Acidic and basic contaminants in

amine treating
S
ources of heat stable salt (HSS) contami-
nants in amine treating solutions are
discussed and their effect on reaction equi-
libria in acid gas-amine systems is analysed. The
requirement of solution electro-neutrality is used
to validate solution analyses. Case studies
demonstrate the important role of HSSs in
determining amine plant performance, that
simulations must incorporate ionic contami-
nants into their calculations, and the need to
simulate the whole plant rather than isolated
units.
Introduction
A common and proven way to remove the acidic
components CO
2
and H
2
S from gas streams is by
absorption into aqueous solutions containing
one or more amines. Use of alkanolamines dates
from the patent awarded to Bottoms in 1930.
However, despite nearly 80 years of commercial
use, further developments in amine treating
continue to appear each year. Triethanolamine
(TEA) was the frst alkanolamine applied
commercially but today it has been completely
supplanted by a large number of other amines
used singly or in mixtures with other amines,
as well as with non-amine additives. Along
with these changes, our ability to simulate and
predict the performance of amine treating
processes has increased enormously in recent
years.
Commercial software packages for amine plant
simulation use column models that range from
simple equilibrium stages, equilibrium stages
modifed for reaction kinetics, equilibrium stages
with computed stage effciencies, right through
to true mass and heat transfer rate models.
However, regardless of the underlying principles
on which each one is based, simulations of
amine plants have traditionally assumed the
solvent to be perfectly clean, meaning that it
Ralph H Weiland and Nathan A Hatcher Optimized Gas Treating, Inc.
contains only water, amines, and acid gases. In
some cases, the solubility of light hydrocarbons
and inert gases may be taken into account. Until
recently, however, the process effect of organic
and inorganic acids that enter amine solutions
when tail gas from Claus sulphur plants is
treated, or which are almost universally encoun-
tered in oil refning operations, has been
ignored. These components are called heat stable
salts (HSS) or heat stable amine salts (HSAS). A
few are acids, both organic and inorganic, that
occur in treating solvents at a total concentra-
tion that does not typically exceed ten to ffteen
thousand parts per million by weight, and is
often very much lower than this. Apart from
being blamed for corrosion problems, HSSs can
have a profound effect on process performance.
Ignoring them or misunderstanding their chem-
istry can lead to very bad processing decisions,
to recommendations to make expensive equip-
ment changes that, in fact, make the process
worse than ever, and to continuing to surround a
number of modern proprietary solvents with an
undeserved mystique. Indeed, HSSs are some-
times purposely added to an amine solvent to
boost the level of treating performance well
beyond that attainable in a clean system. They
can be both friend and foe how to view them
and deal with them in a given circumstance
depends on understanding the chemistry, vapour
liquid equilibrium constraints, and the mass
transfer rate processes taking place in absorp-
tion and regeneration columns. Outside the
laboratory, clean solvents exist in many plants
for only a short time immediately following
initial system charging and startup. In real
processes, solutions are always contaminated to
some degree. The work reported here attempts
to provide a scientifc understanding of HSS
chemistry and the effect of these components on
the process.
www.digitalrening.com/article/1000953 June 2014 1
Heat Stable Salt Process Chemistry
Over time, solvents accumulate contaminants
primarily from the gases being treated or
through the use of makeup agents (water and
amine) that are not completely pure.
Contaminants of interest here are frequently the
anions of organic and inorganic acids, also called
heat stable salts (HSSs). Anions commonly
found in amine solutions include thiosulphate,
oxalate, sulphite, sulphate, glycolate, propionate,
acetate, thiocyanate, formate, and chloride,
which usually enter the solution as a result of
absorption from the gases or liquids being
treated. Contaminants may also be cationic such
as alkali metal ions (sodium, potassium, calcium,
and magnesium) that accumulate from makeup
water hardness or through deliberate addition in
the form of hydroxides or carbonates to deproto-
nate amine associated with HSAS anions. As
contaminants, all of these ions can have a
profound effect, sometimes positive, frequently
negative, on amine treating unit performance.
Other HSS anions such as sulphate or phosphate
are purposely added to enhance H
2
S removal in
tail gas treating units (TGTUs), a practice that,
with greater understanding of the basic science,
is becoming increasingly common.
Sources of Heat Stable Salts
There are three major sources of acidic and basic
contaminants in most amine systems: cracking
unit operations (refnery FCCs and Cokers for
example), oxidation products from Claus sulphur
plants, and certain amine reclaiming practices
(see Hatcher et al., 2006).
After a period of use, especially in treating
sour gases generated from refnery cracking
operations (Cokers, FCCs), trace amounts of acid
anion contaminants can build to signifcant
levels in the solvent. The most commonly found
acid anions are formate and thiocyanate, which
result from the absorption of hydrogen cyanide.
Formate is formed by the hydrolysis of the
cyanide ion to ammonium formate, while thiocy-
anate forms from dissolved oxygen reacting with
H
2
S, followed by reaction of the oxysulphur
anion with cyanide ion. Higher molecular weight
organic acid anions come from the hydrolysis of
higher molecular weight nitrile compounds.
Ammonium ion produced by the hydrolysis reac-
tion will yield an H
+
to the amine; the resulting
ammonia molecule is then stripped from solu-
2 June 2014 www.digitalrening.com/article/1000953
tion by steam in the regenerator and it
accumulates in the overhead condensing system.
This leaves the protonated amine/HSS anion
pair in the amine solution. The reactions are:
RCN + 2 H
2
O NH
4
+
+ RCOO
-
(R=H, alkyl group) (1a)
NH
4
+
+ R
1
R
2
R
3
N R
1
R
2
R
3
NH
+
+ NH
3
(1b)
2 HCN + O
2
+ 2 H
2
S + 2 R
1
R
2
R
3
N 2 R
1
R
2
R
3
NH
+
+ 2 SCN
-
+ 2 H
2
O (1c)
Hydrogen from gasoline reformers can also
contain HCl which will react directly via acid-
base neutralisation with the amine. Thus, for a
strong acid HnX where X is an n-valent anion
(Cl , SO4=, etc.) the reaction with amine is
H
n
X + n R
1
R
2
R
3
N n R
1
R
2
R
3
NH
+
+ X
n
(2)
Thiosulphates generally result from the reac-
tion of dissolved oxygen with H
2
S or from SO
2
reaction with H
2
S in Claus tail gas units when no
HCN is present. Sulphates can either be formed
from absorption of sulphuric acid or from
further oxidation of thiosulphates. The reactions
are
2 H
2
S + 2 O
2
+ 2 R
1
R
2
R
3
N 2 R
1
R
2
R
3
NH
+
+ S
2
O
3
=
+ H
2
O (3a)
2 H
2
S + 4 SO
2
+ H
2
O + 6 R
1
R
2
R
3
N 6 R
1
R
2
R
3
NH
+
+ 3 S
2
O
3
=
(3b)
S
2
O
3
=
+ 5/2 O
2
2 SO
4
=
(3c)
The third source of contaminants comes from
deliberate or inadvertent introduction through
solvent reclaiming practices. With increasing
time-on-stream, acidic contaminants continue to
build until a point is reached where something
must be done to recover the permanently
neutralised, bound amine to restore free amine
circulation capacity or to prevent the amine unit
from self dissolving through corrosion. Some
operators employ addition of a stronger base
(typically NaOH or KOH) to neutralise the
amine heat stable salt, that is, remove the proton
from the protonated amine and attach the
anion to the alkali metal ion. This forms the
sodium or potassium salt of the HSS, i.e., NaX
or KX, and releases R
1
R
2
R
3
NH+ or R
1
R
2
NH
2
+
back into the free amine forms R
1
R
2
R
3
N or
R
1
R
2
NH. Using NaOH as an example,
NaOH + R1R
2
R
3
NHX R1R
2
R
3
N + NaX + H
2
O (4a)
2 June 2014 www.digitalrening.com/article/1000953
However, unless carried out very precisely,
neutralisation with strong bases can have disas-
trous, unforeseen consequences. The danger lies
in the possibility of over neutralisation. If the
solvent becomes over neutralised, the excess
caustic permanently binds H
2
S and CO
2
:
NaOH + H
2
S Na+ HS

+ H
2
O (4b)
NaOH + CO
2
Na+HCO
3

(4c)
The resulting elevated concentrations of HS
and HCO
3

remain chemically bound in solution

no matter how hard the solvent is stripped in the
regenerator, i.e. there is a permanent high lean
loading. In some cases, an order-of-magnitude
increase in lean loading has been observed. As a
result, the regenerated solvent has a much high-
er-than-expected residual acid gas loading which
cannot be reduced by the application of further
heat, and it may become impossible to meet
treating requirements.
Effect of HSSs on Chemical Reaction Equilibria
Clean solvents consist of water solutions of
one or more amines together with the acid gases
CO
2
and H
2
S. A number of reaction equilibria
are set up, all involving ionic species, and all
depending on the presence of water. For a
system containing a single amine together with
H
2
S and CO
2
in aqueous solution, six equilib-
rium ionic reactions occur:

H
2
O H
+
+ OH

(5)

R
1
R
2
R
3
N + H
2
O R
1
R
2
R
3
NH
+
+ OH

(6)
H
2
S + R
1
R
2
R
3
N R
1
R
2
R
3
NH
+
+ HS

(7)
HS

+ R
1
R
2
R
3
N R
1
R
2
R
3
NH
+
+ S
=
(8)
CO
2
+ R
1
R
2
R
3
N + H
2
O R
1
R
2
R
3
NH
+
+ HCO
3

(9)
HCO
3

+ R
1
R
2
R
3
N R
1
R
2
R
3
NH
+
+CO
3
=
(10)
If the amine is primary (R
2
, R
3
= H) or second-
ary (R
3
= H), it is also capable of reaction with
CO
2
to form the carbamate of the amine, with
the overall reaction:
CO
2
+ 2 R
1
R
2
NH R
1
R
2
NCOO

+ R
1
R
2
NH
2
+
(11)
Thus, the four molecular species H
2
O,
www.digitalrening.com/article/1000953 June 2014 3
R1R2R
3
N, CO
2
and H
2
S dissociate and react with
each other to form seven ions (eight when the
amine is primary or secondary). For this
perfectly clean amine solvent, reaction equilibria
(5) through (10) are shifted towards reactants
(i.e., to the left) by increasing the temperature.
This thermal reversibility of the reactions is what
makes treating with amines economically feasi-
ble. But things can go awry when the gas
contains certain types of components, even in
trace amounts. This is especially the case when
the amine is tertiary (typically methyldiethanola-
mine, MDEA) and selective treating is being
done to remove H
2
S from the gas but to leave as
much CO
2
behind as possible.
H
2
S absorption takes place primarily via reac-
tion (7), with reaction (8) being of little
consequence in the operating pH range of most
alkanolamine solutions. With tertiary amines,
CO
2
absorption relies mostly on reaction (9).
With primary and secondary amines at normal
loadings, however, reaction (11) dominates. The
important observation is that each reaction even
in uncontaminated solutions has protonated
amine, R1R
2
R
3
NH+, on its right hand side:
H
2
S + R
1
R
2
R
3
N R
1
R
2
R
3
NH
+
+ HS

(7)
CO
2
+ R
1
R
2
R
3
N + H
2
O R
1
R
2
R
3
NH
+
+ HCO
3

(9)
CO
2
+ 2 R
1
R
2
NH R
1
R
2
NH
2
+
+ R
1
R
2
NCOO

(11)
For a lean tertiary amine at regenerator
temperatures (i.e., towards the bottom of the
column and in the reboiler) the acid gas loadings
are intended to be quite low so the concentra-
tions of HS

, HCO
3

and R
1
R
2
NCOO

are small.
In a clean solvent, the concentration of
R
1
R
2
R3NH
+
or R
1
R
2
NH
2
+
will be quite small, too.
However, when the solvent becomes increasingly
contaminated with acid anions that are stronger
than HS
-
and HCO
3

, the concentration of proto-

nated amine (R
1
R
2
R
3
NH
+
or R
1
R
2
NH
2
+
) at the
regenerators lean end becomes greatly infu-
enced by the extent of contamination, as set by
reactions (1), (2) and (3). Unlike the acid
gas-amine reactions (511), reactions (1), (2) and
(3) are not thermally reversible (thus the term
heat stable salt), so the HSSs permanently tie up
part of the amine as R
1
R
2
R
3
NH
+
ion. The amine
is gradually converted to HSAS and becomes
inactivated. Of potentially greater concern than a
small loss of amine strength, is its effect on the
ability to regenerate the rich solvent to satisfac-
tory acid gas lean loadings (moles of acid gases
per mole of total amine) and to use the regener-
ated solvent effectively in the absorber .
The higher protonated amine concentration
drives reactions (7), (9) and to a small extent
(11) more strongly to the left which is benefcial
to regeneration because it produces greater driv-
ing force for the reverse or decomposition
reaction. Thus, all other conditions being the
same, from a mass transfer and phase equilib-
rium viewpoint, solvent contaminated with
relatively strong-acid is always easier to regener-
ate than a clean one simply because the chemical
equilibria are favourably shifted. From the
chemists viewpoint, there is a common ion
effect at work, in which seemingly small
amounts of contamination drive the main
decomposition reactions further towards
completion and result in a solvent that is
stripped to far lower residual concentrations of
the acid gases CO
2
and H
2
S than could ever be
achieved through using large numbers of contact
stages or very high energy inputs.
There is a catch, however. The common ion
R
1
R
2
R
3
NH+ also displaces the reactions (7, 9,
and 11) towards the left under absorption condi-
tions. Again, all other conditions being the same
(same temperature, total amine concentration,
and acid gas loadings), a contaminated lean
solvent entering an absorber will always exert
greater acid gas partial pressures over it than a
clean one. Taken by itself, this means that acid
contamination is always detrimental to absorp-
tion. The outcome for the overall process,
however, depends very much on the exact condi-
tions of operation as to whether the beneft
provided to regeneration by the contaminant
outweighs its penalty to absorption. An assess-
ment of the potential beneft or penalty of HSSs
to the process requires one to evaluate the entire
plant, not just an isolated contactor or regenera-
tor. To do this assessment rigorously, a mass
transfer rate-based amine simulation tool is
required, and the simulator must be capable of
accounting correctly for ionic contaminants.
Turning now to sodium contamination, the
equilibrium of reactions (3), (6), and (7) is
driven to the left at conditions in the absorber
despite the fact that HS

and HCO
3

appear to be
bound with sodium. In fact, they are in the free,
dissociated state and cause the solution to have
potentially high concentration of both ions. The
net result can be failure to meet treated gas
specifcations by a wide margin which cannot be
overcome using more reboiler energy or
increased solvent circulation rate because the
bisulphide and bicarbonate ions are tightly tied
up with and stabilised by the sodium. Thus, the
excess sodium remains associated with HS

and
HCO
3

and the situation must remain that way

until HSSs again build up to a level suffcient to
displace these ions fully and consume all the
excess caustic.
To understand and quantitatively simulate the
effect of HSSs and alkali metal ions on treating
performance, one must be able to model the
regenerator with just as much confdence and
accuracy as the absorber the regenerator sets
lean solution quality, the main controlling factor
at the lean end of the absorber where treat itself
is determined. It is equally crucial to account for
ionic contaminants and to use a comprehensive
solvent analysis in any accurate process assess-
ment. Before looking at specifc examples of HSS
contamination, it is worth pointing out the heat
stable salts are not always contaminants
sometimes they are added on purpose.
Dibble (1985), in a now expired U.S. patent,
proposed the use of certain mineral acids to
promote the stripping of acid gases from amine
solutions. Included was phosphoric acid which is
used as an additive to MDEA in some proprie-
tary solvents for deeper removal of H
2
S,
especially from Claus sulphur plant tail gas.
Because of the low total gas pressure, the
normally high CO
2
to H
2
S ratios, and the desire
to remove H
2
S to levels of a few parts per
million, this is a diffcult application for amine
treating. Phosphate ions and sulphate ions as
either the acids or their ammonium salts act in
the very same way as the HSSs already
discussed. If added as an ammonium salt, the
proton of the ammonium ion transfers to the
amine, and the released ammonia, being volatile,
is removed from the amine unit via overhead
condenser refux water blow down. The proto-
nated amine acts to drive reactions (7) and (9)
strongly to the left at low acid gas loadings just
as it does when HSSs are present accidentally as
contaminants.
Validating Solution Analyses
4 June 2014 www.digitalrening.com/article/1000953
In many cases, it has been our experience that
amine treating solutions are rigorously analysed
only when the treating unit itself has diffculty
meeting treating specifcations, exhibits severe
foaming, or begins to experience serious corro-
sion. In the present context, the concern is
solution degradation, typifed by the need, for
example, to determine the appropriateness and
likely outcome of solution reclaiming, i.e., of
removing the HSSs either by distillation or ion
exchange, or negating their effect through caus-
tic neutralisation. Accurate simulation is an
outstanding tool for answering reclaiming and
contamination questions in terms of process
performance, but unless the simulation is based
on accurate solution analysis, its results can be
quite misleading. The accuracy of such an analy-
sis can be validated by checking whether it
corresponds to an electrically neutral solution.
Since electro neutrality involves all the ions,
including bisulphide, it can be used to assess the
accuracy of lean solution acid gas loadings. The
following example is drawn from a refnery
treating operation in which the amine solution
was found contaminated not just with HSSs but
also with caustic soda that had accidentally
entered the amine circuit.
An actual laboratory analysis is shown in the
frst two columns of Table 1. The question is
whether the reported data are self consistent.
Ions were measured directly by ion chromatog-
raphy so one has no option but to assume that
sodium and the HSS anions are correctly
reported. However, H
2
S is deter-
mined via potentiometric titration
with silver nitrate or via starch
iodine titration. Since H
2
S is a vola-
tile component in the sample, it
cannot be taken as a given that the
sample being analysed is unchanged
from the time of sampling. It does
not seem to be generally appreci-
ated, for example, that great care
must be taken in sampling, sample
transportation to the laboratory,
and the analysis itself if reliable H
2
S
content in liquid amine samples is
to be assured. Unless every precau-
tion is taken to exclude the samples
from contact with air during each
step along the way from sampling to
fnal analysis, as much as 80 to 90%
of the lean H
2
S loading can disappear, only to
reappear as thiosulphate. When common prac-
tice is to sample directly into an empty
(air-flled) bottle, leave an air space for expan-
sion at the top, then ship the sample (often with
vigorous agitation) for several days and analyse
it in an air environment, the validity of most, if
not all, lean amine H
2
S loading analyses must be
viewed with considerable scepticism.
To analyse the data of Table 1, the raw analysis
in terms of weight percents is converted into
mole fractions of each ionic species, then into
mole fractions of actual charge. The total charge
is forced to zero by adjusting the charge concen-
tration of HS

ion shown in the last column to

the value 0.00407 mole fraction (the sign is
associated with the charge). This indicates that
for electro neutrality the H
2
S mole fraction must
be 0.00407 and the loading must actually be
0.0326 mol/mol (0.00407/0.12476) as opposed
to the range 0.005 to 0.006 mol/mol reported
by the laboratory. An Excel spreadsheet for
doing this type of calculation is freely available
(Optimized Gas Treating, Inc., 2008).
Renery Fuel Gas Treater with HSS
Contaminated Solvent
Refnery amine systems are notorious for solvent
contamination by a variety of heat stable salts
and amine degradation products. If HSSs are not
properly considered during the design phase,
and if they are not taken into account when
troubleshooting and analysing column and plant
www.digitalrening.com/article/1000953 June 2014 5 4 June 2014 www.digitalrening.com/article/1000953
A Typical Solution Analysis
Amine wt% Weight Mole Mole
Fraction Fraction Fraction
Species Species Charge
Free 47.4 0.47400 0.12346
Bound (MDEAH
+
) 0.5 0.00500 0.00130 0.00130
Total 47.9 0.47900 0.12476
Ions ppmw
Sodium (Na
+
) 7717 0.00772 0.01042 0.01042
Formate (HCOO

) 3812 0.00381 0.00263 0.00263

Acetate (CH
3
COO

) 3322 0.00332 0.00175 0.00175

Chloride (Cl

) 27 0.00003 0.00002 0.00002

Sulfate (SO
4
=
) 160 0.00016 0.00005 0.00010
Thiosulphate (S
2
O
3
=
) 5658 0.00566 0.00157 0.00313
Thiocyanate (SCN

) 36 0.00004 0.00002 0.00002

Other Components
Acid gas as HS

0.0050.006 loading 0.00433 0.00407 0.00407

Water 0.49593 0.85471
Total 1.00000 1.00000 0.00000
Table 1
performance, quite erroneous conclusions can be
drawn and expensive changes to plant equip-
ment will be made with absolutely no beneft.
The following example demonstrates the conse-
quences of ignoring the effect of HSSs on amine
plant performance. After a solvent change from
MEA to generic MDEA, a packed column that
had been sweetening a 0.5% H
2
S fuel gas failed
to achieve satisfactory treat. The solvent vendor
expected the column to be capable of reaching 2
ppmv H
2
S in the treated gas but H
2
S leaks in the
range from 20 to 30 ppmv were being measured
regularly, with excursions as high as 80 ppmv.
A conventional commercial simulation tool
predicted that a treated gas containing 1 to 2
ppmv H
2
S should have been produced, in agree-
ment with the solvent vendors expectations. The
simulation basis was a clean solvent and simula-
tions used measured acid gas loadings in the
lean solvent. The presence of HSSs was not
considered, however, nor could this factor be
accounted for by the simulators that were used.
Thinking that perhaps the residence time of
liquid on the packing was too short to allow
suffcient H
2
S absorption (despite the fact that
H
2
S reacts essentially instantaneously upon
absorption into alkaline solutions and that only
CO
2
absorption with its slow reaction could
possibly be affected by contact time), a change of
tower internals to trays was recommended and
the tower was revamped accordingly. The results
were extremely disappointing performance was
no better than with packing.
If a complete change in the type of tower inter-
nals resulted in no signifcant change in overall
performance, obviously the problem was not one
of mass transfer. The answer had to lie with the
solvent. A detailed solvent analysis by the vendor
showed the solvent contained fve different
HSSs, with the total salt level exceeding 2 wt%
on a total solution basis. The packed and trayed
versions of the contactor were simulated using
the ProTreat simulator, with its mass and heat
transfer rate-based column model, and its ability
to account for the contaminated solvent analysis.
Using the measured solution lean loadings but
without HSSs the simulations indicated 0.6
ppmv H
2
S and 0.7 ppmv H
2
S for packing and
trays, respectively. However, the corresponding
simulations with HSSs gave 25 ppmv and 30
ppmv, respectively. A performance test on the
trayed column itself showed a treating level of
26 ppmv H
2
S, in very close agreement with
simulation. The increased equilibrium acid gas
partial pressures over the contaminated solution
are completely in line with expectations from
process chemistry and the common ion effect.
Failure to perform was the result of shifted
phase equilibrium caused by HSSs it had
nothing to do with the column internals.
However, it must be pointed out that for a real-
istic analysis of the effect of HSSs on the
performance of this plant it is absolutely neces-
sary to include the regenerator in the
simulation.
Suffce it to say that if proper account of the
HSS effects had been taken in the frst place, a
costly and ineffective column internals change
out would have been avoided, and action would
have focused instead on providing HSS removal
from the solvent. But to take the right steps, it is
necessary frst to understand the chemistry of
the process. In the present case, the HSSs in the
solvent were undesirable and their presence
meant the difference between meeting sulphur
emissions standards and missing them by a wide
margin. The ability to predict their effect
through truly rigorous simulation helps to avoid
design errors, indicates when reclaiming equip-
ment should be included in the design, and helps
to make periodic reclaiming decisions as to
when, and how vigorously, to reclaim. As shown
in the next example, however, reclaiming can
sometimes be done too aggressively, and it may
result in the plants failure to meet treating spec-
ifcations. In this case, the HSSs are actually
benefcial to treating and removing them too
vigorously will signifcantly reduce treating
ability.
Effect of Reclaiming on a Tail Gas Treating Unit
The refnery tail gas treating unit (TGTU) shown
in Figure 1 is a conventional scheme using 34
wt% MDEA to treat tail gas from a Claus sulphur
recovery unit (SRU). The tail gas is 1.7% H
2
S and
3.4% CO
2
and the contactor has 20-ft of
FLEXIPAK 2Y structured packing to minimise
pressure drop and maximise tower capacity.
TGTUs are typically run on a separate solvent
circuit; however, this one was being run as part
of the refnery MDEA system. An analysis
performed by the solvent vendor showed several
heat stable salts (HSSs) present at the concen-
trations shown in the fgure, with total HSSs at
6 June 2014 www.digitalrening.com/article/1000953
6 June 2014 www.digitalrening.com/article/1000953
0.8115 wt%. Perhaps surprisingly, the unit
nevertheless was producing a vent gas with only
3 ppmv H
2
S, a very low concentration for a
TGTU where 100 ppmv is much more the norm.
The solvent is obviously quite contaminated and
plant operations personnel were considering
reclaiming. The question was asked: If we
reclaim the solvent by removing HSSs, what will
be the impact, if any, on treating performance?
The right place to start answering this kind of
question is a good, reliable simulation capable of
modelling the real equipment and the real,
contaminated solvent.
ProTreat was used to model the complete
plant, including the regenerator with a known
tray count, the detailed solution analysis (HSS
profle), and reboiler duty. The predicted treat of
2.3 ppmv H
2
S was in good agreement with the
measured level of 3 ppmv H
2
S.
If a model that did not account for HSSs had
been used, the predicted treat would have been
over 44 ppmv, more than 10 times the observed
value. This change is in the completely opposite
direction to the frst example where HSSs caused
a serious loss in performance. It should be
noted, however, that unlike the frst example
where the lean solvent acid gas loadings were
taken as the measured values, here the entire
amine circuit, including the regenerator, is being
analysed so the lean solution loadings are also
being predicted. Table 2 shows the simulated
effect of various levels of HSS removal (degrees
of reclaiming) on lean solution quality and H
2
S
treat. This suggests that a serious additional
contribution to the refnerys allowable sulphur
emissions might result from reclaiming, even if
done only moderately. Because of this, and the
facts that treating performance was quite good,
the solutions treating capacity (amine strength)
was also high, and corrosion was not an issue,
the decision was made not to reclaim.
CO
2
slip is hardly affected by reclaiming, but
reclaiming has a tremendous effect on the units
H
2
S leak. Notice also that when the solvent
contains its full compliment of HSSs, lean load-
ings are reduced by a factor of 10 for CO
2
and a
factor of 100 for H
2
S compared with the virgin
solvent. Apart from benchmarking the reliability
of the simulator, a clean solvent, in a TGTU at
least, may not treat to nearly as low a residual
H
2
S level as a contaminated solvent. Of course,
this must be balanced in a practical sense with
the corrosive effects of excessive levels of HSSs.
Perhaps it would be better in this case to avoid
HSS-induced corrosion by employing a dedi-
www.digitalrening.com/article/1000953 June 2014 7

Figure 1 Tail Gas Treating Unit
Effect of HSS Removal on TGTU Performance
% HSS Lean CO
2
Load Lean H
2
S Load H
2
S Leak CO
2
Slip
Removed (mol/mol) (mol/mol) (ppmv) (%)
0 0.000205 0.000071 2.3 66.05
25 0.000320 0.000251 6.5 65.79
50 0.000544 0.000934 17.5 65.55
75 0.001018 0.002790 34.5 65.33
100 0.002044 0.006177 44.3 65.17
Table 2
cated amine circuit and using a few 1000s ppmw
phosphate ion to improve solvent regeneration.
The processing effect of HSSs is really felt in the
regenerator where much lower loadings can be
achieved when the solvent contains HSSs. With
an understanding of the effect of process chem-
istry on reaction equilibria, perhaps this is not
so surprising.
At high acid gas loadings the impact of a small
amount of additional protonation is completely
negligible because the protonated amine concen-
tration is already very high (small change to a
high concentration). But in the reboiler, for
example, the H
2
S loading will be very small (if
low H
2
S leak is to be achieved from the TGTU),
so even a small amount of additional protona-
tion is highly signifcant relative to the very low
concentration of protonated amine normally
present. In fact, the additional protonation can
be 10 to 100 times higher than what would
normally be found in a well-regenerated virgin
MDEA solution. The additional amine protona-
tion caused by HSSs and purposefully added
acids displaces reaction equilibria strongly
towards the formation of free, molecular H
2
S
capable of desorbing from the solution.
Additional (perhaps even artifcially introduced)
protonation enhances solvent regeneration.
The higher H
2
S back-pressure caused by the
increased protonation arising from HSSs,
however, affects absorption negatively. But, it
turns out that in this particular case its benef-
cial effect on reducing the lean loading in the
regenerator far outweighs its negative effect on
back-pressures in the absorber. The result can
easily be a factor of 10 or 20 lower H
2
S leak
when HSSs are present in small amounts. Of
course, caution is needed not to let HSSs build
up too high because they are corrosive. But
caution is also needed not to reclaim too aggres-
sively or treat may be lost. Even if the corrosion
resulting from high HSS concentrations is unac-
ceptable and they must be removed, their
benefcial effect can still be had by replacing
them with a small amount of phosphoric acid.
Simulation will reveal how much phosphoric
acid is enough, how much to reclaim (from a
process standpoint) and what the outcome on
treating unit process performance will be.
The outlet CO
2
concentration is controlled by
the extreme slowness of the reaction between
CO
2
and the amine (actually the low OH ion
concentration in the solution), and not by lean
loading. Thus, the effect on CO
2
slip of even a
factor of 10 reduction in lean solvent CO
2
load is
negligible.
One factor that does not generally seem to be
appreciated, is that using stripping promoters
such as phosphoric acid to reduce H
2
S leak is
effective only if the contactor is lean-end pinched
with respect to H
2
S. Whether a given column is
or is not lean end pinched is unmistakably
shown by mass transfer rate based simulation of
the column. If the column is lean end pinched,
the concentration of H
2
S will remain small and
constant over several metres of packing or
several trays at the top (lean end) of the column.
Reclaiming decisions should start with a good
set of simulations generated using a process
simulator that has high accuracy and reliability.
Mass and heat transfer rate based simulations
meet this criterion. The simulator must also be
able to model the actual system under study,
especially the detailed solution chemistry and
the mass transfer behaviour of the real column
internals being used. Accurate regenerator
modelling is just as important as simulating the
absorber simply because the regenerator sets the
acid gas lean loadings and regeneration is where
HSSs and stripping promoters have their real
effect.
Conclusions and Summary
Troubleshooting amine treating unit problems
can often involve much more than simply plug-
ging plant data into an amine plant simulation
model. In the frst place, amine treating chemis-
try is a rather complex system of ionic reactions,
some thermally reversible, others not, with
common ion effects and great potential for
inconsistencies in plant data. In addition, treat-
ing is often done selectively and this requires the
treating operation to be analysed as a mass-
transfer-rate processsolely basing it on phase
equilibrium is completely inadequate to under-
standing the process and the problem. What is
more, all too often an attempt is made to analyse
only the operation of the absorber without
recognising that its performance hinges heavily
on the quality of the lean amine being produced
by the regenerator. Thus, a frst rate regenerator
model is just as important as a good absorber
simulation.
Careful analysis of plant and laboratory meas-
8 June 2014 www.digitalrening.com/article/1000953
urements can be complemented very effectively
by a simulation model capable of accounting for
all the details and applied to the solution of
treating plant operating problems. In both
examples, it is evident that the complete solution
ion chemistry must be taken into account to
model treating accurately down to low H
2
S levels
at low lean loadings. Any solvent contaminants
must be included in simulations if these are to
predict successfully and accurately the operation
of amine plants.
Understanding the process chemistry is an
essential part of analysing and troubleshooting
amine plants, and using an amine simulator
capable of accounting for the effects of all the
factors, including detailed solution chemistry,
formed an integral part of the analyses. In the
fnal analysis, simulations that do not account
correctly for all the ionic species may give erro-
neous results that can lead to unnecessary and
expensive changes to plant equipment and
solvents. Because the cases studied involved
selective treating, the fact that the simulator was
based on sound mass-transfer-rate principles
enabled the modelling to compliment the analy-
ses without such a true mass-transfer-rate
basis, the simulations would have been of little
or no help. Furthermore, accurately simulating
regeneration was critical to the successful simu-
lation and analysis.
Literature Cited
1 Bottoms, R. R., U.S. Patent 1,783,901 (1930).
2 Dibble, J. H., European Patent Publication Number 0134948,
Application Number 8417586.4, March 27, 1985.
3 Hatcher, N. A., Keller, A. E., Weiland, R. H., Sivasubramanian,
M. S., Proceedings of the 56th Laurence Reid Gas Conditioning
Conference, Norman, Oklahoma, February, 2006.
www.digitalrening.com/article/1000953 June 2014 9 8 June 2014 www.digitalrening.com/article/1000953

LINKS
More articles from: Optimized Gas Treating
More articles from the following categories:
Gas Processing/Treatment
Mass Transfer & Separation
Process Modelling & Simulation