IMPACT OF THERMOPHYSICAL PROPERTIES RESEARCH

ON ACID GAS INJECTION PROCESS DESIGN

Heng-Joo Ng
DB Robinson Research Ltd.
9419 20 Avenue
Edmonton, Alberta, Canada T6N 1E5

John J. Carroll and James Maddocks

Gas Liquids Engineering, Ltd.
#300, 2749 39 Avenue NE
Calgary, Alberta, Canada T1Y 4T8

Abstract
Stricter regulations and increased environmental concerns are making the previously
common practice of flaring acid gas less attractive. Producers are being forced to find
alternative methods to deal with this unwanted by-product. Acid gas injection has
quickly become the method of choice for disposal, especially for small producers.
Presented in this paper is a brief overview of the acid gas injection process. Included in
this discussion is a description of what physical property and phase equilibria data are
required and why.
In addition, some new experimental data for the water content and density of sour gas
mixtures are presented. These data are useful for checking the validity of the
correlations used in the design of an acid gas injection scheme. In addition, such data
are useful for optimizing the interaction parameters in the thermodynamic models
employed to calculate such properties.
1.0

ACID GAS INJECTION

Acid gas injection occurs in three basic steps: (1) compression, (2) transportation by
pipeline, and (3) injection. The acid gas injection process has been reviewed previously
(Wichert and Royan, 1997; Carroll and Maddocks, 1999) so it will only be discussed
here briefly.
Acid gas, the top product off the amine regenerator column, is at low pressure, typically
35 to 50 kPa[g] (5 to 7 psig). The acid gas is made up largely of hydrogen sulfide and
carbon dioxide, but it also contains small amounts of light hydrocarbons and possibly
nitrogen. Because of the nature of the stripping process, the acid gas is saturated with
water at these conditions.

In order to get it through the pipeline and into the injection well, the acid gas has to be
compressed. The compression takes place in a multi-stage compressor with interstage
cooling and liquids knock-out. The interstage exchangers are typically aerial coolers and
are designed to cool the acid gas to 49C (120F). At this temperature, an aqueous
phase is likely to form and is removed using interstage knock-out scrubbers.
Next the acid gas is transported to an injection well in a pipeline. Based on safety
considerations, the pipeline is generally kept as short as possible. On the other hand,
the injection well may be an existing well converted for injection purposes. Then the
length of the pipeline is dictated by the distance from the plant to the well of choice.
Regardless, when in operation, the line is filled with acid gas and a leak would release a
large volume of the deadly gas. By keeping the line short as possible, the amount of gas
released can be kept to a minimum.
The design of the pipeline follows the standard procedures for pipeline design.
However, depending upon the circumstances the line may carry a gas, a liquid, or a
mixture of both. Therefore, the procedure used to design the pipeline must be able to
handle these multiphase cases.
Finally, the fluid enters the injection well where it eventually enters a deep formation for
the ultimate disposal. The calculation of the injection profiles, and hence the injection
pressures, was a topic covered in previous papers: Carroll, and Lui (1997) and Carroll
and Maddocks (1999). However, it is worth repeating the accurate calculation of the
injection pressure is only possible with accurate fluid density estimates.
2.0

PHASE EQUILIBRIA AND PHYSICAL PROPERTIES

There are several important physical property and phase equilibria to be considered in
the design of an acid gas injection scheme. These include, but are not limited to: (1)
non-aqueous phase behavior, (2) water content of the acid gas and the possible
formation of an aqueous liquid, (3) hydrate formation, and (4) the density and viscosity
of the acid gas.
Perhaps the most significant problem with the calculations for acid gas injection is that
much of the process occurs at conditions that are near the critical point of the mixture.
Typically the range of temperature is 0.75 < TR < 1.25, where TR is the reduced
temperature, the ratio of the temperature to the critical temperature. The vicinity of the
critical point is a region where it is difficult to accurately calculate the physical
properties. Not only do the property estimations tend to be poor, but in addition, the
properties are strong functions of the temperature and the pressure. In this region, small
changes in pressure or temperature can have a dramatic effect on the physical
properties.
This means that properties are required for gas, liquid, and supercritical fluid states. In
addition, the calculations must make a smooth transition between the various states. As
demonstrated by Carroll and Lui (1997), not all models make this smooth transition.

2.1

Non-Aqueous Phase Behavior

Knowledge of the vapor-liquid equilibrium in the non-aqueous phases is important

throughout the design of the acid gas injection. For example, it is important in the design
of the compressor to know under what conditions the acid gas will liquefy. We do not
want liquids forming in the compressor, nor do we want non-aqueous liquids forming in
the interstage knock-outs.
Here we are basically concerned with the vapor-liquid equilibrium in the acid gas
mixture. We are also concerned to some extent with the effect of water on this
equilibrium. The traditional tools, notably the Soave (1972) and Peng-Robinson (1976)
equations of state or their more recent modifications, can be applied to these mixtures.
It is well known that even these commonly employed tools require at least a single
binary interaction parameter to obtain accurate results. The desired method of obtaining
these interaction parameters is by fitting experimental data.
2.2

Water Content

The water content of acid gas is also a significant consideration in the design of an acid
gas injection scheme. Sour water is highly corrosive and can cause other processing
problems.
Unlike sweet gas where the water content is a continually decreasing function of the
pressure (i.e., the higher the pressure the less water the sweet gas holds), acid gases
and their mixtures have a minimum water content. We can take advantage of this
minimum in the water holding-capacity and eliminate the need for a dehydration unit.
Only if the water-holding capacity of the acid is well established and predictable can we
exploit this property. For this reason, accurate experimental data are highly valuable.
The estimation of the water content of both sour and acid gases is more difficult than
those for sweet gas. The short-cut methods (such as the McKetta chart) are simple to
use and provide reasonably accurate results for sweet gas. The suggested correction
methods for sour gas make these methods more difficult to use and the results are
much less accurate. In general they cannot be employed for acid gas mixtures.
Ng et al. (1983, 1985, 1995) presented water content data for systems containing both
sour and acid gases in GPA-sponsored research program (Research Report RR-66,
RR-87 and RR-149).
Another significant problem in the calculation of the water content of acid gases is the
fact that they are relatively simple to liquefy (per the earlier discussion). The liquefied
acid gas can hold more water than in the gaseous state. Again, the design engineer
should attempt to take advantage of this property.

2.3

Hydrate Formation

Gas hydrates are ice-like solids that form at conditions where a solid probably would not
otherwise be expected to form. Hydrates are well-known in the natural gas processing
business and are notorious for plugging pipelines and process equipment. Therefore the
conditions where hydrates may form is an important consideration in process design. It
is equally important to have a strategy to deal with hydrates when they do form during
normal operation, as well as start-up, shut-down, and during process upsets.
Of the components commonly present in natural gas, none forms a hydrate more easily
than hydrogen sulfide. The hydrate of H2S forms at the lowest pressures and persists to
the highest temperatures. For a thorough review of the hydrate forming condition for
hydrogen sulfide see Carroll and Mather (1991). Carbon dioxide also readily forms a
hydrate. A brief review of the hydrate forming conditions for CO2 is given by Chun et al.
(1996). Ng et al. (1983, 1985, 1986, 1987) presented hydrate and hydrate inhibition
data for systems containing acid gas mixtures as part of GPA-sponsored research
program (Research Report RR-66, RR-87, RR-92 and RR-106).
Thorough reviews of gas hydrates are also available. Perhaps the most significant of
these is the book by Sloan (1998). He also summarizes the experimental work on gas
hydrates up to that time.
In acid gas injection, one of the methods for combating hydrate formation is the use of
methanol or other hydrate inhibitor inhibitor. The design engineer should carefully
consider the amount of methanol to use. Over injection results in wasted methanol, and
hence wasted money. Insufficient methanol use risks the formation of an equipmentplugging hydrate and potentially an unwanted shut-down to deal with the problem.
2.4

Density and Viscosity

The density and viscosity are also required for the design of an acid gas injection
scheme.
The design of the pipeline follows traditional methods for pipeline design. Therefore, in
order to calculate pressure drops in the lines the Reynolds number, which is a function
of the density and the viscosity, is required.
In addition, as demonstrated by Carroll and Lui (1997) and Carroll and Maddocks
(1999), the density is the key parameter in the calculation of the injection profiles for the
injection wells.
The traditional cubic equations of state (Soave, 1972; and Peng and Robinson, 1976)
are basically correlations of the molar volume (and hence the density) as functions of
the temperature, pressure and, through the mixing rules, the composition. However,
they are notoriously poor at predicting the density of the liquid phase. Many attempts
have been made over the years to improve the accuracy of these equations in this

regard. These modification rely on accurate experimental data to obtain the required
parameters. The design engineer must be aware of this limitation. Failure to do so will
result in improperly sized equipment and lines.
A further complicating factor is that typical viscosity correlations use the density as an
intrinsic parameter. Therefore accurate estimation of the density is important for
accurate prediction of the viscosity.
2.5

Other Properties

The design of an acid gas injection scheme requires other physical property information.
Properties such as: enthalpy, entropy, and heat capacity tend to play a less significant
role or at least a less visible role. However, they are required in the design of the
compressor and heat exchange equipment. These other thermodynamic properties are
usually obtained in the traditional manner (typically using an equation of state). This
also requires the ideal gas heat capacities, which, for the components of interest in acid
gas injection, are readily available.
In the design of heat transfer equipment and heat loss calculations, the thermal
conductivity of the process fluids is also required. However, the design engineer can
often get away with poor estimates of the thermal conductivity, since it usually does not
have a significant effect on the design.
3.0

PROCESS SIMULATION

One of the most important advances in the field of process engineering has been the
advent of the general-purpose process simulators. The power that these software
packages puts on the desktop of the modern engineer was unheard-of thirty years ago.
The downside of these tools is that they have made the process engineer a little lazy.
Users are often so overwhelmed by these tools that they blindly believe the results they
supply. The design engineer should be aware of the old adage garbage in, gospel out.
Always question the results obtained in these simulations.
These observations are no less true in the emerging application of acid gas injection.
The tools provided in the general purpose process simulators are of great help in the
design of an acid gas injection scheme. However, the phase equilibria and physical
properties encountered in acid gas systems require the design engineer to be more
wary of these results. In addition, as noted earlier, the design engineer should be aware
of the limitations of the models used when working in the near-critical region.
The models used in the simulation software require high accuracy experimental data to
obtain the required adjustable parameters. Without the data the models can be in error
and often significantly so. But the design engineer is usually oblivious to this limitation,
rarely questioning what models are being used, never mind whether or not the required
parameters have been optimized.

4.0

EXPERIMENTAL DETAILS

As outlined in this paper physical properties and phase equilibria are keys to the
economic design of an acid gas injection scheme. Presented in this paper are some
preliminary results. These should be very useful for testing and tuning the models
available to the design engineer.
4.1

Experimental Equipment

The experimental measurements were carried out in a variable-volume, visual phase

behavior cell as shown in Fig. 1.
The main body of the cell consisted of a Pyrex glass cylinder 20.3 cm (8.0 in) long with
an internal diameter of 3.2 cm (1.25 in) and a 1.26 cm (0.5 in) wall. The effective
working volume was about 110 cm (6.7 in). This cylinder was housed inside a sturdy
outer steel shell. Vertical opposing tempered flat glass plates mounted in this outer shell
permitted visual observation over the entire working range of the cell. The cell was
equipped with a floating piston so that the fluid being studied could be isolated from
mercury, the hydraulic fluid. The volume, and hence the pressure of the fluids under
investigation, were controlled by a positive displacement pump connected through
mercury to the cell contents. A second transparent fluid, connected through the same
hydraulic system and confined between the steel shell and the cylinder, was used to
balance pressures across the Pyrex cylinder.
The entire cell assembly was mounted on a rocking mechanism inside a temperature
controlled air oven. Temperatures were measured with a Resistance Temperature
Device (RTD). The RTD for the cell contents was inserted into the bottom of the cell and
protruded into the mercury. The signals were read out on a digital temperature indicator
and were reliable to within 0.2F (0.1C). Pressures were measured using a
calibrated Heise pressure gauge.
A specially designed header mounted on top of the cell provided for charging and
sampling. The volume of each phase in the cell was determined using a cathetometer
readable to the nearest 0.01 mm. A volume vs. height calibration for the cell, obtained
by observing the change in elevation of the piston in the cell corresponding to a known
change in volume, was used to transform the height measurements of the vapor-liquid
interface.
4.2

Experimental Procedures and Methods

The experimental acid gas mixtures were prepared gravimetrically from pure
components and stored at elevated pressures in the single-phase condition in stainless
steel pipettes. At the start of a run, the cell was connected to a mercury pump and
evacuated. An appropriate amount of water was then added to the cell, followed by a
suitable amount of the prepared acid gas mixture.

Following the charging process, the pressure and temperature of the cell was adjusted
to the required values and equilibration commenced. The vapor and aqueous liquid
phases were present at each condition.
The most difficult part of the study was the sampling and analysis of each of the two
phases. This was because of the very low concentration of water in the vapor phase,
and limited solubility of hydrocarbon in the aqueous phase. For the vapor, one sample
was taken for the CO2, H2S, and hydrocarbon component determination, and a large
second sample was taken for determining the water content in the vapor phase. In this
second large sample, a solvent with an internal standard was used to extract water from
the vapor phase. The resulting liquid solutions were injected directly into a gas
chromatograph to determine the water content.
For the aqueous liquid phase, a large sample was collected and flashed to atmospheric
pressure and the released gases were analyzed by chromatography.
5.0

EXPERIMENTAL ACCURACY

5.1

Pressure

Pressures were measured using a 10,000 psi (69 MPa) calibrated Heise digital gauge.
Pressures are thought to be known within 3 psi (20 kPa).
5.2

Temperature

The temperatures were measured using resistance temperature devices. They are
believed known to within 0.2F (0.1C).
5.3

Materials

The source and purity of the raw material used for the experimental work is shown in
the following table.
Material

Min. Purity

Supplier

Methane
Propane
Carbon Dioxide
Hydrogen Sulfide
Water

99.95
99.0
99.8
99.8*
Distilled

Matheson of Canada
Praxair, Edmonton
Praxair, Edmonton
Matheson of Canada
Serra Feinbiochemical GmbH & Co., N.Y.

* - This material was distilled once before using, improving the purity as received
from 99.5 to 99.8 mole percent.

5.4

Composition

An HP5890 chromatograph with a combination of TCD and FID detectors was used for
the compositional analyses. The concentration of major components determined by
chromatography is thought to be known to within 0.003 mole fraction. The minor
components such as hydrocarbon and sour components in the aqueous phase are
known to be within 10% of the reported value. The water concentration in the vapor
phase is thought to be known to within 25 ppm.
6.0

EXPERMENTAL RESULTS

Some preliminary results from an on-going, GPA-sponsored research program are

presented here. These new data are for multicomponent mixtures containing CO2, H2S,
methane, propane, and water. These include measurements of the vapor-liquid
equilibrium between the sour gas and an aqueous liquid and the densities of the two
phases.
Table 1 presents the data for a mixture containing 75% methane-propane (95/5 mole
ratio) and 25% H2S-CO2 (1/3 mole ratio) plus water for four pressures at 48.9C
(120F). The K-factors for these mixtures are given in Fig. 2. Table 2 presents the data
for similar mixtures at 93.3C (200F). The K-factors for the data in Table 2 are given
graphically in Fig. 3.
Even though the data presented in this work are not directly applicable to the acid gas
injection process (the mixtures studied contain more hydrocarbon than encountered in a
typical acid gas injection scheme), they provide a severe test of the available
thermodynamic models. And in that sense they are highly valuable.
7.0

DISCUSSION

Acid gas injection is quickly becoming the method of choice for dealing with unwanted
acid gas. Although the design of an acid gas injection scheme follows relatively
conventional procedures, the engineer is dealing with fluids for which there is a relative
paucity of experimental data. Therefore, accurate experimental data, such as those
presented in this paper are required to validate and tune the existing models.
Without accurate data, which lead to accurate models, the acid gas injection schemes
will be over-designed. The problem with such over-design is the increase in cost that
must result from such a practice.
8.0

REFERENCES

Carroll, J.J., Phase diagrams reveal acid-gas injection subtleties, Oil & Gas J., 96 (9),
92-96, (1998a).

Carroll, J.J., Acid-gas injection encounters diverse H2S, water phase changes, Oil &
Gas J., 96 (10), 57-59, (1998b).
Carroll, J.J. and D.W. Lui, Density, phase behavior keys to acid gas injection, Oil &
Gas J., 95 (25), 63-72, (1997).
Carroll, J.J. and J.R. Maddocks, Design considerations for acid gas injection,
Laurance Reid Gas Conditioning Conference, Norman, OK, Feb. 22-24, (1999).
Carroll, J.J. and A.E. Mather, Phase equilibria in the system water-hydrogen sulphide:
Hydrate-forming conditions, Can. J. Chem. Eng., 69, 1206-1212, (1991).
Chun, M.-K., J.-H. Yoon, and H. Lee, Clathrate phase equilibria for water + deuterium
oxide + carbon dioxide and water + deuterium oxide + chlorodifluoromethane (R22)
systems, J. Chem. Eng. Data, 41, 1114-1116, (1996).
Ng, H.-J. and Chen, C.-J., "Vapor-Liquid and Vapor-Liquid-Liquid Equilibria for H2S,
CO2, Selected Light Hydrocarbons and a Gas Condensate in Aqueous Methanol or
Ethylene Glycol Solutions", GPA Research Report RR-149, (1995).
Ng, H.-J., Chen, C.-J. and Robinson, D.B., "Hydrate Formation and Equilibrium Phase
Compositions in the Presence of Methanol: Selected Systems Containing Hydrogen
Sulfide, Carbon Dioxide, Ethane or Methane", GPA Research Report RR-87, (1985).
Ng, H.-J., Chen, C.-J. and Robinson, D.B., "The Effect of Ethylene Glycol or Methanol
on Hydrate Formation in Systems Containing Ethane, Propane, Carbon Dioxide,
Hydrogen Sulfide or a Typical Gas Condensate", GPA Research Report RR-92,
(1986).
Ng, H.-J., Chen, C.-J. and Robinson, D.B., "The Influence of High Concentrations of
Methanol on Hydrate Formation and the Distribution of Glycol in Liquid-Liquid
Mixtures", GPA Research Report RR-106, (1987).
Ng, H.-J. and Robinson, D.B., Equilibrium Phase Composition and Hydrating
Conditions in Systems Containing Methanol, Light Hydrocarbons, Carbon Dioxide
and Hydrogen Sulfide", GPA Research Report RR-66, (1983).
Ng, H.-J. and Robinson, D.B., "The Influence of Methanol on Hydrate Formation at Low
Temperatures", GPA Research Report RR-74, (1984).
Peng, D.-Y. and D.B. Robinson, A new two-constant equation of state, Ind. Eng.
Chem. Fundament., 15, 59-64, (1976).
Sloan, E.D., Clathrate Hydrates of Natural Gases, 2nd ed., Marcel Dekker, New York,
(1998).

Soave, G., Equilibrium constants from a modified Redlich-Kwong equation of state,

Chem. Eng. Sci., 27, 1197-1203, (1972).
Wichert, E., and T. Royan, Acid gas injection eliminates sulfur recovery expense, Oil &
Gas J., 95 (16), 63-72, (1997).

Table 1

Equilibrium Phase Properties for Systems Containing 75% Methane-Propane

(95/5 mole ratio) and 25% Hydrogen Sulfide-Carbon Dioxide (1/3 mole ratio)
in the Presence of Water at 48.9C (120F)

COMPONENT

FEED

CONCENTRATION, MOLE %
LIQUID 1
VAPOR

K1

P = 200 psia = 1.379 MPa

CH4
CO2
H2S
C3H8
H2O

1.8600
0.4901
0.1634
0.0984
97.3881

VOL %
Density(g/cm3)

0.0206
0.0803
0.0695
0.0008
99.8288

73.7208
16.3903
4.8633
3.9171
1.1086

28.5
0.987

71.5
0.013

3.58E+03
2.04E+02
6.99E+01
4.85E+03
1.11E-02

P = 1500 psia = 10.342 MPa

CH4
CO2
H2S
C3H8
H2O

7.6613
2.0196
0.6720
0.4029
89.2442

VOL %
Density(g/cm3)

0.1214
0.3832
0.3226
0.0024
99.1704

74.3182
16.4459
5.0497
3.9636
0.2225

42.9
0.995

57.1
0.108

6.12E+02
4.29E+01
1.57E+01
1.67E+03
2.24E-03

P = 4000 psia = 27.579 MPa

CH4
CO2
H2S
C3H8
H2O

18.2695
4.8153
1.6031
0.9614
74.3508

VOL %
Density(g/cm3)

0.2253
0.5901
0.3942
0.0030
98.7874

72.5156
17.5582
5.8814
3.8595
0.1852

39.9
1.002

60.1
0.287

3.22E+02
2.98E+01
1.49E+01
1.27E+03
1.88E-03

P = 10000 psia = 68.947 MPa

CH4
CO2
H2S
C3H8
H2O
VOL %
Density(g/cm3)

20.8731
5.6663
2.1092
1.1966
70.1547

0.3723
0.6886
0.3366
0.0019
98.6007

71.3409
17.9201
6.4730
4.1379
0.1281

44.9
1.017

55.1
0.431

1.92E+02
2.60E+01
1.92E+01
2.23E+03
1.30E-03

Table 2 Equilibrium Phase Properties for Systems Containing 75% Methane-Propane

(95/5 mole ratio) and 25% Hydrogen Sulfide-Carbon Dioxide (1/3 mole ratio)
in the Presence of Water at 93.3C (200F)

COMPONENT

FEED

CONCENTRATION, MOLE %
LIQUID 1
VAPOR

K1

P = 200 psia = 1.379 MPa

CH4
CO2
H2S
C3H8
H2O

1.0989
0.2896
0.0960
0.0578
98.4577

VOL %
Density(g/cm3)

0.0164
0.0427
0.0286
0.0006
99.9117

69.3268
16.3005
4.7116
3.7313
5.9298

34.5
0.963

65.5
0.011

4.24E+03
3.81E+02
1.65E+02
6.63E+03
5.94E-02

P = 1500 psia = 10.342 MPa

CH4
CO2
H2S
C3H8
H2O

9.5455
2.5188
0.8388
0.5025
86.5944

VOL %
Density(g/cm3)

0.1057
0.2593
0.1857
0.0023
99.4469

72.5179
17.7578
4.7363
3.8701
1.1179

32.4
0.960

67.6
0.089

6.86E+02
6.85E+01
2.55E+01
1.66E+03
1.12E-02

P = 4000 psia = 27.579 MPa

CH4
CO2
H2S
C3H8
H2O

15.3273
4.0455
1.3475
0.8062
78.4735

VOL %
Density(g/cm3)

0.2337
0.4453
0.3463
0.0036
98.9712

71.9676
17.4255
6.2095
3.7309
0.6666

40.7
0.973

59.3
0.227

3.08E+02
3.91E+01
1.79E+01
1.03E+03
6.74E-03

P = 10000 psia = 68.947 MPa

CH4
CO2
H2S
C3H8
H2O
VOL %
Density(g/cm3)

21.3049
6.3600
2.1732
1.2663
68.8956

0.3605
0.6073
0.3549
0.0024
98.6749

69.3699
19.5619
6.3461
4.1667
0.5553

41.7
0.973

58.3
0.396

1.92E+02
3.22E+01
1.79E+01
1.73E+03
5.63E-03

Figure 1

Schematic of Phase Behavior System

Figure 2

Equilibrium Ratio for Methane-Propane-Hydrogen SulfideCarbon Dioxide-Water System (75HC-25AC1) at 120F (48.9C)

T = 120 F (75HC-25AC1)
10000

1000

Equilibrium Ratio, K=Y/X

100

10

C1
CO2

H2S
C3
H2O

0.1

0.01

0.001
100

1000
Pressure, psia

10000

Figure 3

Equilibrium Ratio for Methane-Propane-Hydrogen SulfideCarbon Dioxide-Water System (75HC-25AC1) at 200F (93.3C)

T = 200 F (75HC-25AC1)
10000

Equilibrium Ratio, K=Y/X

1000

100
C1
10

CO2
H2S

C3
H2O

0.1

0.01

0.001
100

1000
Pressure, psia

10000