Berthold Otzisk, Kurita Europe GmbH, Germany,
discusses chemical treatment of an ethylene plant.
W
ithout use of high quality plastics it is
difficult today to imagine modern life.
Ethylene is the most produced organic
compound with a production of
approximately 146 million t in 2014. The commercial
production is based on thermal cracking of petroleum
hydrocarbons with steam, called steam cracking or pyrolysis.
To date more than 265 operational ethylene production
units have been installed. These ethylene units are very
complex plants including cracking, cooling, compression and
separation steps. Several licensors offer different designs and
technologies. In general polymer grade ethylene is produced.
Major byproducts are propylene, butadiene rich C4 streams,
C6-C8 aromatic rich pyrolysis gasoline, hydrogen and fuel
streams.
The production of ethylene and its byproducts is a result
of a free radical reaction. This mechanism does not explain the
complete product distribution, but is very helpful to
understand the formation of ethylene. The radical
polymerisation occurs in three steps:
n Initiation (split into two radicals).
n Propagation (radical reaction with chain transfer).
n Termination (reaction with another radical).
The radical decomposition is one of the most important types
of side reactions as it directly produces ethylene:
R-CH2-CH2-CH2* → R-CH2* + C2H4
The cracked gases leave the cracking furnace at 750 - 875°C
and have to be cooled immediately to preserve their
HYDROCARBON
ENGINEERING Reprinted from April 2015

Figure 1. Block flow diagram naphtha cracker (Linde design).
composition. Transfer line exchangers (TLX or TLE), primary
fractionators (oil quench column) and quench water treatment
play a fundamental role in reducing the temperature.
Corrosion and fouling are significant operational problems,
leading to reduced run time and reduced yields. Figure 1 shows
typical process areas where chemical treatment can help to
increase the ethylene production and run time.
Corrosion control
Neutralising amines, filming amines and oxygen scavengers are
commonly used as corrosion inhibitors to significantly reduce
the corrosion potential. The transfer line exchangers and
dilution steam generation system (DSG) require oxygen free
feed water. Oxygen scavengers will significantly lower the free due to oil/tar fouling, unplanned mechanical cleaning of the
oxygen to low ppb levels.
Organic acids, hydrogen sulfide and low water pH are
corrosion initiators. Directly affected are the overhead systems
of the primary fractionator and water quench column, quench
water separator, process water stripper, DSG blow down,
dilution steam and compressor wash water system (if applied
instead of wash oil). It is still common practice at many
ethylene units to inject a diluted caustic (NaOH) for pH
control. Caustic is a strong base, very cheap and easily
purchased. Disadvantages are the fluctuating concentrations,
when manually diluted with water and the high risk of
unwanted sodium carry over into the dilution steam. The pH
adjustment with diluted caustic requires sufficient practical
experience because of fluctuating concentrations. A perfect
pH control is much easier to realise with suitable amine
blends. For economical reasons it is possible to inject a
combination of caustic and neutralising amine blends to
increase the alkalinity. In that case a constant base
concentration of caustic is injected and the pH is controlled
by additional dosing of the neutralising amine in relation to
the desired pH value. This will help to avoid pH fluctuations.
Modern neutralising amines operate by reacting with any acid
constituent in a straight forward chemical neutralisation. These cracked gas.
active neutralising amine blends form liquid salts, which can be
removed with water. Very effective neutralising amines such as
HYDROCARBON
Reprinted from April 2015 ENGINEERING
KURITA CI-6101 are blends
of high boiling and volatile
amine components. The
high boiling amines will stay
in the water phase to keep
the pH in the steam
generators at the desired
pH level.
When corrosion is
observed in the quench oil
exchangers or air fin
coolers, oil soluble filming
amines will provide perfect
corrosion protection. The
reason for severe corrosion
attack is in many cases the
presence of high organic
acids in the cracked gas
such as formic or acetic
acids. The filming amine will
form a protective layer on
the metal surface, which reduces the corrosion potential
significantly. The use of a hydrocarbon dilution slipstream for
better distribution of the filming amine into the system is
recommended.
Improved oil/water separation
Based on design, the collected quench water from the water
quench column is separated from heavy hydrocarbons in the
oil/water separator. Many times the separated quench water
still contains higher amounts of hydrocarbons, which are
dispersed in the aqueous phase. In case of poor phase
separation in the oil/water separator sporadic loss of quench
water can be observed. Negative impacts are yield reduction
recycle quench water heat exchangers, DSG system or change
out of DOX filter media.
A capable emulsion breaker program such as KURIFLOCK
LC-9610 can be applied to improve the separation of
hydrocarbons and water. Overdose of the emulsion breaker
should be avoided as it may have the tendency to act as an
emulsifier at very high concentrations. A simple beaker test
will provide good information to check, if enough emulsion
breaker was added.
Coke fouling
Fouling is the most frequently observed reason for unwanted
plant shutdowns, unplanned mechanical cleaning and serious
losses in ethylene production. Almost all equipment parts of
the ethylene unit are affected. The first place, where fouling is
observed, are the cracking furnace tubes, the transfer line
exchangers and primary fractionator column and its oil circuit.
Most of the time fouling is related to pyrolytic coke formation.
It is associated with the presence of active sites on the metal
surface. Particles of nickel, iron and chromium or alloys
containing these metals catalyse the formation of filamentous
carbon. The formed coke particles are transported with the
The continuous injection of a sulfiding agent is the
classical method for coke reduction and controlling carbon
monoxide formation. These sulfiding agents form sulfidic
surfaces on the coil surface, blocking the active sites on the
metal surface. Kurita has many years of experience with the
injection of polysulfides at steam crackers and promotes the
KURITA CUT-COKE technology based on Di-tert-butyl
polysulfide (TBPS), which presents a number of advantages
compared to conventional sulfiding agents such as DMS or
DMDS.

Polymer fouling
The cracked gas contains very reactive compounds such as
ethylene, acetylene propylene, butadiene, carbon monoxide,
hydrogen, styrene, etc. Trace amounts of oxygen or oxygen
containing compounds promote polymer or gum formation
and increase the risk of fouling. Butadiene fouling is a common
problem in olefin units, where formation of ‘popcorn’ polymer
is observed. It is a highly reactive molecule and the
dimerisation of two butadiene molecules will make a C8
compound or higher mole weight compounds. This reaction
occurs in the liquid and vapour phase. If oxygen or metal
oxides are present, ‘gums’ formation as a result of unwanted
free radicals polymerisation may also occur.
Deethaniser and depropaniser columns and its reboilers
are often suffering with severe polymer fouling. In many cases
depropaniser fouling is related to butadiene fouling in the
reboiler. Many olefin units have spare deethaniser or
depropaniser reboilers to allow a 3 - 5 year runtime with a
reboiler cycle time of 6 - 18 months. If no fouling inhibitors are
used, fouling of the lower trays in the depropaniser column
can limit the production capacity or can reduce the
equipment lifetime. Effective fouling inhibitors do not prevent
the growth of ‘popcorn’ or gum formation, but will inhibit its
formation. These antifoulants are generally blended chemicals
that may contain one or more of the following active
components:
n Antioxidants.
n Antipolymerants.
n Dispersants.
n Metal deactivators.
Antioxidants react with the peroxide radicals, which are
formed, when oxygen reacts with hydrocarbons. The
antipolymerant stops the initiation or propagation reactions.
Dispersants keep the particles of polymerised material in a
suspended state by preventing their agglomeration and
precipitation. Coke suppressants reduce the catalytic effect of
the hot metal, which could initiate coke particle formation.
Metal deactivators react with metal ions such as Fe2+ or Cu2+
by preventing oxidation and polymerisation reactions.
After cooling and purification of the cracked gases in the
primary fractionator and water quench column further
processing requires compression. It is typically performed by a
turbine driven centrifugal compressor in four to five stages
with intermediate cooling for condensation of water and
hydrocarbons. The cracked gas compressor (CGC) is an
essential equipment part. Polymer fouling in the compressor
can lead to unplanned shutdown and total production loss of
the plant.
For polymerisation and fouling prevention it is mandatory
to keep the cracked gas temperature as low as possible during
Figure 2. Impeller of a cracked gas compressor
(CGC).
the compression steps. As a rule of thumb the polymerisation
rate doubles per 10 Kelvin and the ethylene plants keep the
maximum cracked gas temperature below 100°C.
Wash oil (i.e. aromatic rich oil) or wash water is applied to
reduce fouling of compressor internals and intercooling
equipment. The use of a very efficient antifoulant programme
such as KURITA AP-2402, injected in low ppm dosing rates into
the wash oil or wash water, will significantly reduce the fouling
potential. Figure 2 shows the treated impeller of a cracked gas
compressor after opening. It was running more than five years
at full capacity without any interruptions.
Aldol condensation
The cracked gas contains contaminants like acid gases such as
hydrogen sulfide and carbon dioxide or peroxides, carbonyls
and aldehydes, affecting product quality and further
processing. Hydrogen sulfide and carbon dioxide can be
removed by once through caustic wash and regenerative
solvent scrubbing. The acid gas removal equipment must be
installed upstream of the drying section to avoid formation of
hydrates and ice in the following fractionation steps (cold
section). A caustic scrubber is typically located between the
3rd and 4th compressor stages, or between the 4th and 5th
stages. The caustic scrubber is operating with a cascading flow
of a NaOH solution, reaching its saturation on formed
carbonates and sulfides in the bottom section. The NaOH
liquid is removed at the knock out drum to be routed back to
the caustic scrubber. Bottoms from the scrubber (spent
caustic) are sent to the wastewater treatment plant. The spent
caustic is usually oxidised and neutralised in a wet caustic

HYDROCARBON
ENGINEERING Reprinted from April 2015
 oxidation unit or it is neutralised and stripped,
before it is routed to the sewer system.
Caustic scrubbers are frequently subject
to fouling and major problems are:
n Loss of acid gas removal efficiency.
n Increased deltaP over the absorber
sections.
n Need for gasoline/steam wash.
n High gasoline consumption for polymer
extraction in spent caustic.
n High organic load in wastewater.
Two different fouling mechanisms have
been identified, which are primarily aldol
condensation and radical polymerisation
reaction of dienes. Aldehydes and ketones
can form resinous polymers through Aldol
condensation under alkaline conditions. Figure 3. Aldol
Acetaldehyde (ethanal) will form acetaldol condensation in a
(3-hydroxybutanal). Further reaction at caustic solution.
higher temperature leads to the formation
of crotonaldehyde and hexa-2,4-dienal. Finally
polyacetaldols are formed and the first indication is a
colour change of the caustic solution. ‘Yellow oil’ and ‘red
oil’ are usual insider expressions indicating an aldol fouling.
HYDROCARBON
Reprinted from April 2015 ENGINEERING
Figure 3 shows a coloured caustic sample,
where Aldol condensation already started.
The aldol condensation can be completely
stopped by using a special antifoulant such as
KURITA AP-2501. This antifoulant is a carbonyl
scavenger, which reacts with aldehydes and
ketones to form water soluble alcohols. This
prevents polymerisation to heavier molecules
and eliminates the need for frequent skimming
of red oils. It noticeably reduces the foaming
potential and entrainment of caustic and
amine solutions.
Conclusion
The production of ethylene by steam
cracking is a complex procedure including
cracking, cooling and separation steps.
Corrosion and fouling which can generate
very high production losses are monitored
regularly at many ethylene units. The use of
efficient chemical programmes such as corrosion
inhibitors, antioxidants or antifoulants can considerably
help, reaching the defined targets and extension of the
runtime.
 MITIGATING
COKE
FORMATION
Berthold Otzisk and Christoph Eichman, Kurita Europe GmbH, Germany, discuss
methods for reducing coke formation in ethylene production.
T
he leading technology for the production of
ethylene is based on thermal cracking of gaseous
and liquid hydrocarbons. Typical gaseous
feedstocks are ethane, propane and butane.
Common liquid feedstocks are naphtha, gas oil and
hydrocracker residue. In the presence of steam they are
fed to the cracking furnace. This gas phase reaction is
called pyrolysis or steam cracking. The production of
ethylene is a very complex process including cracking,
cooling, compression and separation steps.
The cracking furnace is the heart and starting point of
the ethylene production. The hydrocarbon stream is
preheated in the convection zone (500 – 680 °C) of the
furnace, where it is mixed with steam. Steam dilution
lowers the hydrocarbon partial pressure of the light
gaseous, cracked compounds to reduce their
recombination and polymerisation reactions. It also
lowers the partial pressure of high boiling and high
molecular mass aromatics, which reduces their tendency
to deposit and minimises coke deposition on the furnace
tubes. This favours the formation of primary reaction
products and reduces the decoking time of the furnace.
Primary reaction products are ethylene, propene,
acetylene, hydrogen and methane. The hot gases leave the
convection zone and enter the radiant zone (750 – 875 °C).
The hot cracked gases are cooled down immediately in
the transfer line exchangers (TLX, TLE) to avoid undesired
side reactions of the reactants. The transfer line
exchanger is located very close to the cracking furnace
outlet.
Coke fouling
Coking is a side reaction from thermal cracking. It is a major
operational problem in the radiant section of steam
cracking furnaces and transfer line exchangers. The coke
formation is a very complex mechanism. In general, there
are four mechanisms, which describe the coke formation:
n Free radical mechanism.
n Catalytic mechanism.
n Condensation reactions of heavy polynuclear
compounds.
n Mechanical entrainment of coke particles.
HYDROCARBON
ENGINEERING Reprinted fro m July 2014
 Gaseous free radicals form pyrolytic coke, which
deposits on the radiant coil and transfer line exchanger.
Catalytic coke formation is associated with the presence of
active sites on the metal surface. Particles of nickel, iron and
chromium or alloys containing these metals catalyse the
formation of filamentous carbon. The carbon deposits
displace the metal particles and dissolve into the metal
under high temperature. Condensation reactions occur at the
metal surface or coke surface. High boiling hydrocarbons
condense at the interface, where they dehydrogenate to
form coke. Coke particles, formed in the radiant section, can
be detached and entrained to the transfer line exchanger,
where they accumulate. Coke fouling in the transfer line
exchanger is a severe problem as the outlet temperature
rises and less heat is recovered for steam generation.
The coke build up results in a negative impact on
product yield, reduction of heat transfer efficiency and
increased energy consumption. The external tube skin
temperature continuously rises while the pressure drop
increases. This influences the process selectivity, declines the
Figure 1. Coke inhibitor injection into the steam.
Table 1. Comparison of sulfiding agents
Dimethyl sulfide Dimethyl disulfide
Abbreviation DMS DMDS
Sulfur content 51% 68%
Flash point, ˚C - 45 16
Thermal decomposition, ˚C > 250 > 200
Vapour pressure (20 ˚C), hPa 527 38
Reprinted from July 2014 HYDROCARBON
ENGINEERING
reaction volume and finally leads to even more rapid coke
formation. Before the external tube skin temperature reaches
its maximum material operating limit, the formed coke has to
be removed frequently by controlled combustion with steam
and air. This decoking process is a non-productive downtime
for the furnace, which negatively affects the mechanical
strength and tube life. The run length of a furnace is defined
as the operation between decoking cycles. Depending on
design and diameter of the coils, the run length may vary
between 10 – 15 days for small diameter coils and
30 – 90 days for larger diameter coils. Several methods to
mitigate coke formation are available, which extend the time
between decoking cycles.
Novel materials
Steel manufacturers are developing new metallurgies to
minimise coke formation. Several novel tube designs have
been achieved in the last years. Two interesting examples are
the spiral mixing element welded on the internal wall of the
cracking coil and the development of cast finned tubes.
Both designs improve the heat and mass transfer, which
saves fuel costs and increases the tube life as carburisation is
reduced. Cast finned tubes provide a very short residence
time, which improves the product selectivity.
Tube coating
There are a number of precoating technologies available,
which are applied on the radiant coil during manufacture. By
forming a very thin layer of diffusion barrier, catalytic coke is
reduced. Typical applications are glass ceramic coatings onto
the tube surface. In many cases this kind of precoating
technology improves the on stream time of a furnace to
some hundred days before decoking is required. Special
attention must therefore be paid to ensure that the heat
transfer does not deteriorate.
Another kind of pretreatment is the injection of an
additive after the decoking procedure, when the cracking
furnace is under steam flow. Aqueous metal salts or
proprietary phosphorus, tin and silicon based compounds are
typically used. They also work by forming a layer of diffusion
barrier. Control of the pretreatment effluents is
recommended as they can be trapped in the pyrolysis oil of
the primary fractionator or downstream processes.
A patented technology is the preliminary reduction of
the steel surface within a defined timeframe. The steel
surface is treated with steam and hydrogen or together with
inert gases such as nitrogen, helium, argon, etc., and steam.
The second step of the patented treatment procedure is the
injection of several coke inhibiting compounds to passivate
the steel surface.
Sulfiding
agents
Di-tert-butyl polysulfide The continuous
TBPS injection of a
54%
sulfiding agent is the
historical method
100
for coke reduction
> 160 and controlling
1 carbon monoxide
formation. Small amounts of sulfur are used to prevent coking
and undesired chemical reactions. These sulfiding agents are
believed to decompose to form sulfidic surfaces on the coil
surface. Others characterise the contribution of the sulfur
compounds with an improved product selectivity and
reaction speed, resulting in a decrease of the decoking
frequency.
As shown in Figure 1 the coke inhibition agent is injected
into the steam in front of the furnace. Special attention must
be paid to the fact that sulfiding agents should not be in
contact with condensed water to avoid unwanted fouling in
the steam line.
DMS and DMDS
Dimethyl sulfide (DMS) and dimethyl disulfide (DMDS) are
well known, cost effective and proven sulfiding agents for
steam cracking operations. Typical dosing rates are between
20 – 100 ppm sulfiding agent during normal operating
conditions.
DMS is an organosulfur compound with the formula
(CH3)2S. It is a flammable liquid with a flash point of -45 °C.
DMS has a sulfur content of 51 wt% and is the most abundant
sulfur compound. It has a very unpleasant, cabbage like
odour. The olfactory threshold is around 3 ppb. DMS was one
of the first sulfur bearing additives for the presulfiding
process of catalysts. Meanwhile, it plays a minor role as a
coke inhibitor for steam cracking operations.
Dimethyl disulfide (DMDS) is a very common sulfiding
agent. It has the molecular formula C2H6S2 and is the simplest
disulfide with a very unpleasant, garlic like odour. DMDS is a
stable pale yellow liquid with a low flash point of 16 °C and a
sulfur content of 68 wt%. There are also special DMDS
formulations available for purchase, where small quantities of
a perfume are added. This improves the odour of DMDS so
that it is less unpleasant for the production personnel.
Due to the strong odours and low flash points of DMS
and DMDS, special handling, storage and transportation are
mandatory. Operators are required to wear special clothing.
Storage under nitrogen pressure in closed containers is
required to avoid fire hazards.
Under standard cracking conditions the complete
decomposition temperatures of DMS and DMDS are believed
to be above 550 °C. It has been reported, that these organic
sulfides do not readily undergo gas phase decomposition and
adsorb intact on the metal surface. It has also been reported,
that DMDS can affect the quench tower acidity, where the
DMDS sulfur is converted to SOx. This will result in additional
costs associated with the quench water pH control.
TBPS
Kurita has many years of experience with the injection of
polysulfides at steam crackers and promotes the use of
Di-tert-butyl polysulfide (TBPS). This additive has the
molecular formula (main constituent) C8H18S4 with a sulfur
content of 54 wt%.
TBPS has a number of significant advantages in
comparison to DMS and DMDS. First among the operating
personnel notices a low TBPS odour, which is similar to gas
oil. This polysulfide has a far higher flash point of 100 °C,
which reduces the risk of potential flash fires. It is classified as
non-hazardous and requires no special handling and storage.
TBPS has a very low vapour pressure at ambient temperature.
Having a viscosity of 14 mPas at 20 °C it can be finely
dispersed or atomised into the steam to the cracking furnace.
The injection system for TBPS is simple to realise as it can
be stored in a storage or smaller day tank. Storage under
nitrogen is not required. The injection into the dilution steam
can be done with a small dosing pump or by using nitrogen
pressure. The injection flow into the dilution steam can be
controlled with a coreolis meter or other suitable
measurement technology. The thermal decomposition
temperature of TBPS starts above 160 °C as shown in Table 1.
Naphtha feed from the crude distillation unit already
contains sulfur and requires no further coke inhibition agent
after pretreatment of the furnace tubes. Ethane, LPG and
unconverted oil from hydrocrackers (UCO) require a
continuous dosing of TBPS after decoking. Typical TBPS
dosing rates are 20 - 100 ppm during normal operation. As a
pretreatment procedure TBPS is injected into the dilution
steam for 2 - 4 hrs after decoking operation.
The continuous injection of TBPS significantly reduces the
coke formation with an increased run length of the cracking
furnace. Ethylene producers report that the negative impact
on quench tower acidity that is usually observed with other
sulfiding agents and attributable to SOx formation is not
observed under TBPS treatment. Typical control settings as an
indication for the successful TBPS treatment are the CO and
CO2 concentrations. Some ethylene plants keep the CO
concentration after caustic scrubber below 0.05 mol%, as CO
can poison the hydrogenation catalyst. The CO2
concentration before caustic scrubber should be below
400 vppm and after caustic scrubber below 1 vppm
(otherwise ethylene would be off spec).
Conclusion
The decoking process negatively affects the mechanical
strength and tube life. The tube service life time is
economically important as the furnace capital costs are
around 20% of the total ethylene costs. Novel tube materials
and coatings are successful technologies for an extended run
length of the furnace with less decoking procedures.
Sulfiding agents like DMS and DMDS are the historical coke
inhibition agents, which reduce the CO and coke formation
with an increased run length of the cracking furnace. Kurita
promotes the use of the polysulfide TBPS, which is a modern
sulfiding agent. TBPS provides a very good performance
combined with an increased operational safety, less odour,
good storage stability and easier handling.
References
1. DIP TAN, C., TERRY, R., and BAKER, K., Department of Chemistry,
Northeastern University, Boston, Massachusetts 02115, ‘The
impact of selected sulfides on coke formation on various metal and
alloy surfaces’.
2. European Patent EP1631699, Nova Chemicals (International) S.A.
1700 Fribourg (CH), ‘Passivation of steel surface to reduce coke
formation’, Representative: WATSON, R.J. et al, Mewburn Ellis
LLP, 33 Gutter Lane, London, EC2V 8AS (GB).
3. SRINIVAS, R.S. and HUMBLOT, F., Arkema Inc, Patent application
number: 20090283451.
4. HELLER, F.E. and ROBERTS, C.D., The Lubrizol Corporation,
'Atomized Polysulfide used in Ethylene steam cracker', Patent
7154016, Application No. 10475457 filed on 03/06/2002.
HYDROCARBON
ENGINEERING Reprinted fro m July 2014
up!
Berthold Otzisk, Michael Urschey and Ingo Königs,
Kurita Europe, present tailor made corrosion
inhibitor programmes for ethylene production plants.

E
thylene is the most commonly produced petrochemical product. The
bulk of annual commercial ethylene production is based on thermal
cracking of hydrocarbon feedstocks with steam. Usual raw materials
for pyrolysis are refinery gas, liquid petroleum gas (LPG), natural gas
liquids (NGL), naphtha or gas oil. The production of ethylene is a very
complex process, including cracking, cooling, compression and separation
steps. Fouling and corrosion are observed in almost all process parts and can
cause significant production losses. This article focuses on chemical
treatment concepts for very powerful corrosion control programmes and
protection of ethylene plant components.
Figure 1 illustrates the common application areas where corrosion inhibitor
programmes can be applied in the hot section. Similarly affected by corrosion
are the depentaniser overhead system of the gasoline hydrogenation unit (GHU)
and the hydrocarbon fractionation section (cold section).

Treatment programmes
Primary cracking reaction products are ethylene, propene, acetylene, methane
and hydrogen. Secondary cracking reaction products are C4 to C7
hydrocarbons and heavier products. The cracked gas may also include
corrosive species such as oxygen, oxygenates, CO2, H2S, phenols and organic
acids (i.e. formic, acetic and propionic acid). Well established chemical
treatment concepts for corrosion control include oxygen scavengers,
alkalising amines, filming amines (FFAs) or combinations of them.

HYDROCARBON Reprinted from April 2016

ENGINEERING

Figure 1. Overview of common injection locations for corrosion inhibitor
programmes.
Oxygen scavengers
When the cracked gases leave the furnace, they must be
rapidly cooled down from 750 - 875°C to approximately
200 - 240°C. This process step is very important to stop
further cracking, polymerisation and condensation
reactions. The cracked gases are cooled down in the
transfer line exchangers (TLX or TLE) by indirect
quenching with boiler feed water. This permits heat
recovery to generate high pressure steam. Corrosion by
oxygen in the boiler or steam generator system can be
controlled by the addition of an oxygen scavenger. For
many years, oxygen scavengers were used to avoid
corrosion by dissolved oxygen in the feed water of
boilers and steam generator systems. Well known oxygen
scavengers, with their specific advantages and
disadvantages, are:
n Hydrazine.
n Carbohydrazide.
n Sodium sulfite (Na2SO3).
n Diethylhydroxylamine (DEHA).
n Methylethylketoxime (MEKO).
At refineries and petrochemical plants, hydrazine
(N2H4) has a long history as an oxygen scavenger because
of its excellent performance. It does not produce
corrosive gases and reacts with oxygen to form nitrogen
and water at high temperatures and pressures. Hydrazine
is no longer allowed to be used in many boiler systems
due to its toxicity, and the suspicion of it being
carcinogenic. Meanwhile, more gentle treatment concepts
are given preference, with environmental requirements, in
order to protect and maintain the health of the
employees. Kurita's patented oxygen scavenger,
KURINPOWER A-6060, is a full substitute of the hitherto
used hydrazine. KURINPOWER A-6060 is an effective
oxygen scavenger, easy to handle and, above all, not
carcinogenic. Based on its safe active substance 1-A-4MP,
this oxygen scavenger has been in use for years in many
boiler systems, with high success rates.
KURINPOWER A-6060 not only offers good oxygen
scavenging quality, but also effective metal passivation. Its pH value is exclusively adjusted by the amine. The usage
high temperature stability also makes it suitable for high
pressure applications. Typical areas of application are the
boiler feed water of the transfer line exchangers and the
Reprinted from April 2016 HYDROCARBON
ENGINEERING
stripped water, which is routed to the
dilution steam generator system
(DSG).
Alkalising amines
The reason for severe corrosion
attack is, in many cases, the presence
of organic acids in the cracked gas,
such as formic or acetic acids.
Unfortunately, ammonia or sodium
hydroxide (NaOH) is still used for
neutralisation and pH control of the
oil quench, water quench, process
water stripper (PWS) and DSG. These
alkalising products are available at
very favourable prices, but have a
number of significant disadvantages. For instance,
ammonia is a weak amine, highly volatile and will not
provide good corrosion protection over a large area,
where neutralisation is absolutely essential.
Sodium hydroxide is a very strong base, and small
amounts can drastically increase pH. There is a high risk
for foam formation or sodium carryover if sodium
hydroxide is used exclusively for pH control of the
dilution steam generation system. Unwanted carryover of
sodium into the dilution steam can generate caustic
embrittlement of the boiler system and furnace tubes.
Indicators for an inadequate treatment are:
n Presence of sodium in the dilution steam.
n Foaming in the dilution steam generators.
n High liquid level in the dilution steam generators.
n Poor efficiency of the demister in the dilution steam
drum.
n High conductivity of the dilution steam.
Modern alkalising amines are neutralising amines,
which operate by reacting with any acid constituent in a
straightforward chemical neutralisation, and shift the pH
to a higher level. Often, coordinated alkanolamines are
formulated as ‘ready to use neutralising amine blends’.
They provide the right combination of high boiling and
low boiling amines to ensure an inhibition of a steam and
water phase. Each neutralising amine has unique chemical
and physical properties, therefore, the appropriate
product choice must be considered carefully. Influencing
variables are:
n Base strength.
n Boiling point.
n Vapour-liquid equilibrium (VLE).
n Effect of neutralising amines and its formed salts.
n Compatibility of different neutralising amines.
n Interaction of neutralising amines.
For cost reasons, sodium hydroxide can be injected
with a defined fixed dosing rate – the neutralising amine is
continuously dosed into the system, where the desired
of neutralising amines significantly reduces the risk of
foaming in the dilution steam generators, or carryover of
sodium.
Filming amines
Generally, the corrosion of metals in aqueous
environments is caused by electrochemical processes of
two half cell reactions. This basic corrosion cell with two
half cells requires an anode, cathode, metallic conductor
and electrolytes. If one of them is missing, aqueous
corrosion will not occur. The use of corrosion inhibitor
programmes with filming amines, also called film forming
amines (FFAs), can help to stop or decelerate the function
of a corrosion cell. They will provide perfect corrosion
protection of the distillation equipment by forming a
very thin layer on the metal surface. The film acts as a
barrier against corrosive substances such as oxygen,
oxygenates, CO2, H2S, phenols and organic acids. The free
electron pair of the amine nitrogen has a very strong
affinity to metal surfaces by forming a hydrophobic film,
which significantly reduces the corrosion potential.
Typical application fields are the steam generation
system, the oil quench overhead system or the
hydrocarbon fractionation section (cold section). The
filming amines are not consumed by the reaction with
oxygen, but protect the complete steam generation
system against corrosion attack. If filming amines are
selected, the functions of oxygen scavengers, phosphates
and caustic dispersants can be replaced by the filming
amine. Common corrosion inhibitors are formulations
with polyamines, fatty amines or imidazolines. Modern
treatment programmes often contain oleyl amine or
oleyl propylene diamine. The general chemical formula is
R1-[NH-(R2)-]n-NH2, where n is an integer between zero
and seven, R1 is an unbranched alkyl chain with
12 to 18 carbon atoms, and R2 is a short alkyl chain with
one to four carbon atoms. These filming amines can be
used in combination with alkalising amines.
Kurita's Cetamine® technology is an efficient,
innovative and comprehensive concept based on filming
amines, which meets the requirements of a modern
corrosion protection programme. It is an all in one concept
and gives the combination of three modes of action:
n Protection of the complete system due to film
formation in feedwater tank, feedwater line, boiler
water and steam/condensate system by means of film
forming amines.
n pH value adjustment in feedwater, boiler water and
steam/condensate system by means of alkalising
amines.
n Cleaning effect on metal surfaces with removal of
existing deposits by means of film forming amines
and dispersants.
Figure 2 shows the comparison of an untreated
coupon and a coupon that was treated with a Cetamine®
product beforehand. The surface-active properties create
the thin impermeable layer on the surface, so that water
or corrosive substances cannot come in contact with the
metal. Higher concentration cycles can be achieved in the
boiler system, leading to lower amounts of blowdown
water with a significant reduction in energy losses. An
oxide layer of, preferentially, magnetite (Fe3O4) is formed
when the superheated surface comes into contact with
boiler water at a saturation temperature above 180°C.
Figure 2. Comparison of an untreated and Cetamine®
protected coupon.
There is no necessity of having free oxygen present. A
compact, smooth and thin magnetite layer is desired in
steam generators. This has a significant positive influence
on heat transmission. The coefficients of heat
transmission from the heating tubes into the water phase
are far higher than when a conventional phosphate
treatment programme is used. New studies confirm that
tubes treated with Cetamine® products had a much
smoother and more homogeneous surface than tubes
examined after a sodium phosphate treatment
programme.
Conclusion
Powerful polymer fouling control and corrosion
protection of the ethylene plant components plays a very
important role. At locations where water and corrosive
components are present, the metal surface must be
effectively protected against corrosion attack.
Besides oxygen scavengers and alkalising amines, the
film forming amines are well established treatment
programmes. The use of a film forming amine can replace
oxygen scavengers, phosphates or caustic containing
dispersants. Kurita's Cetamine® technology offers
significant benefits, particularly at a time when energy
costs and savings are so heavily foregrounded.
References
1. OTZISK, B., First line of defense, Kurita Europe, Germany,
Hydrocarbon Engineering, July 2012.
2. HATER, W., DE BACHE, A., and KÖNIGS, I., Film Forming
Technology, Kurita Europe, Germany, Hydrocarbon Engineering,
September 2015.
3. JACK, M., WEERAKUL, S., and LISTER, D. H., The interaction of a
film forming amine with surfaces of a recirculating experimental
water loop, Department of Chemical Engineering, University
of New Brunswick, Canada, Heat Exchanger Fouling and
Cleaning – 2015, Enfield (Dublin), Ireland.
4. HATER, W., KRAFT, P., and FORÊT, C., Int. J. Corros, The influence of
alkalising amines on the film formation by oleyldiamine; Scale
Inhib., 2015, 4, No. 4, pp. 353 - 364.
5. VAN LIER, R., GERARDS, M., and SAVELKOUL, J., Experience with
polyamines in the high-pressure steam system of a naphtha
cracker – from new to proven treatment, VGB Powertech 8/2012,
pp. 84 - 87.
HYDROCARBON Reprinted from April 2016
ENGINEERING

Berthold Otzisk, Kurita, Germany, discusses the
chemical treatment programmes that can be applied
to ethylene production in order to avoid fouling,
corrossion and the formation of unwanted emulsions.
I
n the production process for ethylene by steam cracking,
fouling, corrosion or the formation of unwanted
emulsions are common and can cause production losses.
In such cases, chemical treatment programs can be
applied to avoid such negative effects and help to achieve the
customers’ targets.
Ethylene is an extremely important and large volume
petrochemical with a production of approximately 140 million t
in 2012. It is a colourless flammable gas with a sweet odor.
Ethylene can undergo many types of chemical reactions. It can
be converted to polymers, oligomers, saturated hydrocarbons
and derivatives. The main part is used to produce reaction
products like polyethylene, ethylene oxide, ethylene
dichloride, styrene monomer, acetaldehyde, vinyl acetate or
ethylene glycol. In smaller quantities, ethylene is used as
welding gas, anesthetic agent or to hasten fruit ripening.
The commercial production of ethylene is based on
thermal cracking of petroleum hydrocarbons with steam. This
FOUL
PLAY
process is commonly called steam cracking or pyrolysis. The
production of ethylene is a very complex process including
cracking, cooling, compression and separation steps.
Cracking
The cracking reactor (furnace) is the heart and starting point of
the ethylene production. Typical feedstocks for the thermal
cracking are gases (i.e. LPG), naphtha, fuel oil and gas oil. The
hydrocarbon stream is preheated in the convection section,
mixed with steam and further heated to an initiating cracking
temperature of 500 – 680 °C. The addition of steam has
several beneficial effects. It reduces the hydrocarbon partial
pressure and favours the formation of ethylene, propene,
acetylene, hydrogen and methane. It also reduces the partial
pressure of high boiling components, resulting in reduced coke
deposition on the inside of the furnace tubes and less
downtime of the furnace for decoking. The hot gases leave the
convection zone and enter the radiant section of the furnace
 where cracking into smaller molecules takes place within
0.1 – 0.5 seconds. To avoid undesired side reactions the
reactants having a temperature between 750 – 875 °C, are
cooled down immediately in the transfer line exchangers (TLX,
TLE) before they are routed to oil quench and/or water
quench columns for further cooling.
Coke formation in steam cracking furnaces and transfer
line exchangers is still a major operation problem. Plenty has
already been written about the mechanism of coke formation.
In general, there are four known mechanisms:
Figure 1. Steam cracking process.
Figure 2. Simple overview of a cracking furnace.
Figure 3. Typical design of a DSG system.
HYDROCARBON
Reprinted from April 2013 ENGINEERING
n Pyrolytic coke formation.
n Catalytic coke formation.
n Droplet condensation.
n Mechanical transfer of coke from the radiant zone.
The formation of coke consists of three types of carbon,
which are amorphous, filamentous and graphitic. Particles of
nickel and iron or alloys containing these metals catalyse
the formation of filamentous carbon. It is well known, that
thiocompounds (i.e. sulfiding agents) reduce the coke
formation by passivating the active centres (Ni, Fe).
Dimethyldisulfide (DMDS) is commonly used as sulfiding
agent but has a number of negative aspects and hazardous
properties. It has a bad odor even in case of the new DMDS
evolutions that have been introduced to the industry. The
flash point of 16 °C is very low, and special handling, storage
and transportation are mandatory. Operators are required to
wear special clothing.
Kurita's Cut-Coke technology uses a sulfiding agent with
a linear polysulfide chain and is a fundamental step for
improvement of safety and hazardous aspects. The Cut-Coke
technology has none of the hazardous properties associated
with applying DMDS and provides low odor, excellent
handling properties and a high flash point of 100 °C.
Cooling
The cracked process gases must be rapidly cooled down to
a temperature below 230 °C to stop further cracking and
unwanted conversions. The hot gases are cooled by indirect
quenching in the transfer line exchanger. Heat is recovered
to produce valuable high pressure steam. Transfer line
exchangers are located very close to the furnace outlet.
Oxygen scavengers are injected into the boiler water in front
of the transfer line exchanger to avoid corrosion. New
products based on DEHA or Carbohydrazide have been
introduced to replace hydrazine, which is an unpleasant and
potentially carcinogenic chemical. But these oxygen
scavengers show a lower performance compared to
hydrazine. The patented KURITA 1A4MP technology provides
the same high performance like hydrazine without its
disadvantages and hazardous risks.
Leaving the transfer line exchanger, the hot gases are
further cooled by direct oil quench column (primary
fractionator) and water quench column. Depending on the
design or feedstock qualities some steam crackers route the
hot gases directly from the transfer line exchanger to a
water quench column. Some older ethylene plants use air fin
coolers for cooling instead of a water quench column.
Distillate fuel oil and heavy fuel oil fractions are
typically removed from the cracked gases at ethylene units
with direct oil quench. The gases are routed overhead to
the water quench tower, where the cracked gases are
cooled to near ambient temperature to prevent
polymerisation and formation of unwanted byproducts.
The quench water is collected from the bottom section to
be separated from hydrocarbons in the oil/water separator
before it is cooled and returned to the water quench tower.
Excess water from the quench water column is stripped in
the process water stripper (PWS) to remove volatile
hydrocarbons.

Figure 4. Adol condensation process.
Chemical programs help to keep the process conditions
stable to provide a long run length. The injection of a
suitable antifoulant into the circulating quench oil avoids
fouling deposition and production losses. The injection of a
suitable emulsion breaker in front of the oil/water separator
can help to accelerate the separation process. Filming amines
and neutralising amines can be injected into the overhead
systems of oil quench or water quench columns. They
protect the metallurgy against corrosive components like H2S
or organic acids. Neutralising amines can be used to control
the pH at the process water stripper and dilution steam
generators (DSG). They replace caustic for this application
and do not bear the risk of unwanted sodium carryover,
which can lead to stress corrosion cracking of the furnace
tubes.
Compression
In order to fractionate the gas streams they must be liquefied.
Light gases become liquid by increasing the pressure in a
compressor or by reducing the temperature in a cold box.
Compression is typically performed by a turbine driven
centrifugal compressor in four or five stages after cooling of
the cracked gas. Temperature control and interstage cooling
keep the cracked gas below 100 °C to prevent polymerisation
and fouling of the compressor equipment. Condensation of
hydrocarbons and water occurs during compression.
Depending on design the condensates from the first three
stages are often returned to the water quench tower or
oil/water separator. Downstream of the compressor, methane,
hydrogen, ethylene, propylene and heavier fractions are
separated in a series of fractionator columns.
Unsaturated hydrocarbons such as ethylene, propylene,
butadiene and styrene can add to itself to create long chain
polymers. Free radicals are generated by heat, metals, oxygen
and peroxides, acting as initiators for the polymerisation
reactions. Polymerisation is a severe problem during
compression. Water or aromatic rich oil is used for washing
and removal of fouling materials from intercooling equipment
and compressor internals. Powerful antifoulants such as
KURITA AP-2402 will help to keep the process conditions
stable and minimise negative impacts on the polytropic head
efficiency. The antifoulants are generally blended chemicals
that may contain antipolymerants, antioxidants, metal
deactivators and dispersants.
Acid gases such as H2S and CO2 are removed from the
cracked gas stream between the third and fourth stages
or between the fourth and fifth stages of compression.
For that purpose the gases are routed to the caustic
scrubber, where they are scrubbed with NaOH.
The caustic scrubber operates with a cascading flow
of caustic solution. The cracked gas also contains small
amounts of aldehydes and ketones. Under the alkaline
conditions they tend to react in an Aldol reaction forming
resinous polymers. An indication for this Aldol reaction is
the colour change of the caustic from colorless/light
yellow to dark yellow/reddish color. Figure 2 shows the
stepwise color change from of a fresh caustic (technical
quality) to a spent caustic after Aldol condensation.
To avoid Aldol condensation special antifoulants can
be applied such as KURITA AP-2501, which scavenges
aldehydes and ketones. It reacts with the carbonyls to
form water soluble alcohols and eliminates the need for
frequent skimming of 'red oils'.
Separation
Downstream of the compressor is a series of fractionators
that separate hydrogen, methane, ethylene, ethane,
propylene and heavier products. Before separation, the
gases are routed from the compressor to a molecular sieve
liquid dehydrator. The drying process is necessary to avoid
increased pressure drop and plugging, caused by moisture
at the extremely low temperature in the cold box system.
Depending on design the dried gases are fed to a
deethaniser (C2/C3 separation), depropaniser (C3/C4
separation) and debutaniser (C4/C5 separation). After
partial hydrogenation of the deethaniser C2 fraction, it is
fed to a compressor and cooling system. The
demethaniser separates methane from the C2 fraction,
which again is separated into ethane and the main
product ethylene.
Polmyer fouling is a very severe problem especially in
the reboilers of the deethaniser and depropaniser. If no
antifoulants are used, fouling of the reboiler and trays can
significantly limit the run length and capacity of the
whole plant. The injection of a suitable antifoulant will
help to keep the process conditions more stable without
unwanted polymer formation. Additionally to antifoulant
programs many olefin units have spare reboilers to allow
a run length up to five years.
Conclusion
Ethylene is the most produced organic compound in the
world. The production by steam cracking is a complex
process, which is divided into a hot section and a cold
section, where the volatile hydrocarbons are separated.
Corrosion, fouling and formation of emulsions are
common problems, which may lead to production losses
or unplanned shutdown. Chemical programs based on
antifoulants, corrosion inhibitors or emulsion breakers are
applied to maintain the productivity of the plants. New
proprietary products have been developed to replace
hazardous chemicals like hydrazine and to inhibit fouling
reactions with higher efficiency by eliminating specific
fouling precursors.
HYDROCARBON
ENGINEERING Reprinted from April 2013
 World Academy of Science, Engineering and Technology
International Journal of Chemical, Molecular, Nuclear, Materials and Metallurgical Engineering Vol:10, No:8, 2016

Abstract—
International Science Index, Chemical and Molecular Engineering Vol:10, No:8, 2016 waset.org/Publication/10007069

Keywords

International Scholarly and Scientific Research & Innovation 10(8) 2016 1162 scholar.waset.org/1999.2/10007069
 World Academy of Science, Engineering and Technology
International Journal of Chemical, Molecular, Nuclear, Materials and Metallurgical Engineering Vol:10, No:8, 2016

CO NaOH Na CO H O

H S NaOH Na S HO

CO NaOH NaHCO

H S NaOH NaHS H O

H S Na S NaHS
International Science Index, Chemical and Molecular Engineering Vol:10, No:8, 2016 waset.org/Publication/10007069

B. Physical Characteristics of the Spent Caustic

o
o

C. The Cause and Mechanism of Polymer Formation in

Caustic Tower

A. The Tower Caustic Reactions

CO NaOH NaHCO

International Scholarly and Scientific Research & Innovation 10(8) 2016 1163 scholar.waset.org/1999.2/10007069
 World Academy of Science, Engineering and Technology
International Journal of Chemical, Molecular, Nuclear, Materials and Metallurgical Engineering Vol:10, No:8, 2016

D. The Mechanism of Polymer Formation

International Science Index, Chemical and Molecular Engineering Vol:10, No:8, 2016 waset.org/Publication/10007069

E. Various Mechanisms of Polymer Formation

R H R H

M R H M R H

M ROOH M R O O H

R O R O O

R O O R H R O O H R

R C C R C C Polymer

R R R R

R R O O R O O R

International Scholarly and Scientific Research & Innovation 10(8) 2016 1164 scholar.waset.org/1999.2/10007069
 World Academy of Science, Engineering and Technology
International Journal of Chemical, Molecular, Nuclear, Materials and Metallurgical Engineering Vol:10, No:8, 2016
International Science Index, Chemical and Molecular Engineering Vol:10, No:8, 2016 waset.org/Publication/10007069

CH CHO NaOH CH CH OH CH CHO

CH CH OH CH CHO H O CH CH CHCHO

CH CH CHCHO CH CHO NaOH CH CH CHCH OH CH CHO

CH CH CHCH OH CH CHO H O CH CH CHCH CHCHO

International Scholarly and Scientific Research & Innovation 10(8) 2016 1165 scholar.waset.org/1999.2/10007069
 World Academy of Science, Engineering and Technology
International Journal of Chemical, Molecular, Nuclear, Materials and Metallurgical Engineering Vol:10, No:8, 2016

CH CHO CH CHO CH CHOH CH CHO Oxygenated Comp. Concentration

Acetaldehyde
International Science Index, Chemical and Molecular Engineering Vol:10, No:8, 2016 waset.org/Publication/10007069

R CH COH NH OH R CH CHN OH H O

A. The Relationship between Olefin Feed Impurities and

Tower Effluent

C. The Relationship between Caustic Concentration at the

Bottom of the Tower and Pyrolysis Gasoline Discharged from
B. Injection of DMDS to Olefin Plant Furnaces and Its Effect the Caustic Tower to the Effluent
on the Caustic Tower Effluent

International Scholarly and Scientific Research & Innovation 10(8) 2016 1166 scholar.waset.org/1999.2/10007069
 World Academy of Science, Engineering and Technology
International Journal of Chemical, Molecular, Nuclear, Materials and Metallurgical Engineering Vol:10, No:8, 2016

Diff. Concentration of Acetaldehyde

Caustic Concentration
International Science Index, Chemical and Molecular Engineering Vol:10, No:8, 2016 waset.org/Publication/10007069

International Scholarly and Scientific Research & Innovation 10(8) 2016 1167 scholar.waset.org/1999.2/10007069
 World Academy of Science, Engineering and Technology
International Journal of Chemical, Molecular, Nuclear, Materials and Metallurgical Engineering Vol:10, No:8, 2016
International Science Index, Chemical and Molecular Engineering Vol:10, No:8, 2016 waset.org/Publication/10007069

D. The Relationship between the Composition of the Olefin

Plant Liquid Feed and 1,3butadiene

International Scholarly and Scientific Research & Innovation 10(8) 2016 1168 scholar.waset.org/1999.2/10007069
 World Academy of Science, Engineering and Technology
International Journal of Chemical, Molecular, Nuclear, Materials and Metallurgical Engineering Vol:10, No:8, 2016
International Science Index, Chemical and Molecular Engineering Vol:10, No:8, 2016 waset.org/Publication/10007069

International Scholarly and Scientific Research & Innovation 10(8) 2016 1169 scholar.waset.org/1999.2/10007069
 Three Years of Experience with Polyamines in the High Pressure Steam System of a Naphtha Cracker

Three Years of Experience with Polyamines in the High

Pressure Steam System of a Naphtha Cracker
Roy van Lier, Gerard Janssen, and Jo Savelkoul

ABSTRACT
The design and operation of the high pressure steam system of a naphtha cracker pose specific problems which in
some cases cannot be satisfactorily resolved with conventional water/steam treatment programs. Following a
decade of operational experience with polyamine products at lower pressures at the Geleen (petro)chemical site in
the Netherlands, the ammonia/morpholine treatment of the 12.5 MPa steam system of one of SABIC Europe's
naphtha crackers was converted to a polyamine program in November 2005. In this paper, the characteristics of the
cracker's steam system are first described. Particular aspects of the conversion are then discussed. Finally, the
experience gained and insights obtained into the polyamine treatment during the first three years are elaborated.

INTRODUCTION
SABIC Europe owns and operates three naphtha cracker
complexes: two in Geleen, Netherlands and one near
Teesside, United Kingdom. The UK cracker has a capacity
of approximately 0.85 Mt of ethylene per annum, while the
two Dutch crackers have a combined capacity of about
1.3 Mt of ethylene per year.
Although their designs and operating modes are quite dif-
ferent, the three plants have some general peculiarities of
high pressure (HP) steam generation in crackers in com-
mon. First of all, naphtha crackers are equipped with very
high heat flux transfer line exchangers (TLEs) - cracker-
specific heat recovery steam generators (HRSGs) - which
are susceptible to fouling and corrosion. Secondly, the
design of the steam cycle is minimalistic without mud
drums or large headers for collection and removal of mag-
netite deposits and with little instrumentation to rely on.
Furthermore, the feedwater and boiler water of multiple
TLEs are centrally alkalized. That leaves only the blow-
down rate to adjust boiler water quality in the HP steam
system of a given naphtha furnace. This is all the more
challenging since TLE steaming rates vary considerably
between start-up, steady state operation and decoking of
the furnaces. Finally, preventive chemical cleaning of TLEs
is not an option for reasons of plant availability, equipment
accessibility and costs.
For this cracker, corrosion and steam quality problems
could not be adequately resolved using conventional treat-
ment chemicals. Following a decade of predominantly
positive operational experience with polyamine products at
lower pressures at the Geleen (petro)chemical site and
elsewhere in the Netherlands, the treatment of the
cracker s HP steam system was converted to polyamines
in November 2005.
In this paper, the characteristics of the cracker s steam
system are first explained in general terms. Particular
aspects of the conversion are then addressed. Finally, the
experience gained and insights obtained into the poly-
amine treatment during the first three years are elaborated.
CRACKER STEAM SYSTEMS
In a naphtha cracker, liquid naphtha feedstock is cracked
into ethylene (C2), propylene (C3) and a whole range of
light and heavy hydrocarbons at a temperature of 800 C -
850 C. The cracking process takes place in reactor coils in
refractory lined furnaces equipped with burners. These are
fired using methane produced by the cracker itself and
natural gas.

One of the crackers in the Netherlands, hereafter referred
to as "the cracker" or "the Geleen cracker," produced its
first ethylene in 1978. Its 12.5 MPa steam system was
originally alkalized using ammonia, hydrazine and morpho-
line. The dosing of hydrazine was stopped in 1996.
Two independent and distinctly different steam systems
can be distinguished in a cracker:
1. a low pressure process steam system
2. a high pressure turbine steam system

2008 by PowerPlantChemistry GmbH. All rights reserved.

696 ,
Three Years of Experience with Polyamines in the High Pressure Steam System of a Naphtha Cracker
Low Pressure Process Steam System
The cracking process uses so-called dil(ution) steam or
process steam to lower the partial pressure of naphtha
components, and, thus, the required cracking tempera-
ture. Upon cooling this steam is condensed together with
gasoline fractions. The resulting condensate is separated
from the hydrocarbons either in a drum or in a tower.
The hydrocarbon-laden steam condensate subsequently
proceeds to a stripper to remove as much total organic
carbon (TOC) as possible through injection of clean low
pressure (LP) steam. The stripped condensate becomes
the feedwater for dilution steam production.
At the Geleen cracker, dil steam is produced in a battery of
natural circulation HRSGs with a common drum. The boil-
ers are vertical shell and tube heat exchangers with the
boiler water on the shell side. The tube side medium is hot
quench oil. Drum pressure is 0.7 MPa.
Because of elevated TOC concentrations in the feedwater,
the dilution steam system has long been the subject of
water quality, fouling and corrosion problems. These diffi-
culties are partly inherent to the system in the sense that
the optimal pH for aqueous/organic phase separation sig-
nificantly differs from the pH for proper alkalization of feed-
water. Furthermore, the required hardware (instrumenta-
tion, multiple dosing points) for adequate pH control has
been lacking.
Following long operating periods with ethanolamine (ETA)
and trials with sodium hydroxide for pH adjustment and an
amide-based dispersant for fouling control, the dil steam
system has been treated with a polyamine-polyacrylate
blend since 1998. Although the process steam system is
not the focus of this paper, it is mentioned here as it repre-
sents an area where a lot of site experience with
polyamines has been gained.
12.5 MPa import steam
from utility supplier
HRSGs
(TLEs)
Ammonia
Morpholine
Make-up water Blowdowns
Deaerator
(demineralized
Ammonia
V Reboiler condensates
AEA_ Condensate 14_
Candle filter
polishing unit
Figure 1:
Block diagram of cracker HP steam system from the mid 1990s up to conversion.
High Pressure Turbine Steam System
As explained in previous papers [1,2], the design and oper-
ating regimes of steam generators in the chemical process
industry (CPI) are dictated by the (petro)chemical process
conditions. This very much holds for the HP steam system
of a cracker.
Steam system description A block diagram of the
cracker HP steam system prior to conversion to polyamine
treatment is depicted in Figure 1. Feedwater was and still
is thermally deaerated at ca. 125 C through LP steam
injection. Before conversion to the polyamine program,
morpholine was dosed into the suction line of the feed
pumps. Hydrazine was used until 1996.
Figure 2 shows how the TLEs and peripheral equipment
are integrated into the cracking process. Each furnace has
its own HP steam system in which one drum serves either
2 or 3 TLEs. The metallurgy of the TLEs, like the rest of the
steam cycle, is all-ferrous.
All drums receive the same centrally alkalized feedwater. In
other words, the only degree of freedom left to adjust
boiler water quality in a given system is the blowdown rate.
Blowdown is normally controlled by pH and on rare occa-
sions by the silica concentration.
Periodically, the carbon fouling on the inner surface of
reactor coils in a furnace, transition lines and cracked gas
entrance to the TLEs is removed in a decoking operation.
This entails the oxidation of the coke deposits using a
steam/air mixture. During decoking, the steaming rate of
the TLEs is reduced to approximately 40 % of normal
steam production.
HP steam is superheated using hot flue gas. At the Geleen
cracker, the temperature of superheated steam is con-
trolled at 520 C through feedwater attemperation. The
1.8 MPa export steam
to site users
Export condensate
to utility supplier
Air-cooled
Turbines
condensers
Steam users Turbine condensates
in cracker
Tracing condensates
697
 Three Years of Experience with Polyamines in the High Pressure Steam System of a Naphtha Cracker

HRSG design The HRSGs are vertical TLEs of three

distinctly different designs:
Flue gas
Steam
drum
1. Borsig conventional shell and tube
Riser
2. Borsig tunnel flow
Cooled 3. SHG (Schmidt sche Heissdampf-Gesellschaft) double
Boiler feedwater cracked tube
gas
Liquid naphtha
feedstock The reader is referred to the manufacturers for more infor-
Downcomer mation on TLE designs.
Process steam
The primary purpose of the TLEs is to cool the cracked gas
as quickly as possible from 800-850 C to - 500 C in
HP steam Transfer line order to arrest the cracking process so as to prevent

71
unwanted secondary reactions between the chemically
ot cracked gas
unstable olefins. This "freezing" requirement translates into
a compact design with local heat fluxes of many hundreds
of kW rn -r2 [3,4].

All three vertical TLE designs are absolutely unforgiving in

the sense that the highest heat flux zone coincides with a
Firing physical low-point, where deposits can - and, again, in our
experience, with conventional water/steam treatment pro-
grams will - accumulate!

TLE damage is mostly only discussed at specialized

forums. Examples of TLE failures are reported in general
terms in [3] and in some detail in [5,6].
Figure 2:
Schematic of furnace/boiler configuration (adapted from [3]).
Make-up water quality In Geleen, demineralized
water prepared from River Maas water is used as make-up
superheated steam is expanded over turbines together water for most steam systems. The quality of the demin
with ammonia treated HP import steam, which is returned water in terms of specific conductivity, TOC and silica con-
to the local utility supplier as 1.8 MPa steam. Condensate centration is excellent as most of it is produced in a state-
may be exported as well. The largest (back pressure/con- of-the-art membrane plant.
densation) turbine drives the cracked gas compressor,
while a smaller back pressure and back pressure/conden-
sation turbine drive the C2 and 03 compressor, respec- Steam system monitoring On-line monitoring of
tively. water quality in the cracker s HP steam system is limited
to:
The condensation pressure is maintained at 0.02 MPa. The
Feedwater pH;
exhausted steam is condensed in air coolers. The turbine
condensates are pumped back to the deaerator untreated. Feedwater acid conductivity
They constitute about half of the total condensate return (cation conductivity in the Anglo-Saxon literature);
flow. The other half results from condensation of steam in
Specific conductivity after cation exchangers
reboilers, tracing lines, etc. As these condensates are
(= condensate acid conductivity);
potentially contaminated with hydrocarbons or inorganic
salts, they are treated in a condensate polishing unit (CPU) Specific conductivity after mixed bed polishers;
with two trains of a standard carbon filter - cation
Silica after mixed bed polishers.
exchanger - mixed bed polisher configuration. The CPU
also treats the small flow of blowdowns.
Boiler water grab samples are analyzed twice per week for
pH (25 C), silica concentration, etc. The oxygen concen-
Water exiting the mixed bed polishers used to be alkalized
tration in feedwater after deaeration is measured once per
with ammonia to mitigate (erosion-)corrosion of the treated
month. On-line analyzers are typically calibrated monthly.
condensate line. The combined polished/turbine conden-
sate is pumped through a cartridge filter to remove any
suspended corrosion products.

698 - 07:
h 11 - 07: r j

Three Years of Experience with Polyamines in the High Pressure Steam System of a Naphtha Cracker
ASPECTS OF THE CI _ SION OF
AMMONIA/MORPH( POLYAMINES
Rationale
A summary of the corrosion history of the cracker s HP
steam system was presented as the "Morpholine Case" in
[1]. It is repeated and elaborated here.
In essence, over the years, the HP steam cycle of the
Geleen cracker has been plagued by two corrosion phe-
nomena:
1. Flow-accelerated corrosion (FAC) in the pre-boiler and
condensate systems [7]
2. First condensate corrosion (FCC) in the condensate
system [8]
Both have led to the transport of iron (oxides) and subse-
quent fouling of thermally highly loaded heat transfer sur-
faces in the TLEs with porous deposits. These have acted
as concentration cells for non-volatile boiler water con-
taminants, resulting in boiler tube failures.
Since start-up of the cracker, several measures have been
taken to minimize corrosion and the transfer of iron (oxides)
from the pre-boiler and condensate systems to the TLEs:
Alkalization of polished condensate with ammonia to
combat FAC in the line from the CPU to the deaerator;
Installation of a cartridge filter in the combined return
condensate line to remove suspended iron oxides,
especially from the air-cooled turbine condensates;
Modification of the operating mode of the steam cycle
from "standard" all volatile treatment (AVT) to high-AVT
by increasing the ammonia concentration in the feed-
water from 0.1-0.2 mg kg -1 to 1 mg kg-1 in order to
counterbalance loss of alkalinity due to acidic morpho-
line degradation products and, in this way, limit FCC;
12.5 MPa import steam
from utility supplier
HRSGs
Hydrazine
(TL Es)
Ammonia
Morpholine
Make-up water B owdowns
Deaerator 4 Steam users
(demineralized)
Condensate
polishing unit
Figure 3:
Block diagram of cracker HP steam system at commissioning in 1978.
II
1-111.-1111
Discontinuation of hydrazine dosing to allow a few ppbs
of residual oxygen in the feedwater for stabilization of
the magnetite layer to reduce FAC susceptibility.
Additionally, periodic operation of bottom tubesheet blow-
down valves for removal of deposits from the TLEs has
been strictly adhered to from the early 1980s on.
The evolution of the HP steam cycle configuration and
treatment is graphically summarized in Figures 3 and 4.
Taking out the hydrazine dosing finally led to the layout of
Figure 1.
Despite all of the modifications, difficulties with corrosion
and steam quality persisted. The root cause was the ther-
mal instability of morpholine. With more and more support-
ing literature on the negative effects of acidic breakdown
products of morpholine, the search - at that time with DSM
- for alternative steam cycle treatment chemicals was
intensified in the late 1980s.
Sodium hydroxide and sodium phosphate were not con-
sidered as options since under the "wrong circumstances"
(i) these may themselves become a threat to boiler tube
integrity and (ii) both are prone to cause turbine fouling
due to mechanical carry-over. Moreover, the use of either
NaOH or Na3PO4 would have necessitated major hardware
changes so as not to use Na-bearing water for attempera-
tion of superheated steam.
Several organic iron dispersants were investigated via lab-
oratory testing. Experiments were done with electrically
heated steel probes in alkalized deaerated demineralized
water in a 10 L continuous autoclave at a heat flux of
100 kW m-2 , a bulk water temperature of 300 C and a
water residence time of approximately 1 h. Tests were fol-
lowed by detailed microscopic investigation of the probes
and comprehensive analysis of water samples.
1.8 MPa export steam
to site users
Export condensate
to utility supplier
Air-cooled
Turbines
condensers
V
Turbine condensates
in cracker
Reboiler condensates
Tracing condensates
699
 11 I I 11 I 1 ! Three Years of Experience with Polyamines in the High Pressure Steam System of a Naphtha Cracker
12.5 MPa import steam
from utility supplier
HRSGs
Hydrazine
(TLEs)
Ammonia
Morpholine
Make-up water Blowdowns
Deaerator
(demineralized)
Am onia
V Reboiler condensates
Condensate
Cartridge filter
polishing unit
Figure 4:
Block diagram of cracker HP steam system by the mid 1980s.
Although most dispersing agents tested turned out to be
unfit for use at cracker conditions for reasons of thermal
degradation and/or because they did not demonstrate a
clear beneficial action, one suitable candidate was identi-
fied. The continuous removal of loose magnetite deposits
and complexation of dissolved iron (hydroxide) species
certainly seemed to be a promising approach to the mitiga-
tion of deposition problems in the boiler systems.
A desk study was carried out to evaluate all kinds of neu-
tralizing amines as substitutes for morpholine. This study
comprised rigorous pHT -calculations using an in-house
developed computer program. With the exception of ETA,
all amine alternatives were considered unsuitable for the
cracker HP steam cycle because of thermal instability, lack
of alkalizing power, volatility, excessive costs, limited avail-
ability and/or insufficient operating experience.
In the mid 1990s, DSM was gaining its first site experience
with mixtures of neutralizing and filming amines [1]. It was
felt that the combined alkalization and "cleaning" action of
these so-called "polyamine" blends would be the key to
solving the problems in the cracker s HP steam circuit and,
possibly, in many other industrial steam systems. Indeed,
to date we still believe that only polyamines make it possi-
ble to combat the source and symptoms of iron (oxide)
transport problems integrally.
Despite encouraging test results in the lab and a success-
ful trial run in an old 14.0 MPa power plant boiler in Geleen
just before its dismantling, these new amine blends defi-
nitely needed to prove themselves in a (petro)chemical
plant environment before they were to be used in a cracker.
It was expected that a common interest of potential end-
users, suppliers and researchers would further polyamine
technology in the years to come. In the meantime, an
ammonia/ethanolamine treatment was considered the next
best replacement for the ammonia/morpholine alkalization.
700
1.8 MPa export steam
to site users
Export condensate
to utility supplier
Air-cooled
Turbines --
condensers
Steam users Turbine condensates
in cracker
Tracing condensates
Some 10 years later, the cracker s HP steam system is
treated with polyamines. The ETA program was never
implemented. The amount of fundamental and generic
data on polyamine technology available in the public
domain has not significantly improved over the last
decade.
Polyamine Knowledge Base
In our opinion, both the development and acceptance of
polyamine technology are hindered (i) by the confidential
and proprietary nature of much of the research work, (ii) by
a wait-and-see attitude, or, at times, even an antagonistic
position of technical institutes and major equipment sup-
pliers, and (iii) by a fragmentation of some of the suppliers
knowledge base in recent years.
In our view, the status quo puts (potential) users of
polyamine products into a difficult position: it demands a
high level of expertise and analytical capabilities of the
client and/or strong confidence in the supplier s know-
how. It is hoped that the discussions at the Lucerne con-
ference Interaction of Organics and Organic Plant Cycle
Treatment Chemicals with Water, Steam, and Materials will
contribute to changing the "polyamine position" for the
better.
The advancement of polyamine technology has largely
been driven by the industry. The application of polyamines
has been described for a Polish power plant [9], German
electrical power and co-generation plants [10-16], Belgian
power plants [17-19], steam systems in the international
paper industry [20-23], utility plants at a German oil refin-
ery [24-26] and steam systems in the Swedish [20] and
Dutch [1] CPI . Boil-out and on-line cleaning experience
with polyamine-polyacrylate mixtures in power plant boil-
ers has been recorded in [24-26] and [19,27], respectively.
Not all applications have resulted in positive results [19,28].
:

Three Years of Experience with Polyamines in the High Pressure Steam System of a Naphtha Cracker
The problem with case descriptions is that they contain
information that is not directly transferable just like the
contents of this paper. Unfortunately, little applied let
alone fundamental research work on polyamines is avail-
able in the public domain, the work reported by Bohnsack
[29], Frahne and Blum [30-32] and Steinbrecht [33,34]
being positive exceptions. The pioneering work by Moran
[35] is gratefully acknowledged. It is highly appreciated
that Hater [36] utilizes pure components rather than pro-
prietary amine blends as part of a recent BK Giulini
research project.
Thanks to Bursik s "ambassador role" [37,38], polyamines
can no longer be ignored by technical institutions such as
Electric Power Research Institute, Palo Alto, CA, U.S.A.
(EPRI), TOV and VGB PowerTech (VGB).
Chemistry of Typical Neutralizing/Filming Amine
Blends
The Material Safety Data Sheets of several commercial for-
mulations of polyamine products list three major compo-
nents:
1. Octadecenyl-1,3-propane diamine
2. Cyclohexylamine
3. Ethanolamine
Since one of the principal concerns with these products is
thermal stability, it is worthwhile considering the nature of
potential breakdown products and unwanted interactions.
Many problems can be prevented by selecting the appro-
priate amine blend and by choosing proper dosing points
and equipment.
Octadecenyl-1,3-propane diamine (OPD) This is
the "model component" Bohnsack investigated [29]. He
referred to it as N-oleyl-1,3-propane diamine, a trivial
name; organic chemistry purists would describe OPD as
N-((Z)-octadec-9-enyl)propane-1,3-diamine. OPD is the
principal filming amine in the mixture.
Possible degradation products of OPD include ammonia,
carbon dioxide, short chain amines and octadecylamine.
On the one hand, chemisorption of OPD onto iron oxides
will enhance its thermal stability. On the other hand, such
interaction will increase its retention time in the system,
thus making it more amenable to breakdown.
Cyclohexylamine (CHA) CHA is a multiple-action
constituent in the amine mixture. First of all, it is presumed
to have a stabilizing (possibly emulsifying) effect on the
blend, particularly when dosing and diluting the polyamine
product. Secondly, it is attributed a stabilizing effect on the
polyamine film at the water/steel interface. Finally, due to
its volatility, it is a good steam alkalization agent.
LI
t ,1 I I ;1111 ,
We are not completely sure about the mechanisms behind
these alleged stabilizing effects. In light of this, it would
also be very interesting to know what exactly goes on at
the boiler water/steam interface, where CHA is preferen-
tially released into the steam phase. How is the stability of
the polyamine film affected in that area? Does alkalization
at the waterline default to the neutralizing amines present
in the blend?
According to Frahne, CHA may form dimers and trimers as
well as components like N-cyclohexyl-1,3-propane di-
amine derivatives in the steam system.
Ethanolamine ETA is much less volatile than CHA.
Generally speaking, its oxygen atom is sensitive to
replacement by nitrogen.
The use of ETA has come under scrutiny for reasons of the
interaction of breakdown products with ion exchange
resins leading to capacity losses in CPUs [39,40].
SABIC Polyamine Experience
SABIC, partly through its DSM heritage, has over 10 years
of operating experience with polyamines. This predomi-
nantly positive experience revolves around Helamin prod-
ucts, in particular "90HTurb."
Our own experience has made us especially observant of
the following four polyamine related issues:
1. availability of CHA
2. occurrence of fouling
3. compatibility with catalysts and ion exchange resins
4. compatibility with site utilities infrastructure
Availability of cyclohexylamine Figure 5 shows
how we have built in flexibility to modify the CHA concen-
tration, if necessary, in the steam system.
Questions, of course, are (i) what the appropriate CHA-to-
polyamine ratio really is, (ii) how its optimal value varies
within the steam cycle, and (iii) how it relates to the dosing
point of polyamine into the feedwater line, where the dilu-
tion factor is in the order of 106.
In our experience, to some extent, the "right" CHA-to-
polyamine ratio is a matter of trial and error. For now, main-
taining the pH (25 C) of treated condensate from the CPU
at > 8.5 against (erosion-)corrosion seems to work in terms
of CHA availability for polyamine stabilization as well.
Although most of it is added upstream from the deaerator,
CHA losses to the deaerator s steam plume are minimal.
Dosing of polyamines to condensate exiting the CPU is not
an option anyway since the temperature there is < 90 C.
701
 H .1 Three Years of Experience with Polyamines in the High Pressure Steam System of a Naphtha Cracker
12.5 MPa import steam
from utility supplier
HRSGs
Polyamine
(TLEs)
Make-up water
Deaerator Blowdowns
(demineralized)
A in cracker
Cyclohexylamine
Condensate 14_
Cartridge filter
polishing unit
Figure 5:
Block diagram of cracker HP steam system with polyamine treatment; polyamine product also contains cyclohexylamine.
Occurrence of fouling In our experience, poly-
amine-related fouling issues may be related to:
Overdosing or locally insufficient dilution/dispersion;
Unfavorable amine distribution ratios - with an impor-
tant role for CHA;
Thermal instability resulting in amine exchange (scram-
bling) reactions;
Excessive fouling (iron oxides, organics like lubricating
oil) from previous treatment/operations.
Controlled dosing is of paramount importance: operators
must be trained not to increase the polyamine addition rate
at the first lowering of pH values in the steam circuit.
Typical locations for deposits are strainers and small di-
ameter instrumentation lines. The sticky nature of a fouling
sample may quickly disappear upon contact with CHA!
Whereas (sticky) deposits in a HP steam system are never
a pretty sight, it is important to realize that with
polyamines, because of their "cleaning" effect, some
degree of localized fouling is acceptable as long as the
uninterrupted safe operation of the steam circuit is guaran-
teed. Manageable fouling includes, for instance, cartridge
filter cakes or strainer deposits that accumulate pre-
dictably and can be periodically removed.
Compatibility with catalysts and resins Polyamine
suppliers have documentation on file on this. In general,
there do not seem to be deleterious interactions between
polyamine products and catalysts or ion exchange resins.
Compatibility with site infrastructure The cracker
polyamine program has been discussed with the Geleen
702
1.8 MPa export steam
to site users
Export condensate
to utility supplier
Air-cooled
Turbines
condensers
Steam users Turbine condensates
Reboiler condensates
Tracing condensates
site utility supplier and the biological wastewater treatment
plant. No corrosion or fouling problems have been
reported by site users of polyamine treated steam. This is
all the more interesting since there is no control over the
CHA-to-polyamine ratio once the cracker export steam is
put onto the 1.8 MPa grid.
EXPERIENCE TO DATE
Ammonia/morpholine containers were replaced with cyclo-
hexylamine/polyamine units in November 2005. The initial
polyamine dosing rate was kept low on purpose to assess
the response of the HP steam system to the new chemi-
cals in terms of pH, (acid) conductivity, fouling, etc. No
changes were made to the blowdown regime in place at
the time.
Three years later, the polyamine program has lived up to
many of our expectations. As will be shown further on, the
steam quality has greatly improved with respect to acid
conductivity. Important water and energy savings have
been realized through blowdown reduction. The time
between regenerations of the cation exchangers in the
cracker s CPU has more than doubled. Inspections have
shown TLEs and drums to be effectively protected against
corrosion by a thin, uniform, adherent, non-dusting, slightly
hydrophobic magnetite layer. Turbines taken out of service
for maintenance during the planned cracker shutdown of
April 2007 were exceptionally clean and free of corrosion
damage, although 1 years of operation at that time is too
short a period to give credit to the new treatment for this.
The (conversion to the) polyamine program has also
revealed its dark side. Fouling incidents involving sticky,
shoe-polish-like deposits have occurred in the pre-boiler
and condensate systems. Cation exchange resin lumping
 Three Years of Experience with Polyamines in the High Pressure Steam System of a Naphtha Cracker
and flotation has been experienced in the CPU. One CHA
container had to be replaced because of the formation of a
floating layer, presumably cyclohexylammonium carbon-
ate, through reaction between CHA and carbon dioxide in
the air.
In the text below our practical experience so far is elabo-
rated.
Feedwater Train Fouling
For reasons beyond our control, the feedwater filter was
not in operation for most of the time between November
2005 and the April 2007 turnaround. During that period,
due to their construction with a stagnant zone, the exiting
water compartments of the feedwater heaters acted as
settlers for suspended solids passing the highly turbulent
deaerator (Figure 6).
The sticky nature of the deposits is indicative of an unfa-
vorable CHA:polyamine ratio. Although the critical micelle
concentration of OPD of about 4 mg kg -1 [29] may occa-
sionally have been exceeded, it is not believed that this has
a direct relation to the fouling we experienced.
Some of the sticky deposits must have made their way into
the drums and TLEs since similar fouling was also
observed in blowdown flash tanks and in the CPU, where it
is retained by the carbon filters. Fortunately, no indications
of fouling were found during waterside video camera
inspections of TLE bottom tubesheets and internal drum
inspections.
This all goes to say that feedwater filtration is a prerequisite
for a successful polyamine program at the Geleen cracker.
The cotton fiber cartridges are now replaced once per year.
Figure 6:
Sticky deposits (most removed) and oily substances in feedwater
heater.
il
In the period of May 2007 to May 2008, around 480 kg of
cake accumulated on the cartridges. Although many of the
cartridges touched, the pressure drop over the filter
remained as low as 0.02 MPa.
The collected cake mass is not representative of steady
state operation of the HP steam cycle in the sense that it is
affected by cracker start-up after the April 2007 shutdown.
It is also expected to reflect ongoing "rummaging" of older
loose deposits in the air-cooled condensers. Nevertheless,
it allows some order of magnitude calculations.
At 50 % moisture content and with negligible concentra-
tions of other elements in the filter cake, whether inorganic
or organic, it follows that 240 kg of iron oxide was col-
lected over a one year period. Assuming all iron is present
as (massive) hematite with a density of 5.3 t rn-3 this
translates into an indicative average Fe203 removal rate of
9 pm per year for an estimated total of 5 800 m 2 internal
surface area of the carbon steel air-cooled condensers.
The amount of iron contained in the cake would amount to
84 kg. Given a density of 7.9 t m -3, this represents a car-
bon steel volume of a mere 10 L.
Whatever its exact nature, periodic replacement of filter
cartridges makes this fouling manageable.
Feedwater, Boiler Water, Steam and Condensate
Quality
Table 1 provides an overview of pH values and conductivi-
ties obtained with the polyamine program compared to
those in internal guidelines for the cracker s steam system.
As mentioned before, feedwater acid conductivity is meas-
ured on-line. It now typically varies between 0.10 and
0.15 pS cm-1 as opposed to 0.15-0.30 pS cm -1 when
morpholine was used.
The acid conductivity range measured in a series of boiler
water grab samples compares well with the average of
about 1 pS cm-1 for our internal reference: the boiler water
of a 12.0 MPa ammonia treated system at the other
cracker in Geleen.
The acid conductivity range of 0.26 to 0.47 pS cm-1 in
condensate was measured on-line as part of an investiga-
tion into the quality of the 1.8 MPa export steam to the
grid. For most of the two days of testing, the acid conduc-
tivity was constant at 0.3 pS cm -1 . This is also a major
improvement from the morpholine era with typically
0.6-0.7 and occasionally 1.2 pS cm -1 or even higher num-
bers. The value of around 0.3 pS cm -1 is in agreement
with earlier results (Figure 7).
703
 Three Years of Experience with Polyamines in the High Pressure Steam System of a Naphtha Cracker

Parameter Feedwater Boiler water Steam/condensate

pH (25 C)
target range 9.0-9.5 8.5-9.5 8.5-9.5
measured range 8.4-9.7 7.8-9.4 8.3-9.6
Specific conductivity in pS - cm-1
target range - -
measured range 1.7-12.4 2.3-9.8 4.7-13.7
Acid conductivity in pS cm-1
target range < 0.2 <3 < 0.2
measured range 0.10-0.22 0.7-2.1 0.26-0.47

Table 1:
Summary of water/steam quality parameters.

It was determined that the TOC concentration of the
1.8 MPa steam at a fixed polyamine product dosing rate
and variable CHA dosing rates ranged from 2.0 to
4.7 mg kg-1 with pH (25 C) between 9.5 and 9.7. In addi-
tion to CHA and filming amines the steam contained
0.5 mg kg-1 each of ammonia (from the utility supplier s
import steam) and ETA.
Interestingly, there was a discrepancy between the results
obtained with Sievers 800 and Shimadzu VCPH TOC
analyzers, the former giving consistently lower results.
With sampling errors and the effect of time between sam-
pling and analysis ruled out, does this leave us with a dif-
ferent response of polyamines to the operating principles
of these two analyzers?
From the combination of the low acid conductivity of
0.3 pS cm-1 and the relatively high TOO concentration of
several ppms it follows that most of the organic carbon in
the steam is associated with amines ("good" TOC) and
little C can be attributed to organic acids ("bad" TOC).
Indeed, no formic, propionic or butyric acid was detected.
The reported acetic acid concentration was not credible
since it would have resulted in a much higher acid conduc-
tivity. At the time of the analyses, the 002 content of the
steam was below 0.2 mg kg-1.
In general terms, the most unfavorable data in Table 1 were
acquired just after the switch to the polyamine program
with deliberate underdosing of filming amine. Stable op-
eration was reached within weeks. Today, a pH (25 C) of
- 9 is maintained throughout the system with typical spe-
cific conductivities for feedwater, boiler water and turbine
condensate of 3.5 to 5, 4 to 6 and 5.5 to 7 pS cm-1,
respectively.

0.9 9.9 In the past 3 years, there have been incidents of overdos-
Morpholine Morpholine
0.8
injection off injection on ing of polyamine: excess concentrations of 5.2 mg kg-1
9.8
have been reported. Over time, we have been able to
c..) 0.7 9.7 decrease the variability in the free polyamine concentration
range from 0.5-2.5 to typically 1-1.5 mg kg -1 as meas-
0.6 9.6
ured via the colorimetric-turbidimetric Bengal Rosa
5 0.
0.5 9.5
method. We are aware that this is still rather high and
a
0
leaves the steam system at increased risk of fouling. For
0 0.4
0
9.4 this reason and economics, of course, a residual
polyamine content of approximately 0.5 mg kg -1 is pre-
0 9.3
ferred.
0.2 It I I 1i if 1 r .9.2
0 0 0 0 E One of the serious shortcomings of a polyamine treatment
Ii ai cd as al
a co CI ID (0 CV 0 CV 0
is the fact that the targeted excess filming amine concen-
Time [h] trations coincide with the detection limit of the current ana-
lytical methods. There is a need for an improved practical
Figure 7: polyamine analysis.
Effect of morpholine dosing on acid conductivity and pH of 1.8
MPa steam [8].

704
Three Years of Experience with Polyamines in the High Pressure Steam System of a Naphtha Cracker
Condensate Polishing Unit Issues
The polyamine program has allowed us to increase the
time interval between regenerations of the cation exchang-
ers in the CPU from 6 to 13 days. On top of considerable
financial savings, fewer regenerations in the largely manu-
ally operated CPU have liberated operator time for other
work in the cracker.
During start-up of the cracker after the 2007 turnaround,
3 m3 of macroporous, strongly acidic cation exchange
resin had to be replaced because of unexplained lumping
and flotation problems in one cation exchanger. The prob-
lem has not recurred and the performance of the cracker s
CPU has been very good otherwise.
Ever since the ammonia/morpholine to cyclohexylamine/
polyamine conversion, the site utility supplier treating the
1.8 MPa condensate of cracker export steam has had to
double the number of regenerations of mixed bed ex-
changers. This seems to be due to increased carbon di-
oxide loading of the anion exchange resin. To date, the root
cause of this phenomenon remains unclear.
The ultimate CPU related challenge would be to by-pass
the cracker s condensate treatment plant under normal
operating conditions. The CPU would be kept on stand-by
and operated only in case of water/steam quality prob-
lems.
Blowdown Reduction
In the course of the polyamine program, the continuous
blowdown of the HP steam cycle has been reduced by
75 %. This represents significant water and energy sav-
ings.
So far, the discontinuous "bottom" blowdown frequency of
Borsig TLEs has not been adapted. It is expected that this
ultimately will be the only blowdown required for control-
ling the boiler water quality between decokings. Trials with
no continuous blowdown are being carried out.
Major blowdown reduction would not have been possible
without the excellent demineralized water quality at our
site. Indeed, our boiler water silica limit of 0.25 mg kg-1
5i02 is rarely exceeded, making the pH the primary blow-
down controlling parameter until on-line analyzers for
boiler water acid conductivity are in place.
Chemical Handling Issues
The use of a 50 % CHA solution was abandoned because
the low flash point of its corresponding vapors, 34 C, was
incompatible with the ATEX zoning of the dosing area in the
plant. Highly concentrated CHA solutions seem to be sen-
sitive to the formation of a creamy cyclohexylammoniu car-
bonate(?) film when exposed to air.
CONCLUSIONS AND FINAL REMARKS
At one of SABIC Europe s naphtha crackers in Geleen,
Netherlands, the ammonia/morpholine treatment of the HP
steam cycle was converted to a polyamine program in
November 2005. Three years later, SABIC experience
shows that a 12.5 MPa cracker steam system can be
safely, reliably and cost-effectively operated using
polyamines.
The choice for polyamines has been a very conscious one
based on a decade of operating experience at lower pres-
sures, several in-house research projects and many
exchanges with other users of polyamine products and
suppliers.
With polyamines, the focus of the alkalization program
changes from bulk water chemistry to steel/water interface
chemistry. The polyamine approach is to keep boiler tubes
clean from porous magnetite deposits resulting from iron
transport due to flow-accelerated corrosion and first con-
densate corrosion in pre-boiler and condensate systems.
In this way, corrodents do not get a chance to concentrate
in high heat flux areas in boilers. This makes steam sys-
tems treated with polyamines more resistant to water qual-
ity upsets and - within reason - permits safe extension of
the narrow bulk water chemistry operating windows dic-
tated by EPRI, TOV, VGB, etc.
Polyamines are no panacea. In particular, unpredictable
fouling in the form of sticky oily or waxy deposits is and
remains a concern with polyamine-based steam cycle
treatment programs. Such fouling may result in plant
"trips" due to plugging of, for instance, instrumentation
lines or pump strainers. Typically, the steam system is most
vulnerable to fouling problems in the early stages of the
conversion phase from a conventional to a polyamine
treatment program.
It is most regrettable that the public polyamine knowledge
base is so limited today. It is felt that the "open ends" asso-
ciated with the application of polyamines deter many
potential users.
We expect EPRI, TOV, VGB, etc., to finally accept and
embrace polyamines as a credible alternative to conven-
tional water/steam treatment programs and to incorporate
polyamine programs in future guidelines. We urge poly-
amine users to accurately document and publish their
cases. We call on technical institutes - together with sup-
pliers - to initiate new research projects in the field of
polyamine/steel interface chemistry, fouling mitigation,
thermal stability and other related topics.
705
 Three Years of Experience with Polyamines in the High Pressure Steam System of a Naphtha Cracker
REFERENCES
[1] Savelkoul, J., Lier, R. van, PowerPlant Chemistry
2005, 7(12), 733.
[2] Duisters, H., Savelkoul, J., Proc., 50 th Safety in
Ammonia Plants & Related Facilities Symposium,
2005 (Toronto, ON, Canada). AlChE, New York, NY,
U.S.A.
[3] Heins, L., Schlizio, H., VGB Kraftwerkstechnik 1981,
61(2), 147.
[4] VGB R 450 Le, Second Edition, 2004.
Enqvist, K., Luoto, M., Proc., EPC Spring Meeting,
[5]
2001 (Houston, TX, U.S.A.). AlChE, New York, NY,
U.S.A., Paper no. 77d.
[6] Enqvist, K., Proc., EEPC Annual HSE Meeting, 2006
(D sseldorf, Germany). Global Technology Forum,
Surrey, UK.
[7] Dooley, R. B., PowerPlant Chemistry 2008, 10(2), 68.
[8] Savelkoul, J., Janssen, P., Verhoef, H., PowerPlant
Chemistry 2001, 3(6), 326.
[9] Moran, F., Rosenbluth, G., Proc., 1 ()ernes Journ es In-
formation Eaux, 1992. Association de Professionnels
du Traitement des Eaux et des Nuisances, Ecole
Sup rieure d Ing nieurs de Poitiers, Poitiers, France,
Paper No. 47.
[10] Beetz, G. E., Fernw rme International 1992, 21(11),
616.
[11] Moldenhauer, D., Energieanwendung + Energietech-
nik 1993, 42(4), 208.
[12] Schulze, U., Hopp, G., Schulze, P., Proc., VGB-
Sondertagung - Organische Konditionierungs- und
Sauerstoffbindemittel, 1994 (Lahnstein, Germany).
VGB-Kraftwerkstechnik GmbH, Essen, Germany,
VGB-TB 450, Paper V16.
[13] Hopp, G., VGB PowerTech 2001, 81(11), 68.
[14] H vel, B., Proc., VGB-Sondertagung - Organische
Konditionierungs- und Sauerstoffbindemittel, 1994
(Lahnstein, Germany). VGB-Kraftwerkstechnik
GmbH, Essen, Germany, VGB-TB 450, Paper V19.
[15] Schontag, W., Proc., Feedwater and Boiler Water
Treatment in Industrial, Co-Generation and Refuse
Incineration Plants and in Units with Heat Recovery
Steam Generators, 2000 (Mannheim, Germany).
PowerPlant Chemistry GmbH, Neulussheim, Ger-
many, Paper #4.
[16] Schontag, W., Modern Power Systems 2001.
[17] Roothooft, R., Eyckmans, M., Verheyden, K., De
Pourcq, D., VGB Kraftwerkstechnik 2000, 80(12), 102.
[18] Verheyden, K. S., Ertryckx, R. A. M., De Wispelaere,
M., Poelemans, N., PowerPlant Chemistry 2003, 5(9),
516.
706
[19] De Wispelaere, M., Proc., Interaction of Organics and
Organic Cycle Treatment Chemicals with Water,
Steam and Materials, 2005 (Stuttgart, Germany).
Electric Power Research Institute, Palo Alto, CA,
U.S.A. and PowerPlant Chemistry GmbH, Neuluss-
heim, Germany.
[20] Allard, B., Chakraborti, S., Kannappan, A., Kemisk
Tidskrift 1982, 9, 107.
[21] Allard, B., Chakraborti, S., Svensk Papperstidning
1983, R 186.
[22] Loay, M., Grabli, A., PowerPlant Chemistry 2006,
8(5), 303.
[23] Vogt, T., Besl, G., Stecklina, M., PowerPlant
Chemistry 2007, 9(8), 500.
[24] Kelm, W., Kranz, H. D., Vrhel, D., PowerPlant
Chemistry 2000, 2(10), 604.
[25] Kelm, W., Proc., Feedwater and Boiler Water
Treatment in Industrial, Co-Generation and Refuse
Incineration Plants and in Units with Heat Recovery
Steam Generators, 2000 (Mannheim, Germany).
PowerPlant Chemistry GmbH, Neulussheim, Ger-
many, Paper #6.
[26] Kelm, W., Bursik, L., Proc., Interaction of Organics
and Organic Cycle Treatment Chemicals with Water,
Steam and Materials, 2005 (Stuttgart, Germany).
Electric Power Research Institute, Palo Alto, CA,
U.S.A. and PowerPlant Chemistry GmbH, Neu-
lussheim, Germany.
[27] Turyna, R., Proc., Organische Konditionierungsmittel
f r den Wasserdampfkreislauf, 2007 (Heidelberg,
Germany). PowerPlant Chemistry GmbH, Neuluss-
heim, Germany, Paper #6.
[28] Winter, H., Gaisser, K. E., VGB PowerTech 2005,
85(7), 94.
[29] Bohnsack, G., VGB Kraftwerkstechnik 1997, 77(10),
841.
[30] Frahne, D., Proc., Feedwater and Boiler Water Treat-
ment in Industrial, Co-Generation and Refuse
Incineration Plants and in Units with Heat Recovery
Steam Generators, 2000 (Mannheim, Germany).
PowerPlant Chemistry GmbH, Neulussheim,
Germany, Paper # 3.
[31] Frahne, D., Blum, T., PowerPlant Chemistry 2006,
8(1), 21.
[32] Frahne, D., Blum, T., Proc., Organische Konditio-
nierungsmittel far den Wasserdampfkreislauf, 2007
(Heidelberg, Germany). PowerPlant Chemistry
GmbH, Neulussheim, Germany, Paper #3.
[33] Steinbrecht, D., VGB PowerTech 2003, 83(9), 120.
[34] Steinbrecht, D., Topp, H., Proc., Organische Kon-
ditionierungsmittel f r den Wasserdampfkreislauf,
2007 (Heidelberg, Germany). PowerPlant Chemistry
GmbH, Neulussheim, Germany, Paper #9.
r
Three Years of Experience with Polyamines in the High Pressure Steam System of a Naphtha Cracker
[35] Moran, E, Rocher, S., Apave 1977, No. 198.
[36] Hater, W., Olivet, D., Proc., Fachtagung Dampf-
erzeuger, Industrie- und Heizkraftwerke, 2008
(Papenburg, Germany). VGB PowerTech, Essen,
Germany, Paper V11.
[37] Bursik, A., Bezzoli, P., Graf, A., PowerPlant Chemistry
2003, 5(6), 373.
[38] Bursik, A., PowerPlant Chemistry 2004, 6(9), 549.
[39] Apblett, A., Kuriyavar, S. I., Foutch, G. L., Tran, R,
PowerPlant Chemistry 2004, 6(9), 523.
[40] Raught, D. P., Foutch, G. L, Apblett, A., PowerPlant
Chemistry 2005, 7(12), 741.
ACKNOWLEDGMENTS
The authors wish to thank the management of SABIC
Europe for permission to publish this paper.
This paper is an edited and extended version of a paper
presented at the 2nd International Conference Interaction
of Organics and Organic Plant Cycle Treatment Chemicals
with Water, Steam, and Materials, Lucerne, Switzerland,
November 4-6, 2008.
THE AUTHORS
Roy van Lier (M.Sc., Raw Materials Processing, Delft
University of Technology, Delft, Netherlands, M.A.Sc.,
Hydrometallurgy, The University of British Columbia,
Vancouver, Canada) previously worked as a process metal-
lurgist in the international mining industry on various
assignments involving multiple metals, but primarily zinc.
In 2001, he joined the DSM Materials and Corrosion
Department in Geleen, Netherlands. Here, he worked
together with Jo Savelkoul on all kinds of corrosion and
water treatment issues in various (petro)chemical plants.
Since 2005 he has succeeded Jo Savelkoul, working as a
corrosion and water treatment specialist with the SABIC
Competence Center Manufacturing in Geleen. Roy van Lier
is an active participant in several (inter)national networks
including the Corrosion Group of the Dutch Chemical
Industry (CCNCI) and the VGB Working Group on Industrial
Power and Co-Generation Plants.
Gerard Janssen (B.S., Mechanical Engineering and
Process Technology, Higher Technical School, Heerlen,
Netherlands) worked with DSM in Geleen, Netherlands,
from 1980 to 2005. During that period, he held engineering,
maintenance and process technology focused positions as
a staff member at different chemical plants at the Geleen
site. In 2005, he joined the SABIC cracker staff as an
energy/utility engineer. Since the beginning of 2008, he has
been working on longer term and strategic energy and util-
ity related projects with the SABIC Site Improvement
Department in Geleen.
Jo Savelkoul (B.S., Physical and Inorganic Chemistry and
Environmental Science, Higher Technical School, Heerlen,
Netherlands) worked with DSM in Geleen, Netherlands,
from 1961 to 2004. From 1979 to 2004 he was a senior
corrosion and water treatment specialist, and head of the
Water Laboratory of the DSM Materials and Corrosion
Department. Jo Savelkoul has spent many years working
on the development of water treatment programs, monitor-
ing and test methods for cooling water, and the mitigation
of corrosion in industrial steam systems. He has published
several papers on these topics. Jo Savelkoul has been a
member of the Dutch NEN (National Standardization
Institute), the NCC (Netherlands Corrosion Center) and
international ISO task forces for various activities in the
field of industrial water treatment. Since his retirement in
2004, he has been working as a private consultant in the
fields of corrosion and water treatment.
CONTACT
Roy van Lier
SABIC Europe BV
Site Improvement
Competence Center Manufacturing
Geleen
Netherlands
E-mail: roy.van.lier@SABIC-europe.com
z-1^-
707
To Whom It May Concern

Geleen, September 12, 2014

The high-pressure steam system of one of SABIC’s naphtha crackers in Geleen,

Netherlands, has been successfully treated with products based on film forming
amines since 2005. In 2009, the cracker changed from competitor products to a
Cetamine® treatment.

A substantial amount of information on this application is in the public domain1,2.

During a planned shutdown in October 2013 one of the cracker’s condensation/

backpressure turbines was opened for inspection and maintenance. It had been in
continuous operation for 12 years.

The turbine was found to be exceptionally clean and in excellent mechanical state.
The pictures above show partial views of the high-pressure/medium-pressure
section (left) and low-pressure section of the rotor, all in “dirty” state.

Roy van Lier, M.Sc., M.A.Sc.

Chief Scientist, Technology & Innovation

Gerard Janssen, B.Sc.

Senior Utilities & Rotating Equipment Engineer, Technical Department

1
Van Lier et al., VGB PowerTech 2012 (8), 84
2
Van Lier et al., PPChem 2008 10 (12), 696
Preventing ammonium salt fouling
and corrosion
Best practices and opportunities to reduce the risk of salt
deposition and damage
BERTHOLD OTZISK, FAUSTO MAGRI, JELLE ACHTEN and SANDER HALSBERGHE
Kurita Europe
S
alt deposits and corrosion
can lead to damage or high
energy losses when crude
oils are processed. Usually,
these salts are ammonium chlo-
ride (NH4Cl) or ammonium
bisulphide (NH4HS). This arti-
cle describes a new approach
to prevent chloride corrosion
or further deposition of ammo-
nium salts in order to avoid or
minimise corrosion and fouling
potential.
The use of a powerful water
washing system is certainly a
good step in the right direc-
tion to wash out as many salts
as possible. Ammonium salts
are generally readily soluble in
water, but can often not be com-
pletely removed in the pres-
ence of hydrocarbons. Process
units suffering ammonium salt
fouling or corrosion are crude
distillation units, FCC units,
hydrocrackers, hydrotreaters
and reformer units.
Higher amounts of chlo-
rides in residue feedstocks or
low main fractionator top tem-
peratures to produce low sul-
phur gasoline are two reasons
for salt fouling in FCC units.
Sometimes a tower washing
programme can be performed
www.eptq.com
to remove water soluble salts.
For this purpose, water is fed
into the top reflux, or the over-
head temperature is lowered
until enough water can con-
dense to the top trays. This
usually takes several hours or
even days. The feed rate must
be significantly reduced by
20-30% during this time. The
produced naphtha, sometimes
also the light cycle oil (LCO),
goes off-specification and has to
be reprocessed with increased
costs. In addition to the lower
throughput, these costs may be
particularly high.
The addition of an oil solu-
ble salt dispersant additive can
help to prevent salt deposition
and keep the salts transporta-
ble, but that practice may also
lead to several disadvantages.
They work by chemically bind-
ing to the salt deposits, where
the lipophilic functional group
keeps microcrystalline salts in
the hydrocarbon phase until
enough free water is present
so that the salts can be solubi-
lised later. The dispersed salts
are carried out of the system
with the hydrocarbon prod-
uct flow. Such salt dispersants
are often surface active com-
ponents with filming amine
properties that can stabilise
emulsions. In the event that not
all of the bound salts are solu-
bilised in free water later, they
may remain in the hydrocarbon
phase of the product steam, to
be transported to downstream
processes.
ACF technology
Kurita’s patented ACF tech-
nology pursues a completely
different approach. Liquid for-
mulations of a very strong
organic base called ACF are
used. These are water sol-
uble and do not react with
hydrocarbons. The organic
base ACF has a very low pKb
value close to zero, which is
an indicator of a very strong
base. It reacts preferentially
with strong acids such as
hydrochloric acid (HCl) or its
corresponding ammonium salts
(NH4Cl). The reaction prod-
uct is ACF-Cl, which is a liq-
uid salt with a neutral pH of 7.
ACF salts have very high mois-
ture absorption characteristics
(highly hygroscopic) and can
be removed easily from the sys-
tem with free water. One more
benefit is that ACF salts have a
PTQ Q4 2017 1

ACF
NH3 HCI
Metal surface
NH4 Cl deposit
High corrosion potential
Figure 1 Removal of ammonium salt deposition and corrosion prevention
very low corrosivity. This is a
significant difference compared
with conventional neutralising
amine salt reactions where the
formed neutralising amines are
almost always very corrosive
and require higher amounts of
wash water for removal.
The favoured reaction of
ACF with HCl is a significant
advantage in process units with
naturally high H2S concentra-
Expected
fouled
section MC
reflux
ACF
Naphtha
stripper
Figure 2 FCC main fractionator top section
2 PTQ Q4 2017
phate ((NH4)2SO4). In all cases,
ACF displaces the weaker base
HCI
ammonia by forming a liq-
uid salt. In the absence of free
water, the salts are transported
NH3 ACF at temperatures below 180°C
NH3 together with the hydrocar-
Metal surface bon stream. When they come
into contact with water, the
ACF ACF HCI salts can be removed with the
aqueous phase. This chemical
Metal surface
programme can be applied con-
Liquid ACF-Cl
NH3 HCI Excellent fluidity tinuously to prevent fouling
Low corrosion potential and corrosion or it is used at
Metal surface higher dosing rates to remove
already formed salt deposits.
Case study 1: FCC main
fractionator
Since it is common practice for
tions, such as hydrotreaters the top temperature in FCC
or hydrocrackers. Only after main fractionator columns to
the conversion into ACF-Cl be lowered to produce more
can weaker acids like hydro- low sulphur gasoline, negative
gen sulphide (H2S) or its salts impacts are often observed. In
react with the ACF by forming most cases, salt fouling is the
the corresponding ACF reac- reason and this can also lead to
tion product. The reaction with corrosion if the salts are present
ACF is shown schematically on the metal surface as a sticky,
in Figure 1, which is also valid highly concentrated and viscous
for other ammonium salts such solution.
as NH4HS or ammonium sul- Negative impacts are:
• Increased pressure drop
• Plugged trays and product
draws
Wet gas
condensation • Increased corrosion rates
• Flooding of the main fraction-
ator top section
Wet gas
condensation • Efficiency losses between gas-
ACF
SWS oline and LCO separation
• Wet gas compressor efficiency
losses.
During a field trial, the
Naphtha
circulation online cleaning performance of
ACF was tested in a Western
European refinery. Chemical
injection into the main fraction-
ator overhead reflux is common
to dissolve the deposited salts
from the top trays. According
to the refiner’s experience,
naphtha circulation was also
www.eptq.com
affected where salt deposits on
the lower trays were assumed
to be present. Therefore ACF
was additionally injected into
the naphtha circulation in order
to remove salt deposits from
the heat exchangers and from
the lower trays of the fraction-
ation column. Figure 2 shows
the two dosing points where
ACF was continuously injected
during the online cleaning. The
reaction with the chloride salts
is a stoichiometric balance and,
according to the expected high
amounts of salt deposits, a few
litres per hour of ACF were
dosed at both locations to mobi-
lise high amounts of salts.
After the dosage had been
started, first successes were
observed within less than 30
minutes. Figure 3 shows some
recorded results from the con-
trol room with basic data before
the treatment and significant
changes during the online
cleaning. The differential pres-
sure of the main fractionator
column as a direct key indica-
tor for fouling, the FCC feed
rate, and the heat transfer coef-
ficient of the naphtha circula-
tion are displayed graphically
in Figure 3. The rapid increase
in the heat transfer coefficient
and the main fractionator dif-
ferential pressure drop impres-
sively demonstrate that higher
amounts of salt deposits were
dissolved and mobilised in a
short period of time.
Case study 2: hydrotreater
feed/effluent exchangers
In a naphtha hydrotreater
operating with several feed/
effluent heat exchangers, corro-
sion problems occurred in the
space of a few years, mainly
due to the formation of ammo-
nium salt deposits. As a correc-
www.eptq.com
Usual operation
Heat transfer coefficient
Feed rate
Differential pressure
1st 2nd 3rd
day day day
Figure 3 FCC main fractionator online cleaning results
tive action, ACF was injected
between the second and third
heat exchangers after the reac-
tor with the aim of stopping
corrosion in the third feed/
effluent exchanger and airline
coolers. Since the treatment
was started, large amounts of
chlorides were collected in the
water boot of the accumulator
from where they could be fre-
quently removed. Despite the
higher chloride concentration in
the accumulator, the corrosion
potential was significantly low-
ered since ACF salts have a very
low corrosion potential. Figure 4
shows the location where ACF
is continuously dosed with a
low ppm injection rate into the
effluent product stream.
After some months of opera-
tion, when no damage occurred,
the ACF programme was
To
reactor Feed
From
reactor
ACF
Figure 4 Naphtha hydrotreater
feed/effluent exchangers
ACF
4th 5th 6th
day day day
extended with an additional
function of film forming prop-
erties. The metal surfaces are
thus directly protected from
acid attack and, because of the
smooth and clean metal sur-
face, any salt deposition is pre-
vented. The result is longer time
in operation with clear energy
savings. The duty calculations
and hot approach tempera-
ture (HAT), where the thermo-
dynamics of heat recovery can
be determined by superimpos-
ing the curves showing heat
demand from the reactor feed
and heat release from the reac-
tor effluent, are good key per-
formance indicators.
Figure 5 shows the heat
exchanger duty trends and
calculated energy savings in
comparison to the previous
operational mode over a period
of two years since the applica-
tion was started. The success of
the treatment is demonstrated
by the fact that no further cor-
rosion damage has occurred so
far, as well as by an improved
energy balance of the feed/
effluent exchangers.
Case study 3: debutaniser
cleaning
Severe flooding problems were
PTQ Q4 2017 3

8.0
7.5 7.29
7.0
Duty, MMkcal/h
6.5
6.38
6.0
5.5
5.47
5.0
4.5
4.0
3.5
3.0
0 365
Days of operation
Figure 5 Feed/effluent exchanger energy savings
observed in the debutaniser
column of a naphtha hydro-
treater unit. The results of a
root cause analysis pointed out
that mainly ammonium salts
and corrosion products such as
iron sulphides were the orig-
inators. The debutaniser col-
umn is equipped with valve
trays, where it was expected
that many of the top trays were
heavily fouled and blocked.
This resulted in a high differen-
tial pressure increase, excessive
liquid return to the bottom sec-
tion, and a significantly reduced
throughput. Clear indicators
for flooding were the high lev-
els of C5 hydrocarbons in the C4
section, due to a decline in sep-
aration (off-spec production),
and the high Reid vapour pres-
sure of the isomerate. The debu-
taniser bottom reboilers had
to be stopped to keep the unit
running.
Previously, water wash-
ing had been carried out with
only mediocre success, where
not all fouling deposits were
removed and the naphtha
hydrotreater had to be stopped
for one to two days. This was
4 PTQ Q4 2017
ACF treatment
Blank run
5.47
3.65
730
not an option this time because
it would involve a shutdown of
the reformer and isomerisation
unit. Mechanical cleaning of the
debutaniser column would also
result in unwanted shutdowns
of the reformer and isomerisa-
tion unit for three to five days.
For these reasons, online clean-
ing with ACF was selected to
put the debutaniser back on
track.
High amounts of fouling
deposits were expected in the
top section of the debutaniser
column. To reach that area for
direct contact of ACF with the
fouled salts, there was only one
possible location for the injec-
tion, which was the suction side
of the overhead reflux pump.
While about 80% of the liquid
phase returns back to the debu-
taniser top section, about 20% is
passed via a reflux pump to the
LPG unit.
Taking this into account, the
ACF dosage was started and
showed immediate improve-
ments. Over a period of six
days, ACF was injected con-
tinuously, where water sam-
ples were taken several times
per day from the overhead
accumulator to report the sta-
tus of the online cleaning. The
pH, conductivity, chloride
and ammonium concentra-
tions and total nitrogen con-
tent were analysed. During the
online cleaning, the pH value
stayed relatively stable at the
same level of approximately
8.5 because of the presence of
NH3. Directly after the start of
the injection, the conductivity
immediately leapt up as a func-
tion of dissolved chloride salts.
A slight increase in the pH at
the end of the online cleaning
was a direct indicator that injec-
tion could be stopped. This was
also confirmed by the fact that
the debutaniser was again run-
ning under normal operating
conditions.
Conclusions
Each oil refinery is aware of the
high risks of ammonium salt
deposits, which can quickly
lead to corrosion damage to
plant equipment and represent
a ubiquitous risk for the con-
tinuous operation of a process
unit. Water washing is a first
step in the right direction, but
sometimes with only mediocre
success. Therefore there is great
interest in avoiding corrosion
attack and in finding powerful
opportunities for the removal or
prevention of fouling materials
such as ammonium salts.
Kurita´s ACF technology is
a completely new approach,
where corrosive HCl or ammo-
nium salts are directly chemi-
cally binded. The formed ACF
salts have a low corrosivity and
can be removed easily with free
water. This approach is com-
pletely different to technologies
where additives with dispers-
ing properties are used that
www.eptq.com
simply shift the salts to another
location. In a refinery, there are
many opportunities to use ACF
during online cleaning or for
continuous corrosion protection
and fouling removal treatment.
For the operating refinery, this
represents significant added
value since production losses
caused by fouling or corrosion
can be reduced to a minimum.
www.eptq.com
References
1 Melin M, Baillie C, McElhiney G, Salt
deposition in FCC gas concentration
units, PTQ Q4, 2009.
2 W A Al-Naim, Ammonium bisulphide
corrosion in hydrocrackers, PTQ Winter
1999/2000.
Berthold Otzisk is Senior Product
Manager for refinery and petrochemical
applications with Kurita Europe, Germany.
Email: berthold.otzisk@kurita.eu
Fausto Magri is Key Account Manager for
refinery and petrochemical applications
with Kurita Europe, Italy.
Email: fausto.magri@kurita.eu
Jelle Achten is Area Manager for Benelux
with Kurita Europe, Germany.
Email: jelle.achten@kurita.eu
Sander Halsberghe is Project Engineer
for Benelux with Kurita Europe, Germany.
Email: sander.halsberghe@kurita.eu
PTQ Q4 2017 5
 Masakazu Koizumi
and Keisuke Karaki,
Kurita Water Industries Ltd,
Japan, present a new
approach to counter salt
deposition for corrosion
and fouling control.
A
mmonium chloride salt deposition is a common
problem in petroleum refining plants. This is
because chlorides in crude oil become hydrolysed
to hydrochloric acid (HCl) and nitrogen content
contained in oil is converted to ammonia (NH3) in
hydroprocessing units or the like. These substances are
present throughout a refinery’s units and sometimes generate
ammonium chloride salt deposits, which can block the flow
of process liquids and gas in a unit. This can result in under
deposit corrosion as the deposit absorbs moisture.1 Salt
deposition can commonly occur in the upper part of
atmospheric distillation towers (crude distillation unit
[CDU]), in the effluent system of hydrodesulfurisation units
(HDS), in the upper part of fluid catalytic cracking (FCC) or
residue fluid catalytic cracking (RFCC) main fractionators and
delayed coker units (DCU), as well as the debutanisers of
continuous catalytic reformers (CCR), etc. This article will
introduce a new approach to counter salt deposition:
ammonium chloride free (ACF) technology.
Ammonium chloride salt
The generation of ammonium chloride salt (solid) is
generally expressed in the following formula:
NH3 (gas) + HCl (gas) --> NH4Cl (solid) (reaction formula 1)
When the partial pressure of NH3 and HCl reach a
critical level, NH3 and HCl can no longer remain in the gas
phase and deposit as solid. In other words, as the relative
concentrations of NH3 (gas) and HCl (gas) increase or the
system gets cooler (Figure 1), the depositing tendencies of
NH4Cl (solid) increase.2
Sources of NH3 can be traced back to the nitrogen
content in the crude oil or fraction oils. In some cases,
nitrogen in the oil can be related to NH3 and amines that are
added as a neutraliser. On the other hand, the sources of HCl
are generated by chloride contained in crude oil as salts
(MgCl2 and CaCl2), as well as organic Cl added for catalyst
regeneration at the CCR platformer, or in some cases
contained in the crude oil through upstream contamination.
The main problems caused by salt deposition can be
categorised as flow path blockage due to salt deposition
and local corrosion under the deposited salt.
 The blockage of process flow is generally detected by
increasing differential pressure in the piping and distillation
columns. However, when deposition occurs on the trays
inside the tower, the distillation performance can be
significantly impacted (due to the reduction in the active
area of the trays or uneven process flows over the trays or
other areas of the tower).
When these problems continue, the refinery will be
forced to either reduce charge to the plant or proceed to
emergency shut down. While salt deposits can be dissolved
with a sufficient amount of wash water injection, if the
dissolution is insufficient, the salt deposits are left in the
system as a highly concentrated, thick, viscous solution.
Because salts such as ammonium chloride are acidic, a thick
Figure 1. The scheme of ammonium chloride salt salt solution is highly corrosive. Therefore, if salts are not
generation.
completely removed out of the system, local corrosion can
be accelerated. Another option to
counter salt deposition is a
chemical injection method (oil
soluble salt dispersant).3, 4
Previously, oil soluble salt
dispersants were the only option
in the chemical approach.
However, the application of this
dispersant merely shifts the
problem of ammonium chloride
downstream, possibly causing
Figure 2. NH4Cl eliminating performance with ACF (reaction formula 2).
deposition and fouling in a
different location.

An alternative
ACF is a highly effective chemical
approach that is capable of
eliminating deposited salts from
an operating system. The
conventional chemical approach
to eliminating these salts had
been conducted by application of
an oil soluble dispersant. The salt
dispersant is also capable of
removing the salt from one
location to another point in the
system. However, when the salt
disperses into the process stream,
Figure 3. Prevention of NH4Cl generation with ACF (reaction formula 3). the process stream has to be sent
to the slop because the salts
remain in the oil phase. In some
cases, the salts are removed from
one location after high dose
injection of salt dispersants to
the tower, and then found in a
downstream heat exchanger.
ACF is a patented technology
with the three major
characteristics: replacement
action, high moisture absorption,
and low corrosiveness. It is a
strong base substance, capable of
Figure 4. The result of a moisture absorption/liquefaction test in the laboratory. replacing the ammonia from the
salt that it comes into contact

Reprinted from November 2017 HYDROCARBON

ENGINEERING

Figure 5. The corrosion rate of each 50% HCl salt
aqueous solution to carbon steel (90°C, and removed
dissolve oxygen using N2 gas).
Figure 6. Salt deposition point and injection point of
ACF in simple flow chart of RFCC.
Figure 7. The relationship between the differential
pressure of the main fractionator, the ratio of VR in
the feed oil, naphtha 95% distillation temperature
and the dosage of ACF.
with (reaction formula 2 and Figure 2). When ACF comes in
contact with chlorides, ACF-HCl salt is preferentially
produced (reaction formula 3 and Figure 3).
NH4Cl + ACF --> ACF-HCl + NH3 (reaction formula 2)
NH3 + ACF + HCl --> ACF-HCl + NH3 (reaction formula 3)
The technology also incorporates a high moisture
absorption property. Even in the absence of liquid water,
ACF salts can absorb moisture from the surrounding stream
and form a highly fluid liquid form. Therefore, the liquefied
ACF-HCl salts are capable of flowing out of the stream
(Figure 4). These salts have low corrosive features compared
with ammonium chloride salts and other amine
hydrochloride salts (Figure 5). This is due to the fact that
ACF-HCl salt forms a near neutral solution because of the
technology’s basic properties. This helps to minimise the
potential for corrosion after moisture absorption (and
eventual liquefaction).
Case study
In a recent case, the main fractionator of an FCC (RFCC)
unit was experiencing severe delta pressure build up in the
upper section (Figure 6).
The differential pressure increase was detected from
the beginning of operation, and it was believed to be
caused by ammonium salt deposition on one of the trays
due to the high chloride content in the RFCC feed.
The estimated ammonium chloride deposition
temperature was calculated based on chemical analysis
results and operation conditions. The results indicated
positive deposition of ammonium chloride salts in the
location that was experiencing issues.
However, shutting down the RFCC and related units was
the least preferable option.
Therefore, an ACF programme was applied to resolve
the deferential pressure build up in the tower during
operation. Careful assessment of the unit was conducted
to provide maximum ACF performance and the top reflux
line from the overhead (OVHD) section was selected as the
injection point.
To control salt deposition inside the tower with this
technology, there are two main approaches: the shock
dosing method and the mitigating method.
The first method uses the characteristics of ACF most
effectively. By injecting a temporary high dose of ACF,
ammonium chloride salts are effectively removed from the
surface of the trays, causing limitation to operation with
low corrosion risk. As a result, a normal level of operation is
recovered. This method is mainly used in emergencies
where rapid deferential pressure increase is occurring in a
short time period.
Figure 7 shows a case where the differential pressure
rapidly increased over time, due to a change in the
property of the feed oil. In this particular case, the feed oil
of the RFCC was a mixture of atmospheric residue (AR) and
vacuum residue (VR). The sharp build up of differential
pressure was observed when the mixing ratio of VR
increased in the range of 3 – 4% to 10%.
Simultaneously, deterioration in the separation of
naphtha and light cycle oil (LCO) was noted, indicating a
reduction in the effective tray surface area due to salt
deposition.
Increasing the injection dose rate from 8 ppm to
37.5 ppm (as shown in Figure 7) recovered the differential
pressure and separation deterioration of naphtha and LCO
in operation in a relatively short time.
HYDROCARBON Reprinted from November 2017
ENGINEERING

Figure 8. The relationship between the increasing
rate of differential pressure and the dosage of ACF
during a common period of operation.
The ACF-HCl salt produced by replacing the ammonium
chloride salt, absorbed the moisture in the tower,
liquefying the salts on the trays. The generated ACF-HCl
salt flowed out of the tower from the LCO like liquid, but it
was not dissolved in oil, as explained previously. The
liquefied ACF-HCl salt is then theoretically discharged from
the RFCC from the separated free water at coalescers in the
LCO rundown, or at intermediate tanks of the LCO.
Applying the ACF in this plant did not negatively impact
the products.
Reprinted from November 2017 HYDROCARBON
ENGINEERING
The second method is a strategy for preventing the
differential pressure build up by formation of ammonium
chloride salt during normal operation utilising formula 2.
Careful assessment of the unit in combination with the
second method enables the differential pressure to be
controlled with continuous low dosage injection of
approximately 10 ppm or less, depending on salt levels.
Summary
The application of ACF can effectively eliminate and
prevent deposits of ammonium chloride salt from
generating in an operating distillation column. The
technology can be applied in FCC/RFCC, as well as
CDUs, vacuum distillation units (VDUs), etc. Careful
investigation and the right chemical approach to the
right location can enhance the operational flexibility of
the units.
Applying ACF can also control or eliminate corrosion
in the effluent system of HDS, in the overhead system of
the HDS distillation tower, and in the circulation line of
the BTX extraction unit (sulfolane). It can also eliminate
ammonium chloride salt in the burner tip of a furnace.
References
1. TOBA, K., ‘Zairyo-to-Kankyo’, 62, pp. 352 – 358, (2013).
2. WU, Y-M., Oil & Gas Journal, pp. 38 – 41, (January 1994).
3. MELIN, M., BAILLIE, C. and MCELHINEY, G., Grace Davison
Catalagram, issue no. 107, pp. 34 – 39, (2010).
4. BARIETTA, T. and WHITE, S., PTQ, pp. 53 – 59, (Q3 2007).