + pAlk. (6.13)

LSI = pHa – pHs, (6.14)
where
pHs is the Langelier Saturation pH
pK2 is the second dissociation constant for carbonic acid
pKs is the solubility product constant for calcium carbonate

06.qxd 8/8/06 1:57 PM Page 174
pCa is the negative log of the calcium concentration

pAlk is the negative log of the alkalinity concentration as based on a
titration procedure
pHa is the actual pH
An LSI less than zero indicates that the water tends to dissolve scale, and
thus may be corrosive. An LSI greater than zero indicates that the water may be
scale forming.
Since Langelier introduced his equations, the ease of calculating the LSI has
been improved. For example, Equation 6.13 may now be written as:
pHs = (9.3 + A + B) – (C + D) (6.15)
Values for A, B, C, and D can be found in Table 6-4. Also, nomographs have
been developed for calculating the LSI. One of these is shown in Figure 6-15.
The user determines the values for pCa, pAlk, and C, and the sum of these val-
ues is then subtracted from the actual pH to give the LSI. The LSI can also be
approximated from the following equation:
LSI = pH ( [9.4 ( log10Ca – 0.97log10MALK + (log10TDS/10.7) – 3.24exp(-T/191)

(6.16)
where
Ca = Calcium as mg/l, CaCO3
MALK = Total alkalinity as mg/l, CaCO3
TDS = Total dissolved solids as mg/l
T = Heat exchanger water-side skin temperature, ˚F
Although the Langelier Saturation Index has proven to be useful, it does not
always provide the accuracy needed due to inherent errors in the calculations. In
particular, the calculations do not account for ionic interactions in the water.
To improve the accuracy of the calculation, Ryznar developed the Stability
Index:
RSI = 2pHs – pHa (6.17)
The scale of the RSI is different than that of the LSI. Waters with an RSI of
less than 6 are scale-forming, while those above 7.5 exhibit corrosive tendencies.
The LSI and RSI are often used in combination, although the RSI is based on the
same calculations as the LSI and provides no additional improvement with
regard to ionic interactions. Supplement 6-2 outlines a BASIC program that cal-
culates LSI and RSI using Equation 6.15. Like the original calculations by

06.qxd 8/8/06 1:57 PM Page 175
Table 6-4
Data for Rapid Calculation of Saturation & Stability Indexes (Based on Langelier for-
mulas, Larson-Buswell residue, temperature adjustments & arranged by Eskel Nordell)
A B C D
Total Solids Temperature Calcium M.O.
(ppm) (F.) Hardness Alkalinity
A B (ppm or CaCO3) C (ppm or CaCO3) D
50–300 0.1 32–34 2.6 10–11 0.6 10–11 1.0
400–100 0.2 36–42 2.5 12–13 0.7 12–13 1.1
44–48 2.4 14–17 0.8 14–17 1.2
50–56 2.3 18–22 0.9 18–22 1.3
58–62 2.2 23–27 1.0 23–27 1.4
64–70 2.1 28–34 1.1 28–34 1.5
72–80 2.0 35–43 1.2 35–43 1.6
82–88 1.9 44–55 1.3 44–55 1.7
90–98 1.8 56–69 1.4 56–69 1.8
100–110 1.7 70–87 1.5 70–87 1.9
112–122 1.6 88–110 1.6 88–110 2.0
124–132 1.5 111–138 1.7 111–138 2.1
134–146 1.4 139–174 1.8 139–174 2.2
148–160 1.3 175–220 1.9 175–220 2.3
162–178 1.2 230–270 2.0 230–270 2.4
280–340 2.1 280–340 2.5
Saturation Index = 350–430 2.2 350–430 2.6
pH (actual) – (9.3 + A + B) + (C + D) 440–550 2.3 440–550 2.7
Stability Index = 560–690 2.4 560–690 2.8
2 [(9.3 + A + B) – (C + D)] – pH (actual) 700–870 2.5 700–870 2.9
880–1,000 2.6 880–1,000 3.0
Values for A, B, C, and D for LSI Calculations. Adapted from Cooling Tower Fundamentals by the Marley
Cooling Tower Company, Overland Park, KS.
Langelier, it works best when TDS is less than 800 mg/l.

Paul Puckorius has developed his Practical Scale Index (PSI), which incor-
porates temperature and conductivity into the calculations and more explicitly
takes into account the effect of alkalinity on the chemistry of the solution. The
PSI calculation, whose scale is similar to the Ryznar index, is available in slide-
rule form. A water with a PSI of greater than 6.0 is considered to be scale-dis-
solving or corrosive, and less than 6.0 to be scale-forming. Table 6-5 shows a
side-by-side comparison of the indices. Even though the RSI and PSI scale are the
same, the values obtained from the two are often not equal because the PSI cal-
culates a pH equilibrium value (pHe), which is often different than pHa.
The indices may lose applicability for analyzing waters treated with calcium

06.qxd 8/8/06 1:57 PM Page 176
Figure 6-15
LSI Nomograph. Reprinted with permission from the Betz Handbook of Industrial Water Conditioning, Ninth
Edition, BetzDearborn, Inc., Horsham, PA.
carbonate crystal modifiers, because some calcium carbonate formation is per-

mitted. However, the programs are still very useful for determining the scale or
corrosion tendencies of makeup water and are helpful for designing new systems.
More accurate programs are now available for determining calcium carbonate
scale potential. The
Table 6-5
personal computer
Scale Index Values has made these pos-
LSI RSI PSI Conditions without sible. Some water
Treatment treatment software
3.0 3.0 3.0 Extremely Severe Scaling companies sell
2.0 4.0 4.0 Very Severe Scaling s c a l e - p re d i c t i o n
1.0 5.0 5.0 Severe Scaling computer pro-
0.5 5.5 5.5 Moderate Scaling grams. The major
0.2 5.8 5.8 Slight Scaling water treatment
0.0 6.0 6.0 Stable Water chemical firms also
-0.2 6.5 6.5 Very Slightly Scale Dissolving have these comput-
-0.5 7.0 7.0 Slightly Scale Dissolving er programs and can
-1.0 8.0 8.0 Moderately Scale Dissolving readily calculate the
-2.0 9.0 9.0 Strongly Scale Dissolving scaling potential of
-3.0 10.0 10.0 Very Strongly Scale Dissolving cooling waters.

06.qxd 8/8/06 1:57 PM Page 177
Other Scales
Many other scales are possible in cooling water systems. The following are
reported guidelines for maximum levels of calcium and other anions besides
bicarbonate that can be tolerated without treatment.
Calcium phosphate. Calcium is limited by the following equation when
phosphate concentrations are 10 ppm or higher.
• CaCO3 < [105 x (9.8 – pH)]

Calcium sulfate. Calcium (ppm as CaCO3) x [SO4=] < 1,800,000 ppm.
Silica. Silica concentrations are maintained below 150 ppm to prevent for-
mation of silica deposits. The limit can sometimes be extended to 200 ppm or
higher with proper dispersants.
Magnesium silicate
Magnesium (ppm as CaCO3) x [SiO2] < 500,000 at pH of 7

Magnesium (ppm as CaCO3) x [SiO2] < 100,000 at pH of 7.5
Magnesium (ppm as CaCO3) x [SiO2] < 50,000 at pH of 8.5
The reader should note that these are only general recommended guidelines,
and as with all other guidelines in this book, each system must be evaluated indi-
vidually.
One obvious item in this list is magnesium silicate. The allowable concen-
tration greatly decreases as pH increases. Many utilities and industries are oper-
ating cooling towers at higher cycles of concentration to minimize discharge and
conserve water. However, this increases the potential for scale formation.
Scale Control
A number of scale control techniques and chemical programs are available.
These range from traditional programs such as acid feed, to treatment with com-
plex organic polymers. The range of chemicals gives the operator flexibility in
program selection.

06.qxd 8/8/06 1:57 PM Page 178
Acid Feed
For many years, the primary method for controlling calcium carbonate scale
was sulfuric acid addition. Sulfuric acid reduces bicarbonate alkalinity as follows:
H2SO4 + Ca(HCO3)2 CaSO4 + 2CO2 + 2H2O (6.18)
Although this treatment method is not as common as before, it can still be

useful for pretreating waters high in hardness and alkalinity. Before discussing
acid feed rates, it is important to examine alkalinity and its relationship in water.
Carbon dioxide, bicarbonate, and carbonate exist in equilibrium in solution,
with the amount of each dependent upon pH.
H2O + CO2 ⇔ H2CO3 ⇔ HCO3- + H+ (6.19)
HCO3- ⇔ CO3= + H+ (6.20)
At a pH of less than 4.4, only free carbon dioxide is present. Between a pH

of 4.3 and 8.3, carbon dioxide is replaced by the bicarbonate ion until at pH of
8.3 essentially no free carbon dioxide remains. Between pH 8.3 and 12, bicar-
bonate gradually converts to carbonate. At a pH of 12 or above, hydroxide begins
to form. Although pH can be used to measure the relative basicity of a solution,
the complexity and buffering effects of the carbon dioxide/bicarbonate/carbonate
series require additional measurements for precise calculation. Bicarbonate, car-
bonate, and hydroxide (for lime softened waters) levels are determined by a
series of titrations. Titration of an alkaline solution with acid to a pH of 4.3 using
a methyl orange indicator determines all of the alkalinity in solution. This is
known as the M alkalinity. Solutions that contain carbonate ion may be titrated
with phenophthalein as an indicator. The phenophthalein changes color near a
pH of 8.3. This titration determines all of the hydroxide and half of the carbon-
ate ion in solution. If one performs both a P and M titration on a sample, the
amount of bicarbonate, carbonate, and hydroxide may be determined. Table 6-6
illustrates the calculations for these determinations.
Acid feed to a cooling tower is predicated on maintaining alkalinity levels
below those which would cause calcium carbonate precipitation. It may be
tempting to closely monitor the alkalinity of a solution and control the acid feed
so that it always just keeps the pH low enough to prevent calcium carbonate for-
mation. This can be perilous because changes in temperature and makeup water
quality, failure of the acid feed pump, or poor calibration of a pH monitor could
all cause conditions to change to scale-forming. The upset might not be noticed
for some time, further exacerbating the situation. It is more common to establish
a safety margin between the actual chemistry and the point where scaling would

06.qxd 8/8/06 1:57 PM Page 179
Table 6-6
Alkalinity Relationship
to “P” and “M” Readings
Alkalinity Reading Hydroxide Carbonate Bicarbonate

(ppm as CaCO3) (ppm as CaCO3) (ppm as CaCO3) (ppm as CaCO3)
P=0 0 0 M
P=M P 0 0
2P=M 0 2P 0
2P=M 0 2P 0
2P<M 0 2P M–2P
2P>M 2P–M 2(M–P) 0
occur. By keeping the M alkalinity of the recirculating water at approximately 40

ppm, the pH can be controlled in a neutral range near 7.0. This precludes the
formation of carbonate ions and maintains a buffer below the pH (8.3) at which
carbonates would form. Supplement 6-3 outlines the calculation for determining
the acid dosage rate based on the alkalinity content of the makeup water and
desired concentration in the recirculating water.
One major problem with sulfuric acid treatment is that it lowers the pH and
increases the corrosive tendencies of the cooling water, and may increase calci-
um sulfate scaling tendencies. Upsets in the feed generate additional problems,
and many cases are known where a malfunctioning feed system has caused
excess acid introduction, severe pH depression, and corrosion of cooling system
components.
For these reasons, many utilities now operate cooling systems in an alkaline
range in which alternate chemicals are used for scale control. The alkaline range
provides a buffer against upsets. For example, a 40 ppm overfeed of H2SO4 at a
cooling water pH of 7.0 would drop the pH to 4.3. By contrast, a 40 ppm over-
feed at pH 8.0 would drop the pH to only 7.6.
Alkaline Treatment Methods

An option to remove hardness from cooling water is mechanical/chemical
pretreatment of the makeup water. These methods include clarification, ion
exchange, or membrane treatments. Ion exchange and membrane filtration are
usually impractical due to operating difficulties or the excessive amount of chem-
icals that are needed to perform the required functions. Clarification/softening
has been used for pretreatment of cooling tower makeup, especially for very large
systems or where the water is very hard. Clarification increases the pH and alka-
linity of the water, which often must be corrected with acid.

06.qxd 8/8/06 1:57 PM Page 180
Alkaline treatments have become the most popular of all cooling water pro-
grams. In an alkaline program, the pH of the cooling water is allowed to equili-
brate at a natural level, or may even be pushed into a basic pH range by addition
of alkaline compounds such as caustic soda or soda ash. In some instances, how-
ever, acid feed is still required to prevent the recirculating water from becoming
too alkaline.
Alkaline treatment programs are designed to
Figure 6-16
either keep calcium in solution or to modify the
H
O O O crystalline structure of calcium precipitates so that
HO O H they form a sludge-like product, which can be
P C P blown down. Solubilizing compounds include
HO OH
phosphonates and polymers such as polyacrylate.
CH 3
Common phosphonates include hydroxyethyli-
HEDP.
dene diphosphoic acid (HEDP, Fig. 6-16) and
aminoethylenephosphonic acid (AMP). A newer
Figure 6-17 product, phosphono-butane-tricarboxylate (PBTC)
~ CH 2 CH CH 2 CH ~ is also proving to be effective. The phosphate por-
tion of each of these compounds adsorbs onto the
COOH COOH
incipient calcium carbonate crystal nuclei, distort-
Polyacrylate.
ing its shape and slowing crystal growth.
Phosphonates may produce some negative side effects. They are corrosive to
zinc and mildly corrosive to copper and aluminum. Phosphonates are also
degraded by oxidizing biocides, particularly at biocide concentrations above 1
ppm. Not only will this cause the phosphonates to lose effectiveness, but the
breakdown products include orthophosphate, which can combine with calcium
to form calcium phosphate scale. Despite these factors, phosphonates are still the
most popular scale control agents. PBTC in particular has made an impact. It is
reported to be less corrosive than other phosphates to copper, and more resistant
to degradation from oxidizing biocides or UV light.
Polyacrylates (Fig. 6-17) also keep calcium in suspension. The polymer binds
calcium through the partial negative charges on the oxygen atoms contained in the
molecules. A co-polymer of polyacrylate and acrylamido-2-
methylpropane-sulfonic acid has proven to be a very effective dispersant.
Crystal modifiers include polymaleates and co-polymers of these with sul-
fonated polystyrenes. These polymers allow calcium deposits to form, but alter
the structure so that the deposits do not precipitate as a hard scale. The deposits
are removed in the blowdown or with a sidestream filter. The crystal modifiers
are more expensive than other treatments and are not used nearly as often.
Table 6-7 illustrates guidelines for scale inhibitor residuals. With all of these
options, it is often difficult to decide on a treatment program. However, some
trends have emerged. Blended products of phosphonates and polymers seem

06.qxd 8/8/06 1:57 PM Page 181
Table 6-7 very popular. One common

Common Scale Inhibitors blend contains HEDP and PBTC
And Dosage Levels for calcium carbonate control
Inhibitor Dosage Range (ppm) (HEDP also provides corrosion
Phosphonate 3–5 inhibition for steel), and a poly-
Polyphosphate 3–5 mer for calcium phosphate con-
Polymaleics 1–2 trol. Azoles may be added for
Sulfonated Polystyrene 1–2 copper corrosion control.
Phosphate/phosphonate treat-
ments with polymer scale dispersants still remain in use. One major water treat-
ment company has developed a new, short-chain polymer that contains no phos-
phorous. The compound, formed by polymerization of an epoxide, contains
many oxygen atoms along the polymer backbone for inhibiting calcium carbon-
ate scale. The polymer is reported to be very popular and has allowed the com-
pany to substitute it in place of phosphate-phosphonate compounds.
With these programs, or various combinations thereof, it is reported that
calcium, sulfate, and phosphate concentrations may be allowed to reach the fol-
lowing limits:
Calcium phosphate - [Ca] as CaCO3 – 1200 ppm

[PO4] – 30 ppm
Calcium sulfate - [Ca] as CaCO3 – 2700 ppm
[SO4] – 4000 ppm
LSIs as high as 2.8 to 3.0 and PSIs as low as 3.5 may be obtained with the
correct chemical program. It has also been claimed that these new polymer for-
mulations allow higher silica levels of up to 300 ppm. Some water treatment per-
sonnel view this cautiously and still abide by the old silica limit of 150 ppm.
The selection of the best cooling water treatment is dependent upon water
quality, process conditions, and system metallurgy. The choice is sometimes not
easy. Any program should be monitored closely to determine its performance.
Fouling
Fouling is the deposition of suspended solids or buildup of microbiological
organisms within heat exchangers and cooling tower fill. Foulants can be intro-
duced to a cooling water system from a variety of sources. A makeup supply from
a lake or river may contain silt and debris that is stirred up during seasonal
changes or heavy rainfall. Surface water sources also contain many microorgan-
isms that will cause microbiological buildups even in once-through cooling sys-
tems. A cooling tower is an ideal source for foulant introduction because the

06.qxd 8/8/06 1:57 PM Page 182
tower is an efficient air scrubber. Warm water, aeration, nutrients, and sunlight
transform a cooling tower into an efficient bio-reactor. The following sections
discuss fouling control methods.
Nonmicrobiological Fouling Control Methods

Both chemical and mechanical methods may be used to minimize fouling.
One practical mechanical method is sidestream filtration of recirculating water.
Sidestream filters remove silt particles which would otherwise concentrate in the
tower. Lower suspended solids reduce deposit formation in heat exchangers and
tower fill, reduce the demand for polymeric dispersants, and lower the erosion
potential of the water towards copper-alloy tubes.
Usually, only a relatively small percentage (1% to 5% of the recirculating
flow rate) must be filtered, because the entire volume of recirculating water will
pass through the filter several times every day. For example, consider the
100,000 GPM recirculating system mentioned earlier. Assume system volume is
300,000 gallons. A sidestream softener operating at 3% of the recirculating
capacity will filter 3000 GPM. This equates to 4,320,000 million gallons per day,
or 14.4 system volumes daily. Supplement 6-4 outlines the calculations needed
to size a sidestream filter.
Sidestream filters may be of the multimedia or cartridge filter type. One
additional benefit of sidestream filters is that they remove iron oxide particles
that may have been generated by corrosion of the cooling water lines.
Another on-line mechanical cleaning method for condenser tubes involves
the use of sponge balls that are introduced at the condenser inlet and collected
by a grate system at the condenser outlet. The balls are sized at a slightly larger
diameter than the condenser tubes so that they will scrub the entire tube surface
as they pass through. These balls are excellent for removing soft deposits like bio-
logical slime. Given the increasingly stringent restrictions being placed on chlo-
rine and oxiding biocides, these systems merit consideration. However, some dif-
ficulties are inherent with on-line cleaning systems. The most obvious problem
is retrofitting such a system on an existing condenser, for this involves installa-
tion of a collection device on the condenser outlet and recirculation line to the
condenser inlet. Capital costs may be prohibitive, and operating difficulties may
also occur. A malfunctioning collection system will allow the balls to escape. It
has been reported that at one facility the scale inhibitor treatment program
caused the balls to clump into sticky masses, although I would suspect that this
was due to gross overfeed of the inhibitor or some other abnormal condition.
A common method for cleaning condenser tubes off-line is mechanical
scraping. In this process, maintenance personnel insert mechanical scrapers sim-
ilar to the type shown in Figure 6-18 into the tubes and then blow the scrapers

06.qxd 8/8/06 1:57 PM Page 183
Figure 6-18 through with water or com-

TUBE WALL pressed air. The scrapers are
equipped with either metallic or
PRESSURE
SCRAPER < plastic fins to dislodge deposits.
Although this procedure may
take several days, it does provide
Outline of Mechanical Tube Scraper. the desired results. (I personally
had the opportunity to work on a number of these projects, most often for
removal of microbiological deposits but also for calcium carbonate deposits.
Tube scraping always returned the condenser to near-optimum performance, as
based on data provided by the cleanliness factor program outlined in Supplement
2–2.) The advantage of mechanical scraping versus on-line cleaning is much
lower capital cost. However, mechanical scraping obviously requires an outage.
The condenser may operate at reduced efficiency for a long time before an out-
age can be scheduled.
While mechanical methods can be effective for control of foulants, their pro-
hibitive cost or on-line limitations may make them impractical. Chemical treat-
ment is sometimes the best or only option.
Foulant control chemicals are often similar or identical to scale control
compounds. The most notable foulant control chemicals include low molecular
weight polyacrylates, polyacrylamide, polymaleic-anhydride, and copolymers
containing monomers of these compounds combined with specialized
monomers. These dispersants keep foulants suspended in solution by reinforc-
ing their negative surface charge,
Table 6-8
causing them to repel each other.
Common Fouling Inhibitors
Table 6-8 lists typical dosages and Dosage Levels
for foulant treatments. Foulant con-
trol, along with scale control, is gain- Inhibitor Dosage Range (ppm)
ing importance due to the increasing Polyacrylate 4–5
use of high-efficiency, close-spaced Polyacrylamide 0.2–0.5
film fill in cooling towers.
Microbiological Fouling
Microbiological fouling may be the worst problem in cooling water systems.
Microbes can:
• Generate acids and corrodents that attack the underlying tube metal
• Secrete a protective gelatinous layer that coats the tube surface. The secre-
tions, along with silt that becomes trapped within, reduce heat transfer
and cause under-deposit corrosion.

06.qxd 8/8/06 1:57 PM Page 184
• Be pathogenic. In the most often-cited case, 34 American Legion members

died in 1976 of complications caused by the bacteria Legionella pneu-
mophilia, which was present in a hotel cooling system.
Three types of microorganisms may exist in a cooling water system, algae,

bacteria, and fungi. Algae contain chlorophyll and require sunlight to thrive.
These organisms proliferate in wetted areas of the tower exposed to the sun.
Fungi, including mold and yeast, are nonphotosynthetic organisms, which do
not need sunlight for survival. They are most notable for attacking either the cel-
lulose or lignin components of wood, where they induce rot. Bacteria also do not
need sunlight for survival and may thrive in many areas, particularly in con-
densers where the water is warm. Bacterial counts of greater than 10,000,000 per
milliliter are possible in systems with insufficient biocide feed.
Control of Microbiological Organisms

With the exception of condenser tube cleaning, it is virtually impossible to
mechanically control microbiological organisms in large cooling systems.
Therefore, chemical feed of a biocide is required. (Ultraviolet light and site-gen-
erated ozone have been successfully used for small cooling tower and makeup
system disinfection.) Chemical treatment can be split into two basic categories,
oxidizing (chlorine, bromine, chlorine dioxide) and nonoxidizing. Oxidizing
chemicals attack cell components to kill the organism, while nonoxidizing bio-
cides may damage the cell wall or interfere with the cell’s metabolic processes.
Environmental and safety restrictions on the use of biocides are causing sig-
nificant changes in biocide treatment. While the following sections provide a dis-
cussion on the properties and use of the most common biocides, they are also
intended to give the reader an idea of the current trends in microbiological control.
Chlorine
For many years, chlorine was the primary disinfectant for all types of water
systems. Its use has fallen into disfavor in some industries because of safety issues
related to gaseous chlorine and because chlorine is known to react with organics
to produce halogenated organic compounds.
Chlorine gas is the least expensive of all oxidizing biocides. When chlorine
is added to water, it immediately reacts as follows:
Cl2 + H2O HOCl + HCl (6.21)
HOCl is the active ingredient of this mixture, and it attacks organisms very
quickly. It has been calculated that the kill rate of a 1.0 ppm chlorine solution is

06.qxd 8/8/06 1:57 PM Page 185
99% in 30 seconds at a pH of 6.5. The key to chlorine effectiveness is pH. The

stability of HOCl is dependent upon it, as the following reaction shows:
HOCl H+ + OCl- (6.22)
OCl- is much less potent than HOCl. Some scientists theorize that OCl-,
being a charged ion, has much more difficulty penetrating the cell wall and
attacking cell components. The dissociation of HOCl begins at a pH of about 5.2,
reaches 50% at a pH of 7.5, and is fully complete at a pH of 9.4. With the grow-
ing trend towards alkaline cooling water treatment programs, other biocides are
proving to be more effective.
Application of chlorine. Chlorine gas is very toxic and must be handled
with great care. A common method of introducing it into water is through an
eductor system, in which a sidestream of flowing water is used to pull the chlo-
rine gas from the chlorine cylinder into the stream. Chlorine is commonly
shipped in one-ton cylinders, which can be loaded in multiple units into a rack
containing a manifold system for quick transfer between full and empty contain-
ers. Chlorine systems are now required to have ambient air monitoring and alarm
systems that provide safety warnings in the event of a chlorine leak. The maxi-
mum allowable ambient air chlorine limit as established by OSHA is 1.0 mg/l.
Chlorine reacts irreversibly with a number of constituents in water, most
notably ammonia and organics. These reactions, which reduce the amount of free
chlorine available as a biocide, are known as the chlorine demand. The higher
the chlorine demand, the more chlorine that must be added to achieve the same
level of killing effectiveness. Ideally, a 0.1 to 0.5 ppm free chlorine residual is
most effective in controlling microorganisms, while minimizing degradation of
cooling water treatment chemicals or cooling tower materials.
Environmental aspects of chlorine usage are becoming more complicated.
The “technology-based limits” established by EPA for power plants essentially
limited the chlorine residual in the cooling system discharge to 0.2 ppm for a
maximum of two hours per day. These limits are gradually giving way to “ambi-
ent water quality standards” to be determined in the stream at the boundary of a
calculated mixing zone. The new limit is 0.019 mg/l in fresh water for intermit-
tent applications. Individual states may set more stringent limits, and often apply
the limits at “end-of-pipe” instead of allowing for dilution in the mixing zone.
These concentrations are much too low to be effective. A method to main-
tain higher free chlorine residuals is to dehalogenate the discharge with sodium
sulfite (Na2SO3), sodium bisulfite (NaHSO3), or sulfur dioxide gas. These react
with chlorine in a one-to-one stoichiometric fashion:
Cl2 + Na2SO3 + H2O Na2SO4 + 2HCl (6.23)

06.qxd 8/8/06 1:57 PM Page 186
An excess of sulfite is added to drive the reaction to completion. Continuous

halogenation of cooling water influent with subsequent dehalogenation of the
effluent before discharge is a popular macrofouling control technique, particu-
larly at utilities and industries that have to deal with zebra mussels. This is dis-
cussed in more detail later.
There is no question that chlorine has saved countless lives from water-
borne diseases since its introduction as a disinfectant. However, with the devel-
opment of other disinfectants, further restrictions on chlorine use for water treat-
ment may be in the offing. Some environmental groups are calling for a total ban
on chlorine due to health concerns. With regard to cooling water treatment
specifically, chlorine use is being questioned because of the halogenated organics
issue.
Liquid and solid chlorine donors. Where chlorine is still the preferred bio-
cide but safety is of concern, liquid sodium hypochlorite (NaOCl) is an alterna-
tive. Sodium hypochlorite is more expensive than gaseous chlorine, but is much
less hazardous to handle, can be fed directly to the cooling water, and can be
stored in a bulk tank. The percentage of sodium hypochlorite in bulk industrial
solutions ranges from about 10 to 15%. Sodium hypochlorite will decompose
with time, increasing temperature, or in the presence of impurities, particularly
iron and copper. Therefore, it is important to design and specify a system that
minimizes hypochlorite decomposition. Supplement 6-5 outlines a list of speci-
fications and guidelines for ordering and storing sodium hypochlorite.
Sodium hypochlorite, which is the sodium salt of hypochlorous acid, under-
goes the same reactions as chlorine does in water and can produce the same com-
pounds, including halogenated organics. Sodium hypochlorite and solid chlorine
compounds are nowadays viewed somewhat less favorably because of this aspect.
Several solid chlorine compounds are available. Calcium hypochlorite
[Ca(OCl)2], the common swimming pool biocide] is one. More popular for the
industrial water treatment industry are the dichlorohydantoins (Fig. 6-19) and
the chlorinated isocyanuarates (Fig. 6-20). The hydantoins may be manufactured
into powder, granules, pellets, or tablets, which can then be loaded into a dis-
solving vessel that takes a sidestream of water
Figure 6-19 from the cooling loop. The products are
Cl designed to dissolve at a relatively slow, con-
CH 3 stant rate to release chlorine gradually. The iso-
N cyanurates provide an added benefit by helping
CH 3 C
C O to keep chlorine in solution. However, the
C buildup of isocyanuric acid in recirculating sys-
N
O
tems can eventually limit the effectiveness of the
biocide.
Cl
Dichlorohydantoin. Chlorine alternatives. The proliferation

06.qxd 8/8/06 1:57 PM Page 187
Figure 6-20 of alkaline cooling water treatment programs

and the issue of chlorinated organics has caused
Cl O
a steady decline in the use of chlorine and an
N C increase in alternate treatment methods.
O C N Cl Bromine. A popular alternative is bromine.
Bromine, which is also a halogen, is not intro-
N C
duced directly to a cooling water system, but
Cl O rather is generated on-site by reacting a bromide
Chlorinated Isocyanurate. solution with chlorine in the following step-
wise procedure:
Cl2 + H2O HOCl + HCl (6.21)

HOCl + NaBr NaCl + HOBr (6.24)
Bromine can also be generated by reaction of sodium bromide with

hypochlorite.
Bromine offers several advantages over chlorine, particularly in alkaline
waters. First, as Figure 6-21 illustrates, the dissociation of HOBr to OBr- and H+
takes place at a higher pH than the dissociation of HOCl to OCl- and H+ (Eq.
6.22). Thus, bromine’s biocidal efficiency in basic solution is greater. Secondly,
although bromine reacts with ammonia and amines, the reaction is reversible.
Unlike chlorine, bromine introduced to ammonia-laden waters still exhibits free-
halogen properties.
Because bromine is similar to chlorine in biocidal efficiency, and better in
Figure 6-21
Bromine and Chlorine Dissociation Chart. Reprinted with permission from the Betz Handbook of Industrial
Water Conditioning, Ninth Edition, BetzDearborn, Inc., Horsham, PA.

06.qxd 8/8/06 1:57 PM Page 188
alkaline waters, residual concentrations are

Figure 6-22
Br
maintained at similar levels (0.1 to 0.4 ppm).
Bromine may be generated with either chlo-
CH 3
N rine gas or liquid hypochlorite. This is usually
CH 3 C carried out with an eductor system that draws
C O
both the chlorine and sodium bromide into a
C sidestream where the reaction takes place. Many
N
O of the major water treatment firms have devel-
oped bromine generators. Bromine is also sup-
Cl
plied in solid form as hydantoins (Fig. 6-22),
Bromo-Chlorohydantoin
which are introduced to cooling water in the
same fashion as the chlorohydantoins mentioned above.
Bromine will also form halogenated organics. Although these compounds
are believed to be less persistent in the environment than chlorinated organics,
they have caused some to view bromine treatment with concern.
Chlorine dioxide. Chlorine dioxide (ClO2) offers several advantages. These
include:
• Powerful oxidizer
• Not pH sensitive like HOCl
• Does not form halogenated organics
• Does not react with ammonia
The primary drawback is that chlorine dioxide must be generated on-site

because it is much too unstable to be transported. It is also quite a bit more
expensive than chlorine or bromine, and since it does not hydrolyze but remains
as a dissolved gas, is readily aerated out of solution as it passes through a cool-
ing tower.
Two principal methods of ClO2 production are presently available, chlo-
rine/hypochlorite generation and acid activation of sodium chlorite. The main
ingredient in both is sodium chlorite (NaClO2), which is generally supplied as a
25% solution.
2NaClO2 + Cl2 2ClO2 + 2NaCl (6.25a)

2NaClO2 + NaOCl + 2HCl 2ClO2 + 3NaCl + H2O (6.25b)
5NaClO2 + 4HCl 4ClO2 + 5NaCl + 2H2O (6.26)
Of these production methods, the first two have been the most popular for
large cooling systems. However, the latter method offers an advantage because no
chlorine or hypochlorite is needed. Similar to gaseous chlorine or bromine treat-
ment, an eductor system is employed to mix the sodium chlorite with the acti-

06.qxd 8/8/06 1:57 PM Page 189
vating agents. The sodium chlorite may be stored

Figure 6-23
in a bulk tank for feed to the system.
H O
Chlorine dioxide residuals similar to those
C C
established for chlorine are effective in killing
C N CH 3 microorganisms. In fact, chlorine dioxide may be
H
S
even more potent than chlorine, in part because
General Isothiozolin Structure.
it is not affected by pH as is hypochlorous acid.
(Even chlorine dioxide is not efficient if treatment
Figure 6-24 is not started ahead of the microbiological grow-
ing season. See Case History 6-5.)
R
As with the other oxidizing biocides, safety
+
R N R is very important while using chlorine dioxide.
R Sodium chlorite is a powerful oxidizer and will
R = Organic Chain react with carbon-based compounds. Sodium
Quaternary Amine. chlorite is generally safe in its liquid form, but
any spills should be promptly removed because
Figure 6-25
NaClO2 becomes very unstable when it dries.
H Br H Even friction will ignite a dried puddle of sodium
chlorite.
HO C C C OH
Chlorine dioxide can be an excellent bio-
H NO2 N cide, but its high cost and tendency to come out
BNPD.
of solution in cooling towers restrict its useful-
ness.
Ozone. Ozone use has increased in recent years, and the technology has
matured enough that it is being used for treatment of some smaller, industrial
recirculating systems. Ozone (O3) is a generally short-lived, extremely powerful
oxidant that must be generated on-site. It is produced by passing an oxygen or
air stream through a high-voltage current. The air stream is then bubbled into the
recirculating water.
The primary advantages of ozone are that it is a much more powerful oxi-
dizer than even chlorine dioxide and that it does not produce any harmful reac-
tants such as halogenated organics. At this time, ozone is not extensively used for
treatment of utility or large industrial cooling towers. In part this is due to the
fact that cooling water often has enough organic material to create a high ozone
demand that uses up much of the biocide before it can attack microorganisms.
Work is continuing in this field.
Nonoxidizing biocides. Due to the safety and environmental concerns
regarding oxidizing biocides, nonoxidizing biocides have received more atten-
tion. Sometimes the nonoxidizers are used in a joint program with oxidizing
reagents to attack microorganisms.
A number of nonoxidizing biocides are currently available. Some of the most

06.qxd 8/8/06 1:57 PM Page 190
popular being used today include isothiozolone (Fig. 6-23), quaternary amines
(Fig. 6-24), and bromonitropropanediol (BNPD, Fig. 6-25). These biocides work
by either reacting with the cell wall or interfering with the organism’s metabolic
processes. A cell structure is shown in Figure 6-26. Quaternary amines penetrate
into cell walls and disrupt transport of products through the cell wall. BNPD and
isothiozolone react with groups inside the cell or interfere with protein making
processes.
Microorganisms can develop a resistance to the nonoxidizing biocides, so
careful planning may be needed for them to be effective. One method is to use
them as a supplement to oxidizing biocides. A periodic batch dosage can be used
to shock the microbes and kill those which may have survived the oxidizing
chemical. Or, two nonoxidizers could be used on a more continuous basis, but
be alternated periodically (perhaps even daily) so that the microorganisms do not
build up any tolerance.
Just like oxidizing biocides, nonoxidizers can have an effect on the environ-
ment if they leave the plant in the cooling discharge. Various treatment methods
are possible to neutralize many of these chemicals. For instance, some of the
non-oxidizers, isothiozolone and BNPD in particular, are decomposed with sul-
fite. Bentonite clay, if added to the discharge, provides absorption sites for com-
pounds such as the quaternary amines.
Biocide Discharge Limits—Future Trends

Much has been said in the preceding sections about cooling water treatment
chemicals and the restrictions being placed on their discharge to open bodies of
water. These restrictions may become more stringent. Where this will end is any-
one’s guess. Some articles have suggested that the EPA may limit some chemical
concentrations to the lowest analytical detection capability. Rumors also persist
that once-through systems will have to be converted to recirculating systems.
Only time will tell.
Macrofouling
Fouling caused by organisms whose individual members are visible to the
naked eye is known as macrofouling. Macrofouling may be caused by a number
of different creatures, but the two most troublesome fresh-water species by far
have been Asiatic clams and even more importantly zebra mussels.
Macrofouling by clams occurs when the creature dies and the shell is
washed into sensitive flow areas such as a condenser. Asiatic clams proved trou-
blesome because shell sizes typically ranged from about 1/2 inch to 1 inch in size.
Since condenser and heat exchanger tubes often range from 1/2 inch to 1 inch in

06.qxd 8/8/06 1:57 PM Page 191
Figure 6-26
Cell Structure. Illustration by Alyssa Buecker.
diameter, the clam shells almost perfectly fit. A growth in Asiatic clams within a
cooling system could play havoc with operation.
This, however, seems almost minor compared to some of the horrors that
plant personnel have experienced in dealing with zebra mussels.
Zebra Mussels
These creatures were inadvertently introduced to the Great Lakes in the
mid-1980s. They are so called because their shell is alternately patterned with
light and dark stripes. Zebra mussels are freshwater bivalves that may grow to
about 5 cm in length.
Zebra mussels are extraordinarily troublesome, primarily because of their
colonization patterns. Zebra mussels are microscopic when first spawned, being
about 40 microns in length. They also have no shell at first. Within just a few
weeks after hatching, however, the mussels have become mature enough to look
for a place to settle. When the mussel has found a location, it extends fibers,
known as byssal threads, to attach permanently to the surface. Zebra mussels are

06.qxd 8/8/06 1:57 PM Page 192
troublesome primarily for four reasons:
• The microscopic larvae (veligers) can be carried a long way by water cur-
rents.
• The veligers are much too small to be stopped by traveling screens at
water intakes.
• They like locations of flowing water, such as cooling water intakes, where
nutrient supplies are plentiful. However, if water flow rates are higher
than 3 or 4 feet per second they may not settle.
• They will attach to many different hard surfaces, including each other, to
form dense, thick masses.
This latter aspect is what has made zebra mussels particularly annoying.
Densities of up to 80,000 creatures per square foot have been found at locations
in the Great Lakes. The mussels can clog trash racks, traveling screens, auxiliary
cooling water lines, and fire lines. Masses that break loose in the main cooling
water line will plug condenser tubes.
Various control methods have been investigated for control of these creatures.
Some of them include electrification of trash racks, application of fouling or slick
coatings to intake pipes, and even use of acoustics. None have yet been adopted on
an industrial level. More effective have been chemical treatment and thermal shock.
Chemical treatment. If zebra mussels infest a cooling water system, it is
usually because the veligers have entered and then matured within the system.
Plant personnel can combat this by diligently operating the biocide system, even
though feed may not be continuous. Infant zebra mussels have no shell, and it
takes several days for one to develop. Regular oxidant feed to the cooling water
will kill some of the veligers before the shell fully develops.
To kill mature mussels takes a more vigorous effort. Adult mussels can sense
oxidizing biocides and will simply close up if they detect a residual oxidant. They
can stay closed for at least several days, and can certainly outwit an intermittent
oxidant feed. One method is to initiate continuous oxidant feed, and dehalo-
genate the discharge. Eventually, the mussel will have to open, at which time the
oxidant can kill it. Such a treatment can only be used if the plant’s environmen-
tal permit allows for it. However, the EPA has granted variances for this type of
application to facilities threatened by zebra mussels.
An alternative is feed of a nonoxidizing biocide. Mussels cannot detect these
chemicals and will filter the water unwittingly. Nonoxidizers damage zebra mus-
sel cells just as they do microbiological cells. Quaternary amines have proven to
be effective. Again, however, chemical feed must be approved by the environ-
mental protection governing board that oversees the plant.
Thermal shock. Zebra mussels are not very tolerant to really warm water

06.qxd 8/8/06 1:57 PM Page 193
(temperatures above 100˚F), therefore, it is possible to destroy mussels by ther-

mal shock. This method is possible at facilities in which warmed cooling water
from the condenser can be recirculated to the cooling water intake. The key is to
do this in winter, when the mussels have become acclimated to cold water. Good
results have been obtained when mussels acclimated to 50˚F water were sud-
denly subjected to 100˚F water. Since many utilities and industrial plants do not
have this capability, thermal shock is not a universal treatment.
In once-through systems, the major colonization sites include the intake
structure and low flow areas of auxiliary water systems. Since the water velocity
in most pressurized cooling water systems is 7 feet per second, the mussels do
not colonize in most locations of constant flow. A considerable measure of pro-
tection can be obtained by shutting down one intake pump (at low load periods)
and increasing the temperature of the remaining circulating flow with steam from
a portable generator. Alternatively, the circulating water can be dosed with a
nonoxidizing molluscicide for a few hours. These treatments are economical with
minimal environmental impact.
Conclusion
Cooling water treatment is often a complex issue, and is not being made any
easier by environmental regulations. Certainly chlorine as a biocide and metals
(zinc and chromate) for corrosion control have lost popularity. Alkaline cooling
water treatment programs appear to be the trend. Plant personnel must careful-
ly consider any treatment program and take into account such factors as econo-
my, safety, efficiency, and environmental regulations.

06.qxd 8/8/06 1:57 PM Page 194

06.qxd 8/8/06 1:57 PM Page 195
Supplement 6-1
Cooling Tower BASIC Program
The cooling tower BASIC program asks for four variables, recirculating
water flow rate, temperature difference between the warm return water and
cooled water in the tower basin, the evaporation correction factor, and cycles of
concentration. The program will display values for evaporation, blowdown, drift,
and makeup.
CTALCS.BAS
10 CLS:LOCATE 5,1
20 PRINT “COOLING TOWER CALCULATION PROGRAM:
30 PRINT “ENTER THE RECIRCULATING FLOW RATE IN GPM:
50 INPUT R
60 PRINT
70 PRINT “ENTER THE TEMPERATURE RANGE BETWEEN THE WARM RETURN WATER”
80 PRINT “AND THE COOLED WATER IN THE TOWER BASIN (DEGREES F):
09 INPUT DELTAT
100 PRINT
110 PRINT “ENTER THE CORRECTION FACTOR (f) FOR EVAPORATION”

06.qxd 8/8/06 1:57 PM Page 196
120 PRINT “f=0.75 (average)”

130 PRINT “f=0.85 (warm weather operation)”
140 PRINT “f=0.65 (cold weather operation)”
150 INPUT F
160 PRINT
170 PRINT “ENTER THE CYCLES OF CONCENTRATION”
180 INPUT C
190 PRINT
200 PRINT “ENTER THE TOWER MANUFACTURER’S ESTIMATED PERCENT DRIFT”
210 INPUT DPERCENT
220 PRINT
230 E=(F*R*DELTAT)/100
240 DGPM=(DPERCENT*R)/100
250 BD=(E/(C-1))-DGPM
260 MU=E+BD+DGPM
270 PRINT “THE WATER BALANCE AROUND THE TOWER IS:”
280 PRINT
290 PRINT “EVAPORATION = “;:PRINT USING “####”;E;:PRINT “ GPM”
300 PRINT
310 PRINT “BLOWDOWN = “;:PRINT USING “####”;BD;:PRINT “ GPM”
320 PRINT
330 PRINT “DRIFT = “;:PRINT USING “##”;DGPM;:PRINT “ GPM”
340 PRINT
350 PRINT “MAKEUP = “;:PRINT USING “####”;MU;:PRINT “ GPM”
Supplement 6-2
LSI and RSI BASIC Program
The program is based on the values for A, B, C, and D found in Table 6–4.
The user must input this data. The program calculates reasonably accurate val-
ues for both the LSI and RSI in waters that are not too highly concentrated. (TDS
below about 800 ppm.) More sophisticated programs, which take into account
ionic interactions, must be used to calculate the scaling potential of more highly
concentrated waters.
LSI.BAS
10 CLS:LOCATE 5,1
20 PRINT “THIS PROGRAM WILL CALCULATE THE LANGELIER SATURATION INDEX”
30 PRINT “AND AYZNAR INDEX FOR WATER.”
40 PRINT
50 PRINT “ENTER THE MEASURED pH”
60 INPUT PHA
70 PRINT
80 PRINT “ENTER THE TOTAL DISSOLVED SOLIDS AS PPM”
90 INPUT TDS
100 PRINT
110 PRINT “ENTER THE WATER TEMPERATURE AS DEGREES F”
120 INPUT T

06.qxd 8/8/06 1:57 PM Page 197
130 PRINT
140 PRINT “ENTER THE CALCIUM HARDNESS AS PPM CaCO3”
150 INPUT CAH
160 PRINT
170 PRINT “ENTER THE ALKALINITY AS PPM CaCO3”
180 INPUT ALK
190 GOSUB
200 PRINT
210 PRINT
220 PRINT “*********************************************”
230 PRINT “THE LSI = “;:PRINT USING “##.##”;LSI
240 PRINT “THE AYZNAR INDEX = “;:PRINT USING “##.##”;RSI
250 PRINT “*********************************************”
260 PRINT
270 INPUT “ANOTHER ANALYSIS (Y/N)”;ANS$
280 IF ANS$=”Y” GOTO 10
290 IF ANS$=”y” GOTO 10
300 CLS:LOCATE 1,1:SYSTEM
310 IF TDS>=SO AND TDS<=3OO THEN A=.1
320 IF TDS>300 AND TDS<400 THEN A=(.001*TDS)-.2
330 IF TDS>=400 AND TDS<=1000 THEN A=.2
340 IF T >=32 AND T<=34 THEN B=2.6
350 IF T=35 THEN B=2.5S
360 IF T>=36 AND T<=42 THEN B=2.5
370 IF T=43 THEN B=2.45
380 IF T>= 44 AND T<=48 THEN B=Z.4196
390 IF T=49 THEN B=2.35
400 IF T>=SO AND T<=56 THEN B=2.3
410 IF T=57 THEN B=2.25
420 IF T>=58 AND T<=62 THEN B=2.2
430 IF T=63 THEN B=2. 15
440 IF T>=64 AND T<=70 THEN B=2.1
450 IF T=71 THEN B=2.05
460 IF T>=72 AND T<=80 THEN B=2
470 IF T=81 THEN B=1.95
480 IF T>=82 AND T<=88 THEN B=1.9
490 IF T=89 THEN B=1.85
500 IF T>=90 AND T<=98 THEN B=1.8
510 IF T=99 THEN B=1.75
520 IF T >= 100 AND T<=110 THEN B=1.7
530 IF T= 111 THEN B=1.65
540 IF T>=112 AND T<=12Z THEN B=1.6
5S0 IF T= 123 THEN B=1.55
560 IF T>= 124 AND T<=132 THEN B=1.5
570 IF T=133 THEN B=1.45
580 IF T>=134 AND T<=146 THEN B=1.4
590 IF T=147 THEN B=1.35
600 IF T>=1 48 AND T<=160 THEN B=1.3
610 IF T=161 THEN B=1.25
620 IF T>161 THEN B=1.2
630 IF CAH>=10 AND CAH<=11 THEN C=.6
670 IF CAH>=23 AND CAH<=27 THEN C=1
680 IF CAH>=28 AND CAH<=34 THEN C=1.1

06.qxd 8/8/06 1:57 PM Page 198

750 IF CAH>=1 39 AND CAH<=174 THEN C=1.8
770 IF CAH>225 AND CAH<=275 THEN C=2
780 IF CAH>275 AND CAH<=345 THEN C=2.1
840 IF ALK>=10 AND ALK<=11 THEN D=1
850 IF ALK>=12 AND ALK<=13 THEN D=1.1
880 IF ALK>=Z3 AND ALK<=27 THEN D=1.4
940 IF ALK>=88 AND ALK<=11O THEN D=2
980 IF ALK>225 AND ALK<=275 THEN D=2.4
1040 IF ALK>a75 AND ALK<=1000 THEN D=3
1050 LSI=PHA-(9.3+A+B)+(C+D)
1060 RSI=2*((9.3+A+B)-(C+D))-PHA
1070 RETURN
Supplement 6-3
Acid Feed Calculations
Alkalinity like all other ions will cycle up in a cooling tower unless it is con-
trolled. Sulfuric acid converts bicarbonates and carbonates to carbon dioxide,
which escapes from the tower. The amount of acid needed for conversion is
dependent upon the alkalinity in the makeup water and the desired concentra-
tion in the tower. Water treatment experts recommend an alkalinity of 20 to 40
ppm in the recirculating water to provide safe operation. Consider an example
where the makeup alkalinity is 100 ppm (as CaCO3), the cycles of concentration
are 6, and the desired alkalinity in the recirculating water is 30 ppm. The alka-
linity in the makeup would have to be lowered from 100 to 5 ppm.
The accompanying BASIC program will calculate the amount of acid need-
ed for alkalinity reduction. The program asks for makeup flow rate, alkalinity in

06.qxd 8/8/06 1:57 PM Page 199
the makeup water, desired alkalinity in the recirculating water, and cycles of con-
centration. The calculations are made more simple by the fact that alkalinity is
routinely measured and reported as ppm CaCO3. (The molecular weight of
CaCO3 is 100, this is why it is so often used as a standard for water analysis cal-
culations.) Sulfuric acid has a molecular weight of 98 and an equivalency of 2,
so for most practical calculations 1 ppm of sulfuric acid can be considered to
react with 1 ppm of alkalinity as CaCO3. The BASIC program assumes this ratio
and does not correct for the slight difference in molecular weights.
ADICDALC.BAS
10 CLS:LOCATE 5,1
20 PRINT “ACID FEED CALCULATION PROGRAM.”
30 PRINT
40 PRINT “ENTER THE MAKEUP FLOW RATE IN GALLONS PER MINUTE”
50 INPUT GPM
60 GPD=GPM*1440
70 PRINT
80 PRINT “ENTER THE ALKALINITY OF THE MAKEUP WATER”
90 INPUT MUALK
100 PRINT
110 PRINT “ENTER THE DESIRED ALKALINITY IN THE RECIRCULATING WATER”
120 RWALK
130
140 PRINT “ENTER THE CYCLES OF CONCENTRATION”
150 INPUT C
160 MUALKD=RWALK/C
170 MINUSALK=MUALK-MUALKD
180 PRINT
190 LBACID = (GPD*MINUSALK)/(1000*120)
200 PRINT “THE AMOUNT OF SULFURIC ACID REQUIRED PER DAY = “;:PRINT USING
“#####”
;LBACID;:PRINT “ POUNDS”
Supplement 6-4
Sizing a Sidestream Filter
A simple set of equations is available to size sidestream filters. The calcula-
tions are based on percent solids reduction and the cooling tower blowdown
rate. As an example, consider a recirculating water that contains 20 ppm of sus-
pended solids. The desirable concentration is 5 ppm. Blowdown from the tower
is 100 GPM.
First calculate the percent solids reduction, where:
• %Solids Reduction (SR) = (Si - Sf) / Si x 100

06.qxd 8/8/06 1:57 PM Page 200
• %SR = 20 ppm – 5 ppm) / 20 ppm x 100 = 75%
Determine the filtration rate by the following equation:
• %Filtration Rate (FR) = 100

– 1 x BD
100 – SR
• %FR = [(100/(100-75)-1)] x 100 GPM = 300 GPM
Supplement 6-5
Specifications for Sodium Hypochlorite
Sodium hypochlorite will decompose into oxygen, sodium chloride, and
sodium chlorate. The decomposition rate is affected by temperature and by the
action of metals, most notably iron and copper. A specification should contain
the following:
• Iron < 0.5 mg/l

• Copper < 1.0 mg/l
• pH – 11.0 to 11.2
Temperature can have a dramatic impact on hypochlorite decomposition.

For example, the half-life of a hypochlorite solution is reported to be 800 days at
a temperature of 59˚F. At 77˚F, the half-life drops to 220 days, and at 140˚F, the
half-life is only 3 days! Bulk storage tanks of sodium hypochlorite should be kept
as cool as possible, by sun-shading, painting them white, or both.
A recommended material of construction for bulk storage tanks is fiberglass-
reinforced plastic (FRP). Polyethylene was once the material of choice, but prob-
lems have been reported with cracking of the material.
Case History 6-1

Conditions: Once-through cooling to three condensers
Source - large recreational lake
Scaling had never been a serious problem in the condensers, which had 20,
15, and 10 years of operation, respectively. In the summer of 1988, the drought

06.qxd 8/8/06 1:57 PM Page 201
that afflicted most of the Midwest caused the lake level to drop dramatically.
Plant chemists calculated that the concentration of dissolved solids in the lake
water increased to four times the normal amount. However, they and other plant
personnel did not give thought to the possibility of scale formation, as none had
occurred before. During the summer, performance in the 10-year-old condenser
declined slowly but noticeably. When the unit came off line for an autumn out-
age, an inspection team found that the waterside of the tubes was completely
covered with a layer of scale, less than one millimeter in thickness. The scale was
primarily composed of calcium carbonate. The scale was the direct result of the
drought and the concentration of solids in the lake. The plant staff hired a con-
tractor to mechanically scrape the tubes. The problem has not reoccurred since.
An interesting aspect about this situation is that the scale developed in this one
condenser, which is tubed with copper-nickel alloy. No scale formed in the other
condensers, which are tubed with Admiralty metal. Because outlet cooling water
temperatures in the other condensers were at least as high as those in the scaled
condenser, the scaling may have also been influenced by metallurgy. No investi-
gation was conducted into this possibility, however.
Case History 6-2

Conditions: Admiralty-tubed condenser
Once-through cooling
A number of the condenser tubes had failed by long-term, steam-side,
ammonia gouging. Plant personnel decided to replace the tubes in the affected
zones with 90-10 copper nickel. This material had performed well in other con-
densers at the plant.
Within two years, the new copper-nickel tubes began to fail. When a tube
was pulled for examination, numerous pits were observable on the water-side of
the tube. One half-section sample, cut to one-foot in length, contained eight
through-wall failures. The corrosion was traced to the manufacturing process.
The manufacturer used a lubricant containing sulfides. The lubricant was not
removed before the tubes were placed in service.
This example illustrates the extreme affect that sulfides have on copper alloys.

06.qxd 8/8/06 1:57 PM Page 202
Case History 6-3

Conditions: Closed cooling water system
Cooling water for automated welders
The closed cooling water system had been placed on a treatment program
that included sodium nitrite for corrosion inhibition. After implementation of the
program, some of the small-diameter cooling water tubes within the welder
heads began to plug. When the system was opened for inspection, investigators
found significant microbiological growth within the system. They concluded that
the nitrite was acting as a nutrient to microbes that had entered the water. The
treatment was changed to a molybdate-based program with periodic feed of a
nonoxidizing biocide. Since the changeover, the plugging problems within the
welder heads have disappeared.
Case History 6-4

Conditions Six separate closed cooling water systems
Provide cooling water for pump bearings and circulating water
for building heat
Each of the systems had been placed on a treatment program that consisted
solely of sodium nitrite for corrosion inhibition. The sodium nitrite was added in
dry form to the systems through pot feeders. Unlike the previous case history,
microbiological fouling was not a problem. However, in several of the systems
the recommended minimum limit of nitrite could not be maintained for even
short periods of time. This was caused by leaks, which required the constant
addition of fresh makeup water. Failure to maintain sufficient inhibitor levels can
result in severe localized corrosion. Plan personnel suspended nitrite feed to
those systems in which the residual could not be maintained at sufficient levels.
To my knowledge, the systems have not failed in the subsequent 14 years of ser-
vice.
Case History 6-5

Conditions: One-through cooling to three condensers
Source: large recreational lake
Chlorine dioxide was, and still is, the plant’s preference for cooling water
treatment. One spring, after the normal winter shutdown of the C102 system, the

06.qxd 8/8/06 1:57 PM Page 203
system would not start-up because an electrical component had failed. Plant per-
sonnel had no way of knowing in advance about the failure, because the system
was operating properly when it was shut down the previous autumn. The
replacement part could not be procured for over a month, during which time the
plant results staff watched condenser performance steadily decline. The problem,
which was visually confirmed later, was caused by microbiological fouling of the
condenser tubes. The condenser efficiency dropped so low that plant personnel
chlorine-shocked each condenser once during the summer. The process, which
was performed while the condensers were on-line, was interesting. Unit load was
reduced so that one-half of the condenser could be taken out of service. The out-
let valve was closed, and then 25 to 50 pounds of calcium hypochlorite were
blend-filled into the condenser waterbox with diffusion into the condenser tubes.
The chemical was allowed to remain in the condenser for at least four hours, after
which that half of the condenser was valved back into service and the other half
was treated.
The process restored only about 50% of the condenser efficiency. Part of the
problem may have been caused by poor distribution of chemical in the lower
pass of the two-pass condensers. However, part of the problem was also caused
by the slime layer generated by the bacteria. A lot of it did not detach from the
tubes, even though they had been soaked with a concentrated chlorine solution.
That autumn, the plant maintenance crew mechanically scraped the tubes in
each condenser to remove the residual slime and entrained silt deposits.
This case clearly illustrates that proactive treatment of microorganisms is
better than reactive treatment. Once formed, microbiological deposits are diffi-
cult to remove.

06.qxd 8/8/06 1:57 PM Page 204

07.qxd 8/8/06 1:59 PM Page 205
Chapter 7
Sampling
Introduction
The most well-intentioned chemical treatment program may be virtually
worthless without representative sampling. However, sample collection and
transport to on-line instruments requires great care. Some of the key factors that
go into representative sampling include:
• Extracting the sample from the bulk solution so that it represents actual
conditions in the process fluid
• Conditioning the sample to prevent deposition of dissolved constituents
in the sample line
• Maintaining linear velocity of the sample within a suitable range to pre-
vent deposition or entrainment of suspended solids
• Further conditioning the sample to pressures and temperatures that allow
for accurate analyses by on-line instruments

07.qxd 8/8/06 1:59 PM Page 206
While some aspects of sampling have been highlighted in previous chapters,

the following discussion provides a more focused look at this aspect of steam
generation chemistry.
The Need for Sampling

As many of the previous case histories illustrated, chemistry upsets can
occur very suddenly. Except in very low-pressure units, grab sampling by itself
is not recommended, as this only allows the chemist to obtain “snapshot” views
of chemical conditions. Much can happen during the intervening periods. On-
line monitoring has become virtually essential for proper chemistry control.
Tables 7-1 through 7-4 list recommended sample points and sampling para-
meters for electric utility and industrial co-generation steam plants. Some of
these analyses may not be absolute requirements, and readers of this book may
argue for or against some of the analyses that are or have not been mentioned.
However, schemes outlined will allow steam generation personnel to closely
monitor boiler conditions.
Sample Point Selection

The following paragraphs explain the reasoning behind sample point selec-
tion and analyses. First is a discussion of the common sample points and para-
meters outlined in the tables.
Makeup System Effluent

Recommended analyses of the makeup system vary depending on the type
of treatment method. For a cation/anion/mixed-bed demineralizer, analyses of
the anion effluent for sodium, silica, and conductivity provide much meaningful
information and can be used to differentiate between cation and anion problems.
Elevated levels of sodium indicate exhaustion or poor performance of the cation
bed, while high silica concentrations indicate exhaustion of the anion bed. If the
cation bed exhausts before the anion bed, the conductivity of the anion effluent
will increase. However, if the anion bed exhausts first, the conductivity will dip
for a short period before increasing sharply.
Monitoring of the mixed-bed effluent is even more critical, as the mixed-bed
is the final stage in the treatment process and any contaminants will be directly
introduced to the boiler feed system. Typical analyses include sodium, silica, and
specific conductivity.
Although demineralizer analyzers may be mounted in a general water sam-
pling panel, most often they are installed at the demineralizer, with the instru-

07.qxd 8/8/06 1:59 PM Page 207
Sampling 207
Table 7-1
Steam Sample Points and Recommended Analysis
for Utility Drum Boilers Treated with Phosphate
Sample Recommended Grab Sample Frequency

Point On-Line Analysis Analysis
Makeup System
Effluent Sodium Chloride Daily
Silica Sulfate Daily
Spec.Cond. TOC Weekly
Condensate Storage
Tank Effluent* Sodium Daily
Condensate Pump
Discharge Sodium TOC Weekly
Cation Cond.
Dissolved Oxygen
Condensate Polisher
Effluent* Sodium
Cation Cond.
Silica
Deaerator Inlet Dissolved Oxygen
Oxygen Scavenger
Deaereator Outlet Dissolved Oxygen Daily
Feedwater or
Economizer Inlet pH Ammonia Daily
Cation Cond. Iron Thrice-Weekly
Dissolved Oxygen Copper Thrice-Weekly
Oxygen Scavenger
Boiler Water pH Sulfate Daily
Specific Cond. Sodium Daily
Silica Ammonia Daily
Phosphate Dissolved Oxygen Daily
Chloride
Saturated Steam Sodium
Silica
Main Steam/Reheat
Steam** Sodium Sulfate Daily
Silica TOC Weekly
Chloride
Degassed C.C.
* Disregard if no condensate polisher
** For boilers with a reheater, a reheat sample is preferred over main steam.

07.qxd 8/8/06 1:59 PM Page 208
Table 7-2
for Utility Drum Boilers on All-Volatile Treatment

Makeup System
Condensate Storage
Tank Effluent Sodium Daily
Condensate Pump
Cation Cond.
Dissolved Oxygen
Condensate Polisher
Effluent* Sodium
Cation Cond.
Silica
Oxygen Scavenger
Feedwater or
Economizer Inlet pH Iron Thrice-Weekly
Cation Cond. Copper Thrice-Weekly
Dissolved Oxygen
Oxygen Scavenger
Cation Cond. Ammonia Daily
Silica Dissolved Oxygen Daily
Chloride
Silica
Main Steam/Reheat
Silica TOC Weekly
Chloride
Degassed C.C.

07.qxd 8/8/06 1:59 PM Page 209
Sampling 209
Table 7-3
for Once-Through Utility Boilers
On All-Volatile-Treatment

Makeup System
Condensate Storage
Tank Effluent Sodium Daily
Condensate Pump
Cation Cond.
Dissolved Oxygen
Condensate Polisher
Effluent* Sodium
Cation Cond.
Silica
Oxygen Scavenger
Feedwater or
Economizer Inlet pH Iron Thrice-Weekly
Cation Cond. Copper Thrice-Weekly
Dissolved Oxygen
Ammonia
Oxygen Scavenger
Reheat Main Steam/Steam** Sodium Sulfate Daily
Silica TOC Weekly
Chloride
Degassed C.C.

07.qxd 8/8/06 1:59 PM Page 210
Table 7-4
for Cogeneration Drum Boilers Treated with Phosphate,
which Drive Turbines

Makeup System
Effluent Sodium
Silica
Spec.Cond.
Condensate Return pH
Specific Cond.
TOC
Condensate Pump
Cation Cond.
Dissolved Oxygen
Condensate Polisher
Effluent* Sodium
Cation Cond.
Silica
Feedwater or
Economizer Inlet pH Ammonia Daily
Cation Cond. Iron Thrice-Weekly
Dissolved Oxygen Copper Thrice-Weekly
Oxygen Scavenger
Silica Ammonia Daily
Phosphate Dissolved Oxygen Daily
Chloride
Silica
Main Steam/Reheat
Silica TOC Weekly
Chloride
Degassed C.C.
* For boilers with a reheater, a reheat sample is preferred over main steam.

07.qxd 8/8/06 1:59 PM Page 211
Sampling 211
ment outputs wired directly to the demineralizer and/or other control system.
Thus, the demineralizer operator can examine system chemistry data as he or she
operates the equipment.
Less stringent makeup treatment equipment and techniques, and therefore
less sophisticated instrumentation, are often satisfactory for lower-pressure units.
For example, sodium softening is frequently used to produce makeup for boilers
operating at less than 600 psig. At these lower pressures, carryover is much less
severe, so a higher amount of dissolved solids can be tolerated in the boiler water.
Sodium softeners are designed to remove hardness so the most important analy-
ses are those for calcium and magnesium. On-line monitors are available for
these measurements.
Water quality for intermediate- and high-pressure units is achievable not
only with demineralizers, but also with membrane processes, sometimes in
series, or with a mixed-bed polisher. Reverse osmosis (RO) is typically the back-
bone of a membrane treatment system. Because an RO unit uses pressure to
electromechanically filter dissolved solids, on-line measurements usually consist
of pressure, flow, temperature, and conductivity. Fouling, mechanical failure of
equipment, or degradation of membranes will cause significant changes in these
parameters. RO monitors and gauges are typically mounted on the equipment.
Signals may be sent to a local control panel or plant control system.
Monitoring of RO influent pH is frequently also necessary. For example, cel-
lulose acetate membranes must operate within a relatively narrow pH range (4 to
6) to prevent membrane degradation. Thin-film-composite (TFC) membranes
will operate over a much broader pH range, but are affected by even small quan-
tities of chlorine. Where TFC membranes are used to treat chlorinated water, a
dechlorination system is placed ahead of the RO to protect the membranes.
Chlorine residual or oxidation reduction potential (ORP) monitoring of the
dechlorination system effluent provides a safeguard against membrane attack.
Condensate Storage Tank Effluent

Even though makeup system effluent is usually of excellent quality, conta-
minants may be introduced in the storage tank. A daily analyses of sodium will
help the plant chemist keep an eye on water quality in the tank.
Condensate Pump Discharge

This is one of the most critical samples in the entire system. Severe contam-
ination of a boiler is most often caused by a condenser leak. If the condensate
pump discharge is properly sampled and alarmed, leaks can be detected before
a major problem occurs. Typical on-line analyses include sodium, cation con-

07.qxd 8/8/06 1:59 PM Page 212
ductivity, and dissolved oxygen. These will provide a quick indication of a con-
denser tube leak, demineralizer overrun, or an air in-leakage problem.
Given the effects that organics have on boiler water and steam chemistry,
boiler water experts now recommend that TOC be analyzed on a weekly grab
sample basis.
Condensate Polisher Effluent

For those plants equipped with condensate polishers, monitoring of the
effluent is also important. The chemist can determine if the polisher is perform-
ing correctly and has not become exhausted. The typical on-line analyses are the
same as those for the demineralizer, i.e., sodium, cation conductivity, and silica.
Ammonia analyses of hydrogen-cycle polishers can also be valuable, since
ammonia is the most weakly held ion, and its breakthrough indicates the
approach of bed exhaustion.
Deaerator Inlet
The two principal on-line analyses recommended for the deaerator inlet are
oxygen scavenger and dissolved oxygen. These analyses provide data on the per-
formance of the scavenger in the condensate system. Even if the oxygen scav-
enger is injected in the deaerator or deaerator outlet, on-line scavenger monitor-
ing of the deaerator inlet is still useful in determining scavenger carryover
through the steam system.
Deaerator Outlet/Boiler Feed Pump Suction

This sample gives the chemist an extra tool for monitoring deaerator per-
formance. Weekly DO grab sampling is recommended, although daily or even
continuous sampling is easily possible with Chemets or a portable dissolved oxy-
gen monitor.
Feedwater or Economizer Inlet

This is the final sample point before the condensate enters the boiler, and
thus is very valuable. On-line analyses should include pH, cation conductivity,
dissolved oxygen, and oxygen scavenger. Some experts also recommend ammo-
nia analyses, although once-per-shift or daily grab samples seem adequate.
Weekly, or even bi- or tri-weekly analyses of iron and copper are also highly rec-
ommended.
Proper chemistry in the feedwater system is critical to prevent corrosion of
the feedwater piping and high-pressure heaters, transport of corrosion products

07.qxd 8/8/06 1:59 PM Page 213
Sampling 213
to the boiler, and introduction of impurities to the steam attemperator system.

The pH, dissolved oxygen, conductivity, and ammonia samples provide infor-
mation on conformance of feedwater chemistry to recommended guidelines,
while iron and copper analyses indicate how effective the chemical treatment
program is in controlling system corrosion.
Boiler Water
This is another very critical sample, for it is in the boiler that chemistry
upsets typically cause the most damage. Continuous analyses should include pH,
specific conductivity, silica, and phosphate for phosphate-treated units. As chap-
ter 3 discussed, a chemistry upset can rapidly deplete boiler water treatment
chemicals. When this happens, the boiler water pH will decrease (or sometimes
increase) to values outside of recommended ranges. A chemistry upset severe
enough to significantly alter boiler water pH has the potential to cause cata-
strophic tube corrosion. Any monitoring system should be equipped with alarm
capabilities for drum water pH.
For phosphate-treated boilers, phosphate measurements are obviously quite
critical. Phosphate provides the primary protection of the boiler water, and if an
upset depletes the phosphate concentration, severe corrosion or deposition may
result. Phosphate monitoring is also critical for controlling sodium-to-phosphate
ratios in coordinated/congruent treatment programs.
Silica is an extremely important parameter due to its vaporous carryover
qualities and tendency to form tenacious deposits on turbine blades. Silica is usu-
ally the limiting factor when a boiler is started up from outage (Case History 7–1).
Ammonia analyses of phosphate-treated boiler water allow the plant chemist
to calculate correct sodium-to-phosphate ratios. Ammonia analyses are not as
critical for AVT drum units because control is based on ammonia concentrations
in the feedwater.
Chloride and sulfate analyses of boiler water can help the plant chemist
monitor and prevent carryover. Given the problems these two ions can cause in
turbines, and the relative ease with which they carry over, the sampling recom-
mendations seem well justified.
Saturated Steam
A principal reason for sampling saturated steam is to verify that the steam pro-
duced by the boiler meets the turbine manufacturer’s steam quality guidelines. If
main or reheat steam samples are unavailable, then saturated steam sampling
assumes much greater importance. If saturated steam is not the only steam sample,
then saturated steam sampling can be restricted to silica and sodium, as these are

07.qxd 8/8/06 1:59 PM Page 214
most effective for monitoring carryover. Where saturated steam is the only steam
sample, chloride, sulfate, TOC, and conductivity analyses should also be included.
Data from saturated steam and boiler water samples can be used to determine the
percentage of mechanical carryover from the drum water to the steam.
Main/Reheat Steam
Main or reheat steam sample data is often even more informative than satu-
rated steam data because any contaminants introduced by the attemperator sys-
tem will also be measured. For units without a reheat section, main steam is the
only choice, but for units with reheat, the reheat sample is preferred. As with sat-
urated steam, main/reheat samples are used to verify that steam quality is with-
in the turbine guidelines. Recommended continuous analyses include sodium,
silica, conductivity, and chloride, with sulfate and TOC being monitored on a
grab sample basis.
Cogeneration/Combined-Cycle/Industrial
Plant Sampling
At cogeneration, combined-cycle, or industrial steam generating facilities,
sample point selection may be slightly different. For example, heat recovery
steam generators often consist of two or three steam generating circuits and
drums. Steam and drum water samples from each circuit should be obtained if a
comprehensive analysis of the boiler water chemistry is desired.
Sampling of condensate return at industrial plants is very important because
many contaminants may be introduced by process heat exchangers or other
equipment. If, as is very common, the condensate return is treated in a polisher,
the polisher should be sampled similarly to those at electric utilities. However,
more emphasis on TOC analyses is needed when organics or oils have the poten-
tial to enter the condensate. As an example, at one petrochemical plant I had the
chance to visit, all of the condensate return lines to the boilers are equipped with
continuous TOC analyzers. If an analyzer detects organic concentrations above
plant guidelines, the analyzer triggers an automatic dump valve.
These paragraphs have covered the sample points and analyses for various
steam generating systems. The following sections discuss sample collection and
conditioning, topics that are extremely important with regard to chemistry mon-
itoring.
Techniques to Obtain Representative Samples

The factors that influence the reliability of sample collection include sample
nozzle arrangement, sample conditioning, and sample flow rates. Unless each of

07.qxd 8/8/06 1:59 PM Page 215
Sampling 215
these aspects is carefully considered in the sample system design, the reliability
of the samples will be seriously compromised.
Sample Nozzle Design

Sample tap/nozzle design is extremely critical whether the fluid is water or
steam. Researchers have found that samples collected isokinetically (i.e., at the
same flow rate as the bulk fluid) or as close thereto as possible, are most repre-
sentative. The reasons for this are severalfold and are tied into the properties of
the fluid. In most working fluids, which include those at a power plant, fluid
velocities are low enough that a laminar sublayer forms at the pipe walls. The
chemistry of this sublayer is different than that of the bulk fluid. A simple tap
bored into the pipe wall without a nozzle extension allows some of the laminar
fluid to enter with the sample, thus affecting the chemistry.
Figure 7-1 illustrates a liquid sampling nozzle based on ASTM standards.
Figure 7-1
Liquid Sampling Nozzle.

07.qxd 8/8/06 1:59 PM Page 216
Figure 7-2
Generic Outline of a Multiport

Steam Sampling Nozzle
Figure 7-3
Each nozzle is designed with considerations of vortex shedding, resonance, vibration, erosion,
and strength of the attachment to the pipe.
Materials: Ordering Information:
Carbon Steel (C1018 or A105, as specified) 1. Pressure, temperature, and mass flow
Stainless Steel (304 or 316, as specified) rate of the sampled fluid
Low Alloy Chrome Moly Steel 2. Pipe ID, wall thickness, and material.
(F11 or F22, as specified) 3. Desired sample flow.
Other materials available upon request 4. Attachment to the pipe: weld, thread
plus seal weld, flange, etc.
5. Nozzle material.
6. Thickness of thermal insulation.
Pressure-Temperature Rating
lbs. per sq. inch
Temperature - ˚F
Material 70˚ 200˚ 400˚ 600˚ 800˚ 1000˚ 1200˚
Carbon Steel 5950 5750 5450 5250 4000 1750

A.I.S.I. 304 7800 7050 6400 6150 6000 5190 1875
A.I.S.A. 316 7800 7800 7250 7100 6950 5800 2720
F-11 7350 7350 7350 7350 7350 2898 504
F-22 7224 7224 7224 7224 7098 3192 546
EPRI Isokinetic Steam Sampling Nozzle Supplied by Jonas & Consultants. Reprinted courtesy of Jonas &
Consultants, Wilmington, DE.

07.qxd 8/8/06 1:59 PM Page 217
Sampling 217
Features of the nozzle include a beveled entry port, insertion to a depth that sam-
ples the bulk fluid but minimizes the moment of the flowing water on the noz-
zle, and a reinforcing boss to provide strength to the nozzle.
Liquid sampling is usually straightforward, while saturated steam, on the
other hand, is more difficult to sample. These difficulties are caused by the fact
that the fluid is at saturation temperature and may contain fractional water vapor,
or may develop more moisture as the steam cools in the sample line. Thus, the
sample is not totally homogeneous. Multiport sampling is needed to collect rep-
resentative saturated steam samples. At some utilities, the boiler drums have
been designed with multiple sample ports installed in the drum shell. Where this
has not been done, the generic multiport nozzle outlined in Figure 7-2 may be
installed in the saturated steam line. The multiple ports are designed to balance
out variabilities in the constituency of the fluid.
Some debate still exists over sample extraction techniques for superheated
steam. The sample itself is much more homogeneous than saturated steam
Figure 7-4
STEAM FLOW
>
DOUBLE ISOLATION
VALVES
NOZZLE
TO SAMPLE PANEL
>
SAMPLE COOLER
Arrangement of Isokinetic Steam Sampling Nozzles.

07.qxd 8/8/06 1:59 PM Page 218
because superheating converts the fluid to a single phase, and passage through
the superheater or reheater tubes thoroughly mixes the fluid. The debate centers
around isokinetic sampling, whether it is needed and whether it is totally possi-
ble. Figure 7-3 illustrates the generic outline of several isokinetic nozzles
designed by EPRI. The nozzle opening is oriented in the path of the steam flow.
Jonas and Consultants of Wilmington, Delaware offer these isokinetic nozzles for
installation similar to that shown in Figure 7-4.
A few experts are not convinced that even these nozzles offer true isokinetic
sampling. Additionally, concern exists that the nozzle could potentially fracture
and be carried downstream to the turbine. Even so, the trend towards isokinetic
sampling is growing stronger as utility and industrial personnel realize the impor-
tance of representative sampling. However, any user must be aware of the poten-
tial hazards and should install the nozzles according to approved guidelines.
Sample Nozzle
Installation
Except for low-pressure lines such as condensate storage tank discharge,
condensate pump discharge, and deaerator inlet, the procedures for installing
sample taps must conform to appropriate boiler codes for welding and work on
high-pressure lines. The principal code that applies to this work is ANSI B31.1,
which discusses welding procedures for high-pressure pipe. It is absolutely vital
that personnel installing high-pressure taps follow this guideline, both to ensure
the integrity of the sample tap and to prevent problems such as nozzle breakage.
One important requirement of the code is that the pipe be heat treated so that
the welding does not compromise the integrity of the pipe material. These pro-
cedures make sample tap installation into high-pressure lines rather expensive.
Sample Point Location

Certain guidelines apply to sample point and nozzle location. Ideally, boiler
water taps should be installed in a downcomer line, although the continuous
boiler blowdown line is an alternate and frequently used spot. The ASTM rec-
ommends that liquid sample taps be installed at least 25 pipe diameters down-
stream of a chemical injection point if the flow is turbulent, and 50 pipe diame-
ters if the flow is laminar. A rule-of-thumb guideline suggests that the tap be
located at least 10 pipe diameters downstream of a fitting or flow disturbance.
Where possible, taps should be installed in long vertical runs of pipe. When long
vertical runs are unavailable, installation of the nozzle in a long horizontal run is
the next best choice. Jonas and Consultants recommend that steam taps be
installed 35 pipe diameters downstream and 4 pipe diameters upstream of any
flow obstructions, and at a 12:00 position. Regarding liquid samples, I have seen

07.qxd 8/8/06 1:59 PM Page 219
Sampling 219
recommendations for both the 3:00 and 1:30 positions. Either should ensure a
good sample with minimal introduction of solids that may have settled along the
bottom of the pipe.
Primary Sample Conditioning

Once a sample is extracted, it is transported to the on-line instruments and
grab sample ports. Many possibilities exist for the sample chemistry to change if
it is not properly conditioned for flow through the sample line. For example,
steam samples passing unconditioned through the sample line will cool along the
way. This affects the solubility of solids and may cause some of the contaminants
to plate out on the tube walls. The concentration of products reaching the ana-
lyzers will be artificially low. From time to time, pieces of the deposits may break
loose, which will cause an abnormally high spike in readings.
Relatively cool liquid samples such as condensate pump discharge and
deaerator inlet samples are not affected by this phenomenon. However, high-
temperature fluids, especially steam, are usually conditioned to prevent solids
deposition. The lines are equipped with a remote sample cooler located as close
to the root sample valves as possible. Figure 7-4 illustrates this design. Double
isolation valves, required by boiler code, are located within a couple of feet of the
sample tap. These are quickly followed by the cooler, which should be located
as close to the sample tap as possible to minimize any solids deposition. These
initial, or roughing coolers as they are known, should be of adequate size to lower
sample temperatures to approximately 100˚ F.
The most time-consuming procedure when installing remote sample coolers
is plumbing of the cooling water piping. Coolers located at the drum and steam
lines are usually a long distance from, and at a much higher elevation than the
cooling water source. Booster pumps may be required to lift the cooling water to
the coolers. Primary coolers may require up to 12 GPM of cooling water to oper-
ate properly. A loss of cooling water flow could have serious effects on down-
stream components, so the cooling water supply must be reliable.
Recent reports indicate that primary sample conditioning at the root valve
may not be as critical as has been thought. In one side-by-side test of superheat-
ed steam flowing through a conditioned and unconditioned sample line, conta-
minant concentrations were almost identical. Perhaps more research will be con-
ducted on this very interesting possibility.
Sample Flow Rate and Line Size

The optimum linear flow rate of liquid samples (both condensate and con-
densed steam) in the sampling tube is 5 to 6 feet per second. This rate has been

07.qxd 8/8/06 1:59 PM Page 220
found to be best for maintaining the integrity of the samples, and for either pre-
venting deposition of products on the sample tube walls or scouring of previ-
ously deposited products. With this criteria as a guideline, the volumetric flow
rate and line size become dependent on flow requirements of the instruments.
Sample line sizes of 1/4 in. OD and 3/8 in. OD are most common. Volumetric flow
rates in these lines to maintain a 6 foot per second linear rate are 1200 cc/min
and 3300 cc/min, respectively. The Sentry Equipment Corporation recommends
the following line sizes for steam samples.
• Pressure greater than 2000 psig - 1/4 in. sample line

• Pressure between 750 to 2000 psig - 3/8 in. sample line
• Pressure below 750 psig - 1/2 in. pipe
In some cases, the 6 ft/sec linear flow rate taxes the capability of the sample
conditioning equipment. Sampling researchers have found that flow rates can
often be lowered to 3 ft/sec without compromising the integrity of the sample.
When laying out sample line routes, the designer should attempt to mini-
mize the number of elbows and fittings. Horizontal lines should be installed on
level grade or set at a constant slope. Low spots in the lines create problems.
Final Sample Conditioning

Additional conditioning of the samples is needed before they are ready for
analysis. Figure 7-5 shows the flow diagram of a high-pressure and temperature
sample at the sample panel and downstream of the primary cooler. The princi-
pal instruments consist of:
• High-pressure reducing valve

• Pressure regulating valve
• High-temperature shutoff solenoid
• Low-pressure blowdown
• Secondary sample cooler
• Total flow rotameter
• Pressure gauge
• Temperature gauge
• Individual rotameters for the on-line instruments
• Backpressure regulator
• Grab sample discharge
The high-pressure reducer is usually a rod-in-tube device designed to take

pressures from the several hundred to several thousand psi of the rough samples

07.qxd 8/8/06 1:59 PM Page 221
Sampling 221
Figure 7-5
PRIMARY SAMPLE COOLER
HIGH PRESSURE REDUCING VALVE

S
HIGH TEMPERATURE SHUTOFF SOLENOID
PI
LOW PRESSURE BLOWDOWN VALVE
SECONDARY SAMPLE COOLER
TI
FI
> > >
FI FI FI
TO GRAB TO ON-LINE
DRAIN SAMPLE INSTRUMENTS
Flow Schematic of Sample Conditioning for a High-Pressure, High Temperature Sample.
to less than 100 psi. Further fine-tuning of the pressure is then accomplished
with the pressure regulator. The high-temperature solenoid is a safety device that
will cut off sample flow if a preset temperature is exceeded. It protects down-
stream equipment against high-temperatures caused by loss of cooling water to
the roughing coolers.
The low-pressure blowdown line is used to send samples to waste during
unit startup. At unit startup, power plant fluids, especially boiler water, contain
many particulates that could foul the downstream equipment. Sometimes the
blowdown line is installed ahead of the high-pressure regulator so that these par-
ticulates do not foul the high-pressure reducer and pressure regulator. However,
safety factors involved with blowing down high-pressure samples have generat-
ed greater interest in the low-pressure blowdown design.
The sampling system also includes a secondary cooler. The most accurate
readings are obtained if the sample is cooled to 77˚F, ±1 /2˚F. Some instrument
manufacturers claim that their temperature compensation devices eliminate the
need for secondary cooling, but the arguments are not yet convincing. Unless a
manager’s budget does not allow for secondary coolers, the coolers should be
installed to help ensure representative readings.

07.qxd 8/8/06 1:59 PM Page 222
Figure 7-6
Complete Sample Conditioning System. Photo courtesy of the Sentry Equipment Corp., Oconomowoc, WI.
A chilled water supply is usually required for the secondary coolers. The
water can either be passed through the coolers and discharged or can be circu-
lated through an isothermal bath and be reused over and over. Recirculation
saves on water costs but may increase the size of the chiller.
The rotameters/flow controllers are used to set sample flow rates at recom-
mended values. The combination of a main sample rotameter and individual
flow rotameters provides flexibility in adjusting both the total sample flow rate
and the flow rate to each on-line instrument.
The backpressure regulator helps maintain a constant flow to the system,
and minimizes upsets.
Figures 7-6 and 7-7 illustrate how the equipment in a sample panel looks
when it is completely assembled.
Data Acquisition
Figure 7-8 shows the actual layout of the on-line water chemistry system at
the Dallman Power Station at my former utility. On-line instruments include
hydrazine, pH, specific conductivity, cation conductivity, sodium, silica, phos-
phate, and dissolved oxygen. Each analyzer sends continuous signals to a PLC
located in the analysis room. The PLC then communicates to various points in
the plant. It sends control signals to the hydrazine pumps of three generating
units. The pump stroking rate is controlled from signals sent by the deaerator
inlet hydrazine analyzers. The PLC also provides an audible alarm if condensate
pump discharge sodium levels exceed 5 ppb, or if drum pH drifts 0.4 units above
or below a predetermined range.

07.qxd 8/8/06 1:59 PM Page 223
Sampling 223
Figure 7-7
Sample Conditioning Piping. Photo courtesy of the Sentry Equipment Corp., Oconomowoc, WI.
Finally, the PLC is the data collection source for continuous data display
screens located in the plant control room, main laboratory, and engineering
offices. Two of the screens are illustrated in Figures 7-9 and 7-10. The first is the
primary screen in the control room. Should any of the values exceed the ranges
shown, a red warning light will flash in the “Unit” boxes to the right. Not all of
the data collected by the analyzers is included on the screen, because the system
designers, including myself, did not want to overwhelm the operators with num-
bers. We wanted them to only be concerned with the most important readings.
Figure 7-10 is a simplified flow diagram of one of the units showing real-time
water chemistry data. This screen is particularly useful for personnel who wish
to look at any point in the system and immediately analyze the chemistry. Both
screens are color coded so that the plant staff can differentiate between units and
between process lines on the flow diagram.
Conclusion
Proper sampling of steam generation fluid samples is extremely important in
guaranteeing the overall reliability of the plant. Without representative sampling,
upsets could occur that would have potentially catastrophic results. This chapter
hopefully serves as a guideline for the equipment and procedures that go into the
design of a truly useful sampling system.
07.qxd 8/8/06 1:59 PM Page 224
Figure 7-8
Laboratory Annex
Oxygen To Instrument
pH & Cond. Analyzers Analyzers Solenoid Valves
To Dallman
Sample Line Inlets PLC Control Room
(Data Display)
Si & Po4 (Alarm System)
Display Panel
Conductivity
pH, Oxygen &
Sink
Hydrazine Sodium Analyzers
Analyzers Analyzers To Main Laboratory
Communication (Data Display)
Routes = (Data Logging)
To Hydrazine
Pump Stroke
Actuators
Computer
Workstation
On-Line Sample System Arrangement at City Water, Light & Power, Dallman Generating Station
Springfield, IL.
Figure 7-9
Dallman Unit Chemistry
Unit 31
Unit 33 Unit32
Control Control
Unit 33 Range Unit32 Range Unit 31
Drum pH 9.24 9.1–9.5 9.40 9.1–9.7 9.52 Unit chemistry

Drum Silica (ppb) 43 <160 496 <750 O.O.S alarms – when any
Feedwater pH 8.96 8.8–9.3 8.82 8.8–9.3 8.92 box is lit please
D.A. Inlet Hydrazine (ppb) 16 20–30 5 20–30 7 compare on-line
W.C.P.D. Sodium (ppb) O.O.S <2 O.O.S <2 O.O.S data with control
E.C.P.D. Sodium (ppb) 1.00 <2 0.26 <2 0.83 parameters.
W.C.P.D. Cation Cond. O.O.S <1 0.226 <1 0.177
E.C.P.D. Cation Cond. 0.155 <1 0.311 <1 0.196
W.C.P.D. = West condensate pump discharge

E.C.P.D. = East condensate pump discharge
O.O.S. (blue) = Equipment “out of service” due to a unit outage.
O.O.S (other colors)= Equipment “out of service” either because the on-line instrument has cycled to
another unit, or because unit load is low enough that the equipment is off.
Source: City Water, Light & Power
Primary On-Line Data Display Screen at Dallman Generating Station.

07.qxd 8/8/06 1:59 PM Page 225
Sampling 225
Figure 7-10
Sat. Steam
Na= 0.74 ppb Main Steam
Si= .15
Hot Reheat
Drum Water
pH = 9.39 Turbine
Si = 40 ppb
Phos = 0.0S
Spec.
Conc. = 15.1 Cold Reheat
Economizer Inlet D.A. Inlet Condenser

pH = 9.42 D.O. = 6 ppb
D.O. = 0 ppb HYD. = 5 ppb
HYD. = 1 ppb W.C.P.D.
Spec. Na = 0.54 ppb
Cond.= 6.20 D.A. Cat.
Cond.= 0.249
E.C.P.D
Na = 0.0.S.
Cat.
Cond.= 0.210
Feedwater Heaters
Primary On-Line Data Display Screen at Dallman Generating Station.

07.qxd 8/8/06 1:59 PM Page 226

07.qxd 8/8/06 1:59 PM Page 227
Case History 7-1

Conditions: 2400 psig steam generating unit
Scheduled unit outages are often deleterious to unit chemistry when the unit
is restarted. The stresses that occur in the unit during cool down and reheating,
combined with maintenance work that may have been performed, serves to
loosen deposits within the boiler and introduce excess contaminants to the sys-
tem. Serious problems may result if the boiler is started up without regard for
chemistry.
For the boiler outlined in this history, silica concentration is the guideline
upon which unit pressure is increased. At unit startup the pressure is only
allowed to reach 1800 psig. This usually ensures that boiler water and steam sil-
ica concentrations remain below harmful levels. Only when a portion of the sil-
ica has been removed by the boiler blowdown will the pressure be increased.
This is done in 200 psig increments until final pressure is established.
Sometimes, a 200 psig increase will cause silica concentrations to exceed guide-
lines. If this happens, boiler load is reduced until additional silica is removed.

07.qxd 8/8/06 1:59 PM Page 228
Case History 7-2

Conditions: Two 950 psig boilers
Cooling coil attemperation
This discussion is a continuation of Case History 3–5, which outlined a cure
for leaks in two cooling coil attemperator bundles.
As the earlier case history outlined, the cooling coil attemperator bundles
periodically failed due to insufficient layup procedures. The leaks were usually
detected by lab personnel via sample analyses. This was possible because the
steam within the bundles was at a slightly lower pressure than the mud drum
water which surrounded the tubes. If a leak developed, drum water would enter
the steam and introduce trace but detectable levels of phosphate. As soon as lab
personnel noticed such an occurrence they would inform the operations staff,
who would then schedule repair of the bundles during the next available outage.

07aBiblio.qxd 8/8/06 2:00 PM Page 229
Bibliography
Books
Applebaum, S.B., Demineralization by Ion Exchange, New York: Academic Press, 1968.
BetzDearborn, Inc., Betz Handbook of Industrial Water Conditioning, ninth edition,

Trevose, Pennsylvania, 1991.
Byrne, W., Reverse Osmosis: A Practical Guide for Industrial Users, Littleton, Colorado:
Tall Oaks Publishing, Inc., 1995.
Cantafio, A.R., ed., Drew Principles of Industrial Water Treatment, eleventh edition,
Boonton, New Jersey: Drew Industrial Division Ashland Chemical Co., 1994.
Cohen, P., ed., The ASME Handbook on Water Technology for Thermal Power Systems,
New York: The American Society of Mechanical Engineers, 1989.
Dillon, C.P., Corrosion Control in the Chemical Process Industries, second edition,
Houston. Published for the Materials Technology Institute by NACE International, 1994.
Hensley, H.C., ed., Cooling Tower Fundamentals, second edition, Overland Park, Kansas:
The Marley Cooling Tower Company, 1985.
Herro, H.M., and R.D. Port, The Nalco Guide to Boiler Failure Analysis, New York:
McGraw-Hill.
Meller, F.H., ed., Electrodialysis (ED) & Electrodialysis Reversal Technology, Watertown,
Massachusetts: Ionics, Incorporated, 1984.
National Association of Corrosion Engineers, Corrosion Basics: An Introduction,

Houston: NACE International, 1984.
Owens, D.L., Practical Principles of Ion Exchange Water Treatment, Littleton, Colorado:
Tall Oaks Publishing, Inc., 1995.

230 Bibliography
Technical Reports
“Interim Consensus Guidelines on Fossil Plant Cycle Chemistry,” Palo Alto, California:
Electric Power Research Institute, June 1986. CS-4629.
“Water, Steam, and Turbine Deposit Chemistries in Phosphate-Treated Drum Boiler

Units,” Palo Alto, California: Electric Power Research Institute, August 1987. CS-5275.
Seminars
BetzDearborn, Inc., “Betz Laboratories Water Treatment Technology Seminar,” Trevose,

Pennsylvania: BetzDearborn, Inc.
College of Engineering, University of Wisconsin-Madison, “Improving Cooling Tower

Operation and Cooling Water Treatment Seminar,” January 4–6, 1995, University of
Wisconsin-Madison.
Articles and Technical Papers
Borck, W., and G. Bartley, “Equilibrium Phosphate Treatment Conversion at TVA Gallatin
Fossil Plant.” In Proceedings of the 1996 International Water Conference, Pittsburgh,
1996.
Dooley, B., “The Cutting Edge of Cycle Chemistry for Fossil Plants.” In Proceedings of the
1996 International Water Conference, Pittsburgh, 1996.
Harfst, W. “Back to Basics: Using Reverse Osmosis to Pretreat Demineralizer Feedwater.”

Ultrapure Water, April 1994, pp. 45–47.
Harfst, W., “Chlorine or Bromine: Which is Right for Your System?” Power, September
1993, pp. 70–72.
Jonas, O., L. DuPree, T. Gilchrist, G. Lawrence, and M. Stevens, “Copper Deposition and
MW Loss: Problem and Its Solutions.” In Proceedings of the 1996 International Water
Conference, Pittsburgh, 1996.
Jonas, O., and B. Dooley, “Steam Chemistry and Its Effects on Turbine Deposits and
Corrosion.” In Proceedings of the 1996 International Water Conference, Pittsburgh, 1996.

Bibliography 231
Port, R.D., and H.M. Herro, “An Overview of Water-Formed and Steam-Formed
Deposits,” Industrial Water Treatment, May/June 1992, pp. 33–39.
Robinson, J.O., and A.W. Fynsk, “A Practical Guide to Avoiding Steam Purity Problems in
the Industrial Plan,” Proceedings of the 1992 International Water Conference, Pittsburgh,
1992.
Smock, B., ed., “Water Treatment: Boiler Feedwater and Condensate.” Series of reprints
from Power Engineering magazine.
Stockdill D., and G. Reck, “Full Scale Evaluation of Hydrazine Alternatives in a High
Pressure Boiler,” Paper No. 204 presented at Corrosion 94. Published by the National
Association of Corrosion Engineers, Houston.
Stodola, J., “Ten Years of Equilibrium Phosphate Treatment.” In Proceedings of the 1996
International Water Conference, Pittsburgh, 1996.
Strauss, S., ed., S. Keen, and P. Puckorius, “Boiler Water Treatment for Low- and
Moderate-Pressure Plants,” Power Special Report, June 1987.
Thaler, J.O., and P.K. Sinha, “Control Parameter Model Speeds Water Treatment
Calculations,” Power Engineering, November 1987, pp. 31–32.
Wiltsey, D.G., and C.B. Batton, “Hydrazine Alternative Optimizes Corrosion Control: A
Twelve-Year History of Results in Utility Cycles.” Presented at Power-Gen Europe, Paris,
May 25–27, 1993.


Index.qxd 8/8/06 2:07 PM Page 233
Index
Acid feed,
for alkalinity reduction in cooling water .….….….….…178-179, 198-199
Acid phosphate corrosion of boiler tubes .….….….….….….….….….….… 66
Acrylates
boiler water treatment .….….….….….….….….….….….….….….….…68
cooling water treatment .….….….….….….….….….….….….….….…180
Activated carbon
guidelines for selection .….….….….….….….….….….….….….….…151
makeup water pretreatment .….….….….….….….….….…102-103, 106
Admiralty condenser tubes .….….….….….….….….….….…6, 47, 169, 201
Aeration .….….….….….….….….….….….….….….….….….….….….…106
Air inleakage to condensers and condensate/feedwater systems .…10, 46-47
Air removal from condensers .….….….….….….….….….….….….…10-13
Algae .….….….….….….….….….….….….….….….….….….….….…184
Alkaline cooling water treatment programs .….….….….….….….…179-181
Alkalinity,
bicarbonate .….….….….….….….….….….….….….….….….…178-179
carbonate .….….….….….….….….….….….….….….….….….…178-179
control in cooling towers .….….….….….….….….….….….….…178-179
effect on strong acid cation exchanger performance .….….….….…117-118
effect on weak acid cation exchanger performance .….….….….….…124
hydroxide .….….….….….….….….….….….….….….….….….…178-179
“M” .….….….….….….….….….….….….….….….….….….….…178-179
“P” .….….….….….….….….….….….….….….….….….….….… 178-179
All-volatile treatment (AVT) .….….….….….….….….….….…23, 34, 69-70
American Society of Mechanical Engineers boiler and feedwater chemistry guide-
lines .….….….….….….….….….….….….….….….….….….….…24-30
Amines
distribution ratio .….….….….….….….….….….….….….….….….…21
effect on sodium-to-phosphate ratios in boiler water .….….….….….…64

234 Power Plant Chemistry: A Practical Guide
filming .….….….….….….….….….….….….….….….….….….….….…21
function of in steam-generating systems .….….….….….….….….…7, 20-21
neutralizing .….….….….….….….….….….….….….….….….….…20-21
used in boiler layup .….….….….….….….….….….….….….….….….…73
Aminoethylenephosphonic acid (AMP) .….….….….….….….….….….…180
Ammonia,
as a condensate/feedwater treatment .….….….….….….….….….….…7, 20
as a copper-alloy tubed heat exchanger corrodent .…7, 12, 15, 21, 47, 169
effects on condensate polishing .….….….….….….….….….….….….…36
effect on sodium-to-phosphate ratios in boiler water .….….….…63-64, 76
in oxygenated treatment programs .….….….….….….….….….….…22-23
reaction with chlorine in cooling water systems .….….….….….….….…185
Anion exchange resin (see Ion exchange)
Anodic corrosion protection in cooling water systems .….….….….…170-171
Antiscalants for RO systems .….….….….….….….….….….….….….….…140
Attemperation
effects on steam purity .….….….….….….….….….….….….…84, 91, 93
detection of leaks .….….….….….….….….….….….….….….….….…228
Azoles
Copper corrosion inhibitors .….….….….….….….….….….…172-173, 181
Backwash
ion exchange resins .….….….….….….….….….….….….….….…128-129
media in filtration units .….….….….….….….….….….….….….….…105
Bacteria .….….….….….….….….….….….….….….….….….….….….…184
Barium sulfate scale on RO membranes .….….….….….….….….….….…140
Blowdown,
Boiler .….….….….….….….….….….….….….….….….….….….…56, 62
cooling tower .….….….….….….….….….….….….….….….….…162-163
Boilers
circulating fluidized bed (CFB) .….….….….….….….….….….….….…53
drum .….….….….….….….….….….….….….….….….….….….…51-54
field-erected .….….….….….….….….….….….….….….….….….…52-53
heat recovery steam generators (HRSG) .….….….….….….….….….…53-54
once-through .….….….….….….….….….….….….….….….….….….…54
package .….….….….….….….….….….….….….….….….….….….…52-54
Boiler chemical cleaning .….….….….….….….….….….….….….…20, 55, 78
Boiler drum internals .….….….….….….….….….….….….….….….…88-89
Boiler layup .….….….….….….….….….….….….….….….….…72-73, 80-81
Boiler tubes
corrosion .….….….….….….….….….….….….….….….….…57-60, 80-81

deposits.….….….….….….….….….….….….….….….….….….….…57-60
Boiler Water Chemistry .….….….….….….….….….….….….….….…51-82
ASME guidelines .….….….….….….….….….….….….….….….….…24-30
effects of condenser tube leaks .….….….….….….….….….….….…56-58
guidelines to control steam chemistry .….….….….….….….….….….…92
water chemistry upsets and effects on steam chemistry .….….….….…97-99
Boiler water treatment programs .….….….….….….….….….….….…2, 60-73
all-volatile-treatment (AVT) .….….….….….….….….….….….….…69-70
early programs .….….….….….….….….….….….….….….….….…61-62
caustic .….….….….….….….….….….….….….….….….….….….….…69
chelants and polymers .….….….….….….….….….….….….….….…67-68
coordinated/congruent phosphate .….….….….….….….….….….…62-64
equilibrium phosphate (EPT) .….….….….….….….….….….….….…66-67
oxygenated treatment (OT) .….….….….….….….….….….…21-23, 70, 79
phosphate treatment (PT) .….….….….….….….….….….….….….….…67
treatment for heat recovery steam generators .….….….….….….….…70-72
Bromine
cooling water microbiocide .….….….….….….….….….….….….…187-188
pretreatment for makeup water production .….….….….….….….….…102
Calcium
removal in ion exchange units .….….….….….….….….….…108-109, 112
Calcium carbonate scale
in boilers .….….….….….….….….….….….….….….….….….….….….…3
in cooling water systems .….….….….….….….….….….….….…3, 173-176
on RO membranes .….….….….….….….….….….….….….….….….…140
Calcium fluoride scale
in cooling water systems .….….….….….….….….….….….….….….…173
on RO membranes .….….….….….….….….….….….….….….….….…140
Calcium hypochlorite .….….….….….….….….….….….….….….….….…186
Calcium phosphate scale
in cooling water systems .….….….….….….….….….….….…3, 173, 177
Calcium sulfate scale
in cooling water systems .….….….….….….….….….….…3, 173, 177, 179
on RO membranes .….….….….….….….….….….….….….….….….…140
Carbohydrazide
metal passivator .….….….….….….….….….….….….….….….….….…16
oxygen scavenger .….….….….….….….….….….….….….….….….….…16
Carbon dioxide
corrosive effects of .….….….….….….….….….….….….….….…2, 3, 8, 17
produced by decomposition of water treatment chemicals 16-17, 18, 21, 70
Carbonic acid corrosion .….….….….….….….….….….….….….….….….…8

Carryover .….….….….….….….….….….….….….….….….…69-70, 83-100

caused by severe boiler water contamination .….….….….….….….…97-99
in all-volatile-treatment programs .….….….….….….….….….….…69-70
mechanical .….….….….….….….….….….….….….….….…84, 88-89, 100
vaporous .….….….….….….….….….….….….….….….….….….…84, 90
Cathodic corrosion protection in cooling water systems .….….….…170-172
Cation conductivity (see also Sampling)
steam purity guidelines.….….….….….….….….….….….….….….…90, 92
Cation exchange resins (see Ion exchange)
Cationic polymers for clarification .….….….….….….….….….….…103-104
Caustic
regenerant for anion exchange resins .….….….….….….….….…114, 129
specifications.….….….….….….….….….….….….….….….….….….…152
Caustic corrosion
of boiler tubes .….….….….….….….….….….….….….….….….…58, 62
Caustic treatment of boiler water .….….….….….….….….….….….…61, 69
Chelants
in boiler water treatment .….….….….….….….….….….….….….…67-68
Chemical cleaning
Boilers .….….….….….….….….….….….….….….….….….….…20, 55, 78
superheaters .….….….….….….….….….….….….….….….….….….…94
Chemical feed pumps .….….….….….….….….….….….….….….….….…23
Chemical feed systems .….….….….….….….….….….….….….….….….….…
boiler feedwater .….….….….….….….….….….….….….….…17-18, 31-32
hydrazine .….….….….….….….….….….….….….….….….….….…17-18
oxygenated treatment .….….….….….….….….….….….….….….…23, 32
portable containers .….….….….….….….….….….….….….….….….…18
Chloride
boiler water guidelines .….….….….….….….….….….….….….….…92-93
corrosive effects on turbine blades .….….….….….….….….….….….…86
steam carryover product .….….….….….….….….….….….….…69, 84, 86
Chlorine
cooling water microbiocide .….….….….….….….….….….….….…184-186
effect on RO membranes .….….….….….….….….….….….….…137, 139
Chlorine dioxide
cooling water microbiocide .….….….….….….….….…188-189, 202-203
effect on RO membranes .….….….….….….….….….….….….…137, 139
Chromate (for cooling water corrosion inhibition) .….….….….….….….…172

Circulating fluidized bed boiler (CFB) .….….….….….….….….….….….…53

Clarification
makeup water pretreatment .….….….….….….….….….….….…103-105
Closed cooling systems .….….….….….….….….….….….….….…1, 165, 202
microbiological fouling .….….….….….….….….….….….….….….…202
sodium nitrite treatment .….….….….….….….….….….….….….….…202
Coagulation .….….….….….….….….….….….….….….….….….…103-104
effect of residual coagulants on RO membranes .….….….….….….….…140
Cogeneration .….….….….….….….….….….….….….….….….….….….…2
Combined-cycle power plants .….….….….….….….….….….….….….….…2
Condensate/feedwater chemistry .….….….….….….….….….….….… 5-34
carbonic acid corrosion .….….….….….….….….….….….….….….….…8
chemical feed systems .….….….….….….….….….….….….….….…31-32
condensate/feedwater system materials of construction .….….….….….…6
control guidelines .….….….….….….….….….….….….….….…24-31, 33
copper-alloy corrosion .….….….….….….….….….….….….….….…19-20
dissolved oxygen (see also Dissolved oxygen)
chemical control of .….….….….….….….….….….….….….….…13-23
corrosive effects of .….….….….….….….….….….….….….….….…6-8
mechanical control of .….….….….….….….….….….….….…10-13, 80
flow-accelerated corrosion (FAC) .….….….….….….….….….….….…19
monitoring of condensate contaminants .….….….….….….….….…32-34
oxygenated treatment (OT) .….….….….….….….….….….….….…21-23
pH and corrosion control with ammonia and amines .….….….….…20-21
Condensate polishing.….….….….….….….….….….….….….…23, 34-38, 91
ammonia form .….….….….….….….….….….….….….….….….…36, 38
at industrial plants .….….….….….….….….….….….….….….….….…38
deep bed .….….….….….….….….….….….….….….….….….….…34-36
effects of ammonia on run length .….….….….….….….….….….….….…36
powdered-resin .….….….….….….….….….….….….….….….…34, 37-38
precoat process, powedered-resin.….….….….….….….….….….….….…37
regeneration .….….….….….….….….….….….….….….….….….…35-36
resins .….….….….….….….….….….….….….….….….….….….….…35
Condensate polishing at industrial plants .….….….….….….….….….….…38
Condensate return .….….….….….….….….….….….….….….…5, 8, 21, 38
Condenser air in-leakage .….….….….….….….….….….….….….…10, 46-47
detection of .….….….….….….….….….….….….….….….….….…11-12
Condenser air removal compartment .….….….….….….….….….….…10-11
Condenser performance monitoring .….….….….….….….…12, 41-45, 46-47
BASIC program to calculate performance .….….….….….….….….…44-45
Condenser tube leaks .….….….….….….….….….….….….….….…32, 47-48

BASIC program to calculate size of a tube leak .….….….….….…45-46, 49

causes .….….….….….….….….….….….….….….….….….….….….…34
effects on boiler water chemistry .….….….….….….….….….….….…56-60
effects on steam chemistry .….….….….….….….….….….….…76-78, 97
Condenser steam-side corrosion .….….….….….….….….….….….….….…47
Condensers
chlorine shock treatment .….….….….….….….….….….….….…202-203
Congruent and coordinated phosphate treatment .….….….….….…61, 62-64
Cooling-coil steam attemperator corrosion.….….….….….….….….….….…80
Cooling systems, types
closed recirculating .….….….….….….….….….….….….….….….…1, 156
once-through .….….….….….….….….….….….….….….….….…156-157
open recirculating .….….….….….….….….….….….….….…156, 157-165
Cooling towers
approach .….….….….….….….….….….….….….….….….….….….…164
blowdown .….….….….….….….….….….….….….….….….….…162-163
calculations .….….….….….….….….….….….….….….….….….…160-165
BASIC program .….….….….….….….….….….….….….….….…195-196
cycles of concentration .….….….….….….….….….….….….…161-162
evaporation factor “f” .….….….….….….….….….….….….….….…161
evaporation rate .….….….….….….….….….….….….….….…161-163
drift .….….….….….….….….….….….….….….….….….….….…162-163
fans .….….….….….….….….….….….….….….….….….….….…158-159
fill .….….….….….….….….….….….….….….….….….….….….….…160
makeup .….….….….….….….….….….….….….….….….….….…162-163
materials of construction .….….….….….….….….….….….….….….…165
types .….….….….….….….….….….….….….….….….….….….…157-160
crossflow .….….….….….….….….….….….….….….….….….…157-158
counterflow .….….….….….….….….….….….….….….….….…157-159
forced-draft .….….….….….….….….….….….….….….….….…157-159
hyperbolic .….….….….….….….….….….….….….….….….…159-160
induced-draft .….….….….….….….….….….….….….….….…157-159
natural-draft .….….….….….….….….….….….….….….….….…157-159
wet-bulb temperatures .….….….….….….….….….….….….….…163-165
Cooling Water Chemistry .….….….….….….….….….….….….….…155-203
corrosion .….….….….….….….….….….….….….….….….….…166-173
anodes .….….….….….….….….….….….….….….….….…167-168, 169
cathodes .….….….….….….….….….….….….….….….…167-168, 169
copper alloy corrosion .….….….….….….….….….….….…168-169, 201
electrochemistry nature of .….….….….….….….….….….….…167-169
galvanic .….….….….….….….….….….….….….….….….….…169-170

influence of hydrogen ions .….….….….….….….….….….….…167-168

iron corrosion .….….….….….….….….….….….….….….….…167-168
non-metallic corrosion .….….….….….….….….….….….….….….…170
pitting .….….….….….….….….….….….….….….….….….….….…168
temperature effects on .….….….….….….….….….….….….….….…169
corrosion inhibition .….….….….….….….….….….….….….….…170-173
anodic inhibitors .….….….….….….….….….….….….….….…170-172
molybdate .….….….….….….….….….….….….….….….….…171-172
nitrites .….….….….….….….….….….….….….….….….….….…171
orthophosphate .….….….….….….….….….….….….….….…171-172
potential problems with .….….….….….….….….….….….….….…171
silicates .….….….….….….….….….….….….….….….….….….…171
cathodic inhibitors .….….….….….….….….….….….….….….…170-172
phosphonates.….….….….….….….….….….….….….….….…171-172
polyphosphates .….….….….….….….….….….….….….….…171-172
potential problems with .….….….….….….….….….….….….….…171
zinc .….….….….….….….….….….….….….….….….….….…171-172
cathodic protection .….….….….….….….….….….….….….….….…170
chromates and discontinued use of .….….….….….….….….….….…172
coating .….….….….….….….….….….….….….….….….….….….…170
common inhibitor dosage levels .….….….….….….….….….….….…172
copper alloy corrosion inhibition .….….….….….….….…172-173, 181
environmental considerations .….….….….….….….….….….….….…172
sacrificial anodes .….….….….….….….….….….….….….….….….…170
fouling .….….….….….….….….….….….….….….….….….….…181-184
microbiological .….….….….….….….….….….….….….….….…183-184
non-microbiological .….….….….….….….….….….….….….…181-183
fouling control methods
chemical treatment .….….….….….….….….….….….….….….….…183
common silt inhibitor dosage levels .….….….….….….….….….….…183
future trends .….….….….….….….….….….….….….….….….….…190
macrofouling .….….….….….….….….….….….….….….….…190-193
Asiatic clams.….….….….….….….….….….….….….….….….…190-191
zebra mussels.….….….….….….….….….….….….….….…190, 191-193
chemical treatment .….….….….….….….….….….….….….….…192
thermal treatment .….….….….….….….….….….….….….….…193
microbiological .….….….….….….….….….….….….….….….…184-193
non-oxidizing chemicals .….….….….….….….….….….….…189-190
bromonitropropanediol .….….….….….….….….….….….….…190
isothiozolone.….….….….….….….….….….….….….….….….…190
quaternary amines .….….….….….….….….….….….….….….…190

oxidizing chemicals .….….….….….….….….….….….….….…184-189

bromine .….….….….….….….….….….….….….….….….…187-188
calcium hypochlorite .….….….….….….….….….….….….….…186
chlorine .….….….….….….….….….….….….….….….….…184-186
chlorine dioxide .….….….….….….….….….….….….….…188-189
hydantoins .….….….….….….….….….….….….….….….…186, 188
isocyanurates.….….….….….….….….….….….….….….….….…186
ozone .….….….….….….….….….….….….….….….….….….…189
sodium hypochlorite .….….….….….….….….….….….…186, 200
solid chlorine donors .….….….….….….….….….….….….….…186
off-line condenser tube scraping .….….….….….….….….….…182-183
on-line condenser tube cleaning systems .….….….….….….….….…182
sidestream filtration .….….….….….….….….….….….….….….….…182
calculation of filter size .….….….….….….….….….….…182, 199-200
shock treatment of condenser tubes .….….….….….….….….….…202-203
scale .….….….….….….….….….….….….….….….….….….….…173-177
calcium carbonate .….….….….….….….….….….….…173-176, 200-201
calcium fluoride .….….….….….….….….….….….….….….….….…173
calcium phosphate .….….….….….….….….….….….….….…173, 177
calcium sulfate .….….….….….….….….….….….….….….….…173, 177
Langelier Saturation Index (LSI).….….….….….….….….…173-176, 181
BASIC program .….….….….….….….….….….….….….….….…196-198
magnesium silicate .….….….….….….….….….….….….….….….…177
Practical Scale Index (PSI) .….….….….….….….….….….….…175, 181
Ryznar Stability Index .….….….….….….….….….….….….….…174-175
silica .….….….….….….….….….….….….….….….….…173, 177, 181
scale inhibition .….….….….….….….….….….….….….….….…177-181
alkaline treatment methods .….….….….….….….….….….….…179-181
common chemical dosage levels .….….….….….….….….….….….…181
crystal modifiers .….….….….….….….….….….….….….….…180-181
phosphonates.….….….….….….….….….….….….….….….…180-181
polymer solubilizers.….….….….….….….….….….….….….…180-181
sulfuric acid feed .….….….….….….….….….….….….….….…178-179
BASIC program .….….….….….….….….….….….….….….…198-199
Coordinated phosphate treatment .….….….….….….….….….….…61, 62-64
Copper
steam carryover product .….….….….….….….….….….….…20, 84-85, 90
steam purity guidelines .….….….….….….….….….….….….….….….…90
Copper alloys
corrosion .….….….….….….….….….….….….….….…6-8, 19-20, 47, 90
corrosion control .….….….….….….….….….….…13-23, 31-32, 72-73, 80

materials for heat exchanger tubes .….….….….….….….….….….….…6, 7

Copper deposition
on boiler tubes .….….….….….….….….….….….….….….….….….…20
on turbine blades .….….….….….….….….….….….….….…20, 84-85, 90
Copper-nickel alloys .….….….….….….….….….….….….….….….….…4, 47
Corrosion
carbon steel condensate/feedwater lines and boiler tubes .….….…7-8, 58-60
cooling water systems .….….….….….….….….….….….….….…166-173
copper-alloy heat exchanger tubes .….….….….….…6-8, 19-20, 109, 202
off-line corrosion of steam generating systems.….….….….….….….…72-73
Corrosion products
transport to boiler .….….….….….….….….….….….….….…5, 6, 23, 79
Counterflow cooling towers
Crossflow cooling towers .….….….….….….….….….….….….….…157-159
Cuprous oxide
protective film .….….….….….….….….….….….….….….….….….….…7
Cycles of concentration,
cooling towers .….….….….….….….….….….….….….….….….…161-162
Cyclohexylamine.….….….….….….….….….….….….….….….….….….…20
distribution ratio of .….….….….….….….….….….….….….….….….…22
Deaeration
chemical .….….….….….….….….….….….….….….….….….….…13-14
mechanical.….….….….….….….….….….….….….….….….….….…12-13
Deaerators .….….….….….….….….….….….….….….….….….….….…12-13
Deep-bed condensate polishers .….….….….….….….….….….….….…34-36
Dechlorination of cooling water discharge .….….….….….….….….…185-186
Degasification of makeup water.….….….….….….….….….….….….….…111
Demineralizers (see also ion exchange)
design fundamentals
distributors .….….….….….….….….….….….….….….….….….…133
materials of construction .….….….….….….….….….….….….….…134
valves .….….….….….….….….….….….….….….….….….….…133-134
vessels .….….….….….….….….….….….….….….….….….….….…133
packed-bed type .….….….….….….….….….….….….….….….…134-135
short-bed systems .….….….….….….….….….….….….….….….….…135
Deposits
boiler water .….….….….….….….….….….….….….….….….….…57-60
cooling water .….….….….….….….….….….….….….…173-177, 181-184
Diethylaminoethanol
boiler water treatment .….….….….….….….….….….….….….….…20-21
distribution ratio .….….….….….….….….….….….….….….….….….…22

Disodium phosphate .….….….….….….….….….….….….….….…61-63, 65

Dissolved oxygen
chemical control of .….….….….….….….….….….….….….….….…13-23
corrosive effects of .….….….….….….….….….….….….….….….…3, 6-8
mechanical control of .….….….….….….….….….….….….….…10-13, 80
Distribution ratio
of neturalizing amines .….….….….….….….….….….….….….….….…21
Drift from cooling towers .….….….….….….….….….….….….….…162-163
Duplex alloys .….….….….….….….….….….….….….….….….….….….…6
Early condensate in steam turbines .….….….….….….….….….….….….…92
EDTA .….….….….….….….….….….….….….….….….….….….….…67-68
Electrodeionization .….….….….….….….….….….….….….….….…147-148
Electrodialysis .….….….….….….….….….….….….….….…2, 136, 146-148
Electrodialysis reversal .….….….….….….….….….….….….…136, 146-148
Equilibrium phosphate treatment (EPT) .….….….….….….…61, 66-67, 100
Erythorbic acid as an oxygen scavenger .….….….….….….….….….….…18
Ethylene diamine tetraacetic acid (EDTA) .….….….….….….….….….…67-68
Exfoliation
reheater tubes .….….….….….….….….….….….….….….….….….…100
superheater tubes .….….….….….….….….….….….….….….….….….…93
Feedwater heaters .….….….….….….….….….….….….….….…5, 6, 17, 18
Ferric oxide hydrate, product of oxygenated treatment .….….….….….….…23
Filming amines .….….….….….….….….….….….….….….….….….…20-21
Filtration
activated carbon .….….….….….….….….….….….….….….….….….…106
guidelines for selection of multi-media .….….….….….….….….….…150
multi-media .….….….….….….….….….….….….….….….….….….…105
sidestream treatment of cooling water .….….….….….….….…182, 199-200
Flocculation .….….….….….….….….….….….….….….….….….…103-104
Flow accelerated corrosion (FAC) .….….….….….….….….….…2, 15, 19, 20
Foaming in boiler drums .….….….….….….….….….….….….….….….…89
Fouling and control in cooling water systems (see Cooling Water Chemistry,
Fouling)
Fungi .….….….….….….….….….….….….….….….….….….….….….…184
Hardness
effect on weak acid cation exchanger performance .….….….….….….…124
Heat recovery steam generators (HRSG) .….….….….….….….….…2-3, 70-72
High Purity Makeup Water Treatment .….….….….….….….….….…101-153
(see also Ion Exchange, Reverse osmosis,
and Electrodialysis)
electrodialysis and electrodialysis reversal .….….….….….….….…146-148

ion exchange .….….….….….….….….….….….….….….….….…106-135

pretreatment .….….….….….….….….….….….….….….….….…101-106
reverse osmosis .….….….….….….….….….….….….….….….…136-146
Hideout (of phosphate in boiler water) .….….….….….….….…63, 64-66, 72
Hot lime softening .….….….….….….….….….….….….….….….….….…106
Hydantoins .….….….….….….….….….….….….….….….….….…186, 188
Hydrazine
breakdown products .….….….….….….….….….….….….….….….….…15
feed points .….….….….….….….….….….….….….….….….….….….…16
feed systems .….….….….….….….….….….….….….….….….….…17-18
metal passivator .….….….….….….….….….….….….….….….….…14-15
oxygen scavenger .….….….….….….….….….….….….….….….…2, 14-16
safety concerns .….….….….….….….….….….….….….….….….….…16
Hydrogen damage of boiler tubes .….….….….….….….….….….….…58-59
Hydroquinone
catalyst for hydrazine .….….….….….….….….….….….….….….….…16
oxygen scavenger .….….….….….….….….….….….….….….….…16, 17
Hydroxyethylidene diphosphonic acid (HEDP) .….….….….….….….….…180
Hyperbolic cooling towers .….….….….….….….….….….….….….…159-160
Industrial steam chemistry .….….….….….….….….….….….….….…94-95
Ion exchange .….….….….….….….….….….….….….….….….…2, 106-135
anion .….….….….….….….….….….….….….….….….….….….….…110
backwash requirements .….….….….….….….….….….….….….…128-129
cation .….….….….….….….….….….….….….….….….….….…110, 111
costs vs. reverse osmosis .….….….….….….….….….….….….….….…142
mixed-bed .….….….….….….….….….….….….….….…110-111, 130-131
organic scavenger .….….….….….….….….….….….….….….….…38, 108
performance monitoring
strong acid cation effluent .….….….….….….….….….….….…130-131
strong base anion effluent .….….….….….….….….….….….…131-132
mixed-bed .….….….….….….….….….….….….….….….….….….…132
regeneration .….….….….….….….….….….….….….….….…111-114, 129
co-current .….….….….….….….….….….….….….….….….…112-113
countercurrent .….….….….….….….….….….….….…112-113, 119-120
mixed-bed .….….….….….….….….….….….….….….….….….….…131
rinse .….….….….….….….….….….….….….….….….….….….…129-130
sizing vessels .….….….….….….….….….….….….….….….….…126-128
sodium softening .….….….….….….….….….….….….….….….….…110
strong acid cation process
capacity calculations .….….….….….….….….….….….….….…115-121

exchange characteristics .….….….….….….….….….….….….….…111

regeneration.….….….….….….….….….….….….….…111-113, 119-121
strong base anion process
exchange characteristics .….….….….….….….….….….….….…113-114
regeneration .….….….….….….….….….….….….….….…113-114, 123
weak acid cation process
exchange characteristics .….….….….….….….….….….….….…123-124
regeneration.….….….….….….….….….….….….….….….….…114, 125
weak base anion process
exchange characteristics .….….….….….….….….….….….….….…125
regeneration.….….….….….….….….….….….….….….….….…114, 126
Ion exchange resins
exchange sites .….….….….….….….….….….….….….….….….….…107
strong acid cation .….….….….….….….….….….….….….….….….…108
sodium form .….….….….….….….….….….….….….….….….….…110
strong base anion .….….….….….….….….….….….….….….….….…109
weak acid cation .….….….….….….….….….….….….….….….…109, 129
weak base anion .….….….….….….….….….….….….….…109-110, 129
structure .….….….….….….….….….….….….….….….….….…107-108
Iron
Iron oxide
formation in conventionally-treated boilers .….….….….….….….…6-7, 55
formation in oxygenated-treatment programs .….….….….….….….….…23
deposition in boilers .….….….….….….….….….….….….….…23, 55-56
mechanical carryover to turbines .….….….….….….….….….…86, 93-94
test for during boiler startup .….….….….….….….….….….….….…55-56
Isocyanurates .….….….….….….….….….….….….….….….….….….…186
Isothiozolone .….….….….….….….….….….….….….….….….….….…190
Langelier Saturation Index (LSI) .….….….….….….….….….….….…173-176
Layup of steam generating systems .….….….….….….….….….….….…72-73
Legionnaire’s Disease .….….….….….….….….….….….….….….….….…184
Lime/soda ash softening .….….….….….….….….….….….….….….….…104
Macrofouling of cooling water
Asiatic clams .….….….….….….….….….….….….….….….….…190-191
zebra mussels .….….….….….….….….….….….….….….…190, 191-193
control methods .….….….….….….….….….….….….….….…192-193
Magnesium

scale- or corrosion-forming compounds .….….….….….….….…3, 57, 78

Magnesium silicate scale in cooling water systems .….….….….….….….…177
Magnetite .….….….….….….….….….….….….….….….….…6-7, 14, 23, 55
Makeup water pre-treatment .….….….….….….….….….….….….…100-106
aeration .….….….….….….….….….….….….….….….….….….….…106
activated carbon filtration .….….….….….….….….….….….….…102, 106
clarification/softening .….….….….….….….….….….….….….….….…102
hot lime softening .….….….….….….….….….….….….….….….….…106
media filtration .….….….….….….….….….….….….….….….…102, 105
microbiocide control .….….….….….….….….….….….….….….…102-103
Manganese .….….….….….….….….….….….….….….….….….….….…106
Manganese greensand .….….….….….….….….….….….….….….….….…106
Mechanical draft cooling towers .….….….….….….….….….….….…157-159
Mechanical carryover .….….….….….….….….….….….….….….….…88-89
Methyl ethyl ketoxime
metal passivator .….….….….….….….….….….….….….….….….….…17
oxygen scavenger .….….….….….….….….….….….….….….….…16, 17
Microbiocides
bromine .….….….….….….….….….….….….….….….….….….….…187
chlorine .….….….….….….….….….….….….….….….….….….…184-186
chlorine dioxide .….….….….….….….….….….….….….….….…188-189
hydantoins .….….….….….….….….….….….….….….….….….…186, 188
sodium hypochlorite.….….….….….….….….….….….….….….…186, 200
isocyanurates.….….….….….….….….….….….….….….….….….….…186
non-oxidizing .….….….….….….….….….….….….….….….….…189-190
ozone .….….….….….….….….….….….….….….….….….….….….…189
Microorganisms
control .….….….….….….….….….….….….….….….….….….…184-190
effect on RO membranes .….….….….….….….….….….….….….….…139
problems in cooling water systems .….….….….….….….….…3, 183-184
Mixed-bed ion exchangers .….….….….….….….….….….….….….…110-111
Molybdate .….….….….….….….….….….….….….….….….….….…170-172
Monel .….….….….….….….….….….….….….….….….….….….….….…6
Morpholine .….….….….….….….….….….….….….….….….….….….…20
Natural draft cooling towers .….….….….….….….….….….….….…159-160
Neutralizing amines.….….….….….….….….….….….….….….….….…20-21
Neutral water treatment (NWT) (see oxygenated treatment)
Nitrilotriacetic acid .….….….….….….….….….….….….….….….….…67-68
Octadecylamine .….….….….….….….….….….….….….….….….….….…21
Off-line condensate/feedwater contamination .….….….….….….….…34, 48
Off-line corrosion .….….….….….….….….….….….….….….….…20, 72-73

Once-through boilers .….….….….….….….….….….….….….….….….…55

Once-through cooling systems .….….….….….….….….….….….…156-157
On-line steam/water chemistry monitoring .….….…78-79, 207-210, 222-223
Organic acids
corrosive effects of .….….….….….….….….….….….….…2, 14, 15, 87-88
produced by decomposition of water treatment chemicals 16-17, 18, 21, 70
Organics
removal by ion exchange .….….….….….….….….….….….….….….…110
steam carryover products .….….….….….….….….….….….….….…87-88
Oxidation reduction potential (ORP)
measurement for feedwater chemistry control .….….….….….….…19, 20
Oxygen (use in oxygenated-treatment programs) .….….….….….….…22-23
Oxygen corrosion .….….….….….….….….….….….….….….….….…3, 6-8
Oxygen scavenging .….….….….….….….….….….….….….….….….…13-23
industrial steam generating systems .….….….….….….….….….….….…18
Oxygenated treatment.….….….….….….….….….….….….…2, 21-23, 70, 79
Ozone.….….….….….….….….….….….….….….….….….….….….….…189
Passivation of metal surfaces.….….….….….….….….….….….…14-15, 16-17
pH
boiler water recommendations .….….….….….….….….….…58, 60, 62-64
control guidelines in condensate/feedwater systems .….….….….….….…20
effect on corrosion of steel .….….….….….….….….….….….….….….…8
effect on RO membranes .….….….….….….….….….….….….….….…139
Phosphates in boiler water treatment .….….….….….….….….….….…61-67
coordinated treatment .….….….….….….….….….….….….….….…62-64
congruent treatment .….….….….….….….….….….….….….….….…62-64
equilibrium phosphate treatment (EPT).….….….….….….….….….…66-67
hideout .….….….….….….….….….….….….….….….….….….….…64-66
phosphate treatment (PT) .….….….….….….….….….….….….….….…67
Phosphate
cooling water treatment .….….….….….….….….….….….….….…170-172
Phosphonates
cooling water treatment .….….….….….….….….….….….…170-172, 180
Phosphono-butane-tricarboxylate (PBTC) .….….….….….….….….….…180
Polyacrylamides .….….….….….….….….….….….….….….….….…180, 183
Polyacrylate .….….….….….….….….….….….….….….….….….…180, 183
Polymaleates.….….….….….….….….….….….….….….….….….….….…180
Practical Scale Index .….….….….….….….….….….….….….….….….…175
Pretreatment of makeup water .….….….….….….….….….….….…101-106
Priming in boiler drums .….….….….….….….….….….….….….….…88-89
Programmable logic controller (PLC) .….….….….….….….….….….…78, 79

Quaternary amines .….….….….….….….….….….….….….….….…104, 192

Reheater tube failures .….….….….….….….….….….….….….….….….…100
Reverse osmosis (RO) .….….….….….….….….….….…2, 99, 105, 136-146
alarms .….….….….….….….….….….….….….….….….….….….….…145
antiscalants .….….….….….….….….….….….….….….….….….….…140
beta factor .….….….….….….….….….….….….….….….….….…144-145
design parameters .….….….….….….….….….….…140-142, 144, 145-146
flow rate monitoring .….….….….….….….….….….….….….….…143-144
flux rate .….….….….….….….….….….….….….….….….….….….…141
fouling .….….….….….….….….….….….….….….….….….….…139-140
hollow fiber elements .….….….….….….….….….….….….….….….…137
membrane cleaning.….….….….….….….….….….….….….….….….…146
membrane design .….….….….….….….….….….….….….….….….…137
membrane flux rates .….….….….….….….….….….….….….….….…141
membrane material .….….….….….….….….….….….….….….….….…137
cellulose acetate.….….….….….….….….….….….….….….….…137-139
thin-film composite .….….….….….….….….….….….….….…137, 139
performance monitoring.….….….….….….….….….….….….….…144-145
conductivity .….….….….….….….….….….….….….….….….….…144
flow .….….….….….….….….….….….….….….….….….….….….…144
temperature .….….….….….….….….….….….….….….….….….…144
permeate .….….….….….….….….….….….….….….….….….….….…137
pH effects upon membranes .….….….….….….….….….….….…139, 144
pretreatment .….….….….….….….….….….….….….….….….….….…139
reject .….….….….….….….….….….….….….….….….….….….….…137
sampling of process streams .….….….….….….….….….….….….….…211
salt rejection .….….….….….….….….….….….….….….….….….….…141
scaling .….….….….….….….….….….….….….….….….….….….….…140
spiral wound membranes .….….….….….….….….….….….….…137-138
system components
piping .….….….….….….….….….….….….….….….….….….….…143
pressure vessels .….….….….….….….….….….….….….….….….…143
pumps .….….….….….….….….….….….….….….….….….…142-143
temperature effects on output .….….….….….….….….….….….….…144
Ryznar Stability Index (RSI) .….….….….….….….….….….….….…174-175
Sampling .….….….….….….….….….….….….….….….….….….… 205-228
Data acquisition .….….….….….….….….….….….….….….….….…222-223
Representative sampling .….….….….….….….….….….….….….…214-222
sample conditioning
primary cooling .….….….….….….….….….….….….….….…217, 219
final conditioning .….….….….….….….….….….….….….….…220-222

sample flow
flow rate recommendations .….….….….….….….….….….….…219-220
line size .….….….….….….….….….….….….….….….….….…219-220
sample nozzles
design of liquid sample nozzles .….….….….….….….….….….…215-216
design of steam sample nozzles .….….….….….….….….….….…216-218
installation recommendations .….….….….….….….….….….….….…218
sample point location to prevent flow interferences.….….….….….….…218
Sample parameters
ammonia .….….….….….….….….….….….….….….….….…33, 207-211
cation conductivity .….….….….….….….….….….….….….…33, 207-211
chloride .….….….….….….….….….….….….….….….….….….…207-211
conductivity .….….….….….….….….….….….….….….….….….….…33
copper .….….….….….….….….….….….….….….….….….….…207-211
degassed cation conductivity .….….….….….….….….….….….…207-211
dissolved oxygen .….….….….….….….….….….….….….….…33, 207-211
iron .….….….….….….….….….….….….….….….….….….….…207-211
oxygen scavengers .….….….….….….….….….….….….….…33, 207-211
pH .….….….….….….….….….….….….….….….….….….…33, 207-211
phosphate .….….….….….….….….….….….….….….….….…78, 207-211
silica .….….….….….….….….….….….….….….….….….…207-211, 227
sodium .….….….….….….….….….….….….….….….…33, 48, 207-211
specific conductivity .….….….….….….….….….….….….….….…207-211
sulfate .….….….….….….….….….….….….….….….….….….…207-211
total organic carbon (TOC) .….….….….….….….….….….….….…33, 38
Sample point selection
boiler water .….….….….….….….….….….….….….….…72, 207-211, 213
condensate pump discharge (CPD) .….….….….….…33, 207-211, 211-212
condensate polisher outlet (CPO) .….….….….….….….…33, 207-211, 212
condensate storage tank effluent.….….….….….….….….….….….….…211
deaerator inlet (DAI) .….….….….….….….….….….….…33, 207-211, 212
deaerator outlet (DAO) .….….….….….….….….….….…33, 207-211, 212
economizer inlet/feedwater .….….….….….….….…33, 207-211, 212-213
guidelines for various types of steam generating systems .….….…207-211
makeup system effluent .….….….….….….….….….….…131-132, 206-211
main steam .….….….….….….….….….….….….….….…94, 207-211, 214
reheat steam .….….….….….….….….….….….….…94, 100, 207-211, 214
saturated steam .….….….….….….….….….….….…94, 207-211, 213-214
Saturated steam .….….….….….….….….….….….….….….….….….….…94
Scale
control in boilers .….….….….….….….….….….….….….….….….…60-73

control in cooling water systems .….….….….….….….….….….…177-181

formation in boilers .….….….….….….….….….….….….….….….…57-60
formation in cooling water systems .….….….….….….….….….…177-181
Sidestream filtration of cooling tower water .….….….….….….…182, 199-200
Silica
collodial .….….….….….….….….….….….….….….….….….….…90, 99
effect on boiler startups .….….….….….….….….….….….….….….…227
leakage from strong base anion exchangers .….….….….….…122, 131-132
removal by clarification/softening processes .….….….….….….….….…104
removal in strong base anion exchangers .….….….….….….….…114, 122
scaling in boilers .….….….….….….….….….….….….….….….….….…3
scaling in cooling water systems .….….….….….….….….….….….….…177
scaling of RO membranes .….….….….….….….….….….….….….….…140
sources of contamination in boilers .….….….….….….….….…34, 48, 99
steam carryover product .….….….….….….….….….….….….…84, 87, 90
Silt Density Index (SDI) .….….….….….….….….….….…139-140, 152-153
BASIC program .….….….….….….….….….….….….….….….….….…153
formula .….….….….….….….….….….….….….….….….….….….…152
Sodium
leakage from strong acid cation exchangers .….….…116-119, 120, 123, 131
measurement for monitoring condenser cooling water inleakage 33-34, 48
Sodium bisulfite
for removal of oxidizing biocides from makeup water .….….….….….…103
Sodium hydroxide
corrodent of boiler tubes .….….….….….….….….….….….….….…58, 62
ion exchange regenerant (see Ion exchange)
steam turbine carryover product .….….….….….….….….….….….….…85
Sodium hypochlorite .….….….….….….….….….….….….….….….….…186
guidelines for ordering and storage .….….….….….….….….….….….…200
Sodium-to-phosphate ratio in boiler water .….….….….…62-64, 75-76, 78-79
BASIC program to calculate ratior .….….….….….….….….….….…75-77
Sodium phosphates
boiler water treatment .….….….….….….….….….….….….…61-67, 75-77
steam carryover products .….….….….….….….….….….….….…87, 100
Sodium salts
steam carryover products .….….….….….….….….….….….….….…84, 85
Sodium softening .….….….….….….….….….….….….….….….….….…110
Sodium sulfite

breakdown products in boilers .….….….….….….….….….….….….…14

for removal of oxidizing biocides from makeup water .….….….….….…103
oxygen scavenger.….….….….….….….….….….….….….….….….…13-14
Solid particle erosion (SPE) of turbine blades .….….….….….….…91, 93-94
Stainless steel .….….….….….….….….….….….….….….….….….….….…6
Steam Chemistry .….….….….….….….….….….….….….….….….…83-95
boiler water guidelines to maintain proper steam chemistry .….….…92-93
carryover .….….….….….….….….….….….….….….….….….….…83-91
mechanical .….….….….….….….….….….….….….….….….….…88-89
vaporous .….….….….….….….….….….….….….….….….….….….…90
chloride and sulfate corrosion of turbine blades
contaminant introduction from attemperators .….….….….….….…91, 93
copper .….….….….….….….….….….….….….….….….…17, 84-85, 90
deposition patterns .….….….….….….….….….….….….….….….…84-88
organics .….….….….….….….….….….….….….….….….….….…82, 88
silica .….….….….….….….….….….….….….….….….….….….…87, 90
sodium phosphates .….….….….….….….….….….….….….….….….…87
superheater exfoliation and solid particle erosion .….….….….…91, 93-94
Steam turbine purity limits .….….….….….….….….….….….….….…90, 92
Strontium sulfate scale
on RO membranes .….….….….….….….….….….….….….….….….…140
Stress corrosion cracking (SCC)
turbine blades .….….….….….….….….….….….….….….….….…69, 85
Sulfate
corrosive effects on turbine blades .….….….….….….….….….….….…86
steam carryover product .….….….….….….….….….….….….….…84, 86
Sulfuric acid
cooling water treatment.….….….….….….….….….….…178-179, 198-199
regenerant for cation exchange resins .….….….….….….….…111-112, 129
specifications .….….….….….….….….….….….….….….….….…151-152
Sulfonated styrene
for boiler water treatment .….….….….….….….….….….….….….….…68
ion exchange resin base .….….….….….….….….….….….….….…107-108
Superheaters .….….….….….….….….….….….….….….….…36, 91, 93-94
Suspended solids
effect on reverse osmosis membranes.….….….….….….….….….…139-140
in cooling water .….….….….….….….….….….….….….….….…181-182
control .….….….….….….….….….….….….….….….….….…182-183
Tank volume calculations – BASIC programs .….….….….….….….….…230

Titanium.….….….….….….….….….….….….….….….….….….….….….…6
Total Organic Carbon (TOC)
effects on boiler water and steam chemistry .….….….….….….….…98-99
in condensate return .….….….….….….….….….….….….….….….….…38
Trisodium phosphate for boiler water treatment .….….….…61-63, 65, 66, 78
Turbidity .….….….….….….….….….….….….….….….….….….….….…105
Ultraviolet light for makeup treatment .….….….….….….….…102, 149-150
Underdeposit corrosion
in boilers .….….….….….….….….….….….….….…55, 56, 58-59, 67, 77
in cooling water systems .….….….….….….….….….….….….….….…168
Unit conversion program (COADE) .….….….….….….….….….….….…229
Vaporous carryover .….….….….….….….….….….….….….….….….….…90
Wet-bulb temperature .….….….….….….….….….….….….….….…163-165
Zebra mussels .….….….….….….….….….….….….….….….…186, 190-193
Zinc .….….….….….….….….….….….….….….….….….….….….…171-172


Power Plant Water Chemistry A Practical Guide

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Power Plant Water Chemistry A Practical Guide

Uploaded by

Copyright:

Available Formats

Ali Sadeghi Digital Library

00fm.qxd 8/8/06 1:49 PM Page i

Ali Sadeghi Digital Library

Ali Sadeghi Digital Library

Condensate/Feedwater Chemistry iii

Ali Sadeghi Digital Library

Director: Mary McGee

Library of Congress Cataloging-in-Publication Data

Printed in the United States of America.

Ali Sadeghi Digital Library

IV.indd 1 8/24/08 12:00:05 PM

Ali Sadeghi Digital Library

Ali Sadeghi Digital Library

1 INTRODUCTION TO STEAM GENERATION WATER

2 CONDENSATE FEEDWATER CHEMISTRY 5

Ali Sadeghi Digital Library

viii Table of Contents

Supplement 2-1: Why Condenser Performance is Important 41

3 BOILER WATER CHEMISTRY 51

Ali Sadeghi Digital Library

Chloride and Sulfate 86

5 HIGH-PURITY MAKEUP WATER TREATMENT 101

Ali Sadeghi Digital Library

Strong Base Anion Exchanger Calculations 121

6 COOLING WATER CHEMISTRY 155

Ali Sadeghi Digital Library

Cooling Tower Calculations 160

Ali Sadeghi Digital Library

Feedwater or Economizer Inlet 212

Ali Sadeghi Digital Library

Ali Sadeghi Digital Library

Ali Sadeghi Digital Library

Steam generation, whether it be for power production or industrial process

The primary water systems in a steam plant include:

● Closed cooling water

● Ash sluicing at coal-fired plants

The contrasting nature between differing systems requires water of varying

Ali Sadeghi Digital Library 1

2 Power Plant Water Chemistry: A Practical Guide

Ali Sadeghi Digital Library

Introduction to Steam Generation Water Chemistry Systems 3

Ali Sadeghi Digital Library

Ali Sadeghi Digital Library

Ali Sadeghi Digital Library 5

6 Power Plant Water Chemistry: A Practical Guide

FEEDWATER HEATERS FEEDWATER HEATERS

Condensate Chemistry—Dissolved Oxygen

Ali Sadeghi Digital Library

3Fe + 4H2O → Fe3O4 + 4H2 ↑ (2.1)

Magnetite is very dark, and in a properly treated system will appear as a

2Cu + H2O → Cu2O + H2 ↑ (2.2)

2Fe3O4 + ¹/₂O2 → 3Fe2O3 (2.3)

Fe2O3 is rust, and, like rust produced by atmospheric corrosion of iron, is

Cu2O + ¹/₂O2 → CuO (2.4)

In the presence of a complexing agent such as ammonia, the oxidized cop-

Cu+2 + 4NH3 → Cu(NH3)4+2 (2.5)

Ali Sadeghi Digital Library

8 Power Plant Water Chemistry: A Practical Guide

Oxygen is an aggressive corrodent. Control of inleakage and removal of dissolved

Carbonic Acid Corrosion in

HCO3- + heat → CO2↑ + OH- (2.6)

CO2 + H2O ⇔ H2CO3 (2.7)

Fe + 2H2CO3 → Fe(HCO3)2 + H2↑ (2.8)