Effect of impurities on thermophysical properties and

phase behaviour of a CCS stream

Antonin Chapoy, Mahmoud Nazeri, Mahdi Kapateh, Rod Burgass, Christophe
Coquelet, Bahman Tohidi

To cite this version:

Antonin Chapoy, Mahmoud Nazeri, Mahdi Kapateh, Rod Burgass, Christophe Coquelet, et al.. Effect
of impurities on thermophysical properties and phase behaviour of a CCS stream. International
Journal of Greenhouse Gas Control, 2013, 19, pp.92-100. �10.1016/j.ijggc.2013.08.019�. �hal-01585883�

HAL Id: hal-01585883

https://hal-mines-paristech.archives-ouvertes.fr/hal-01585883
Submitted on 12 Sep 2017

HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est

archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents
entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non,
lished or not. The documents may come from émanant des établissements d’enseignement et de
teaching and research institutions in France or recherche français ou étrangers, des laboratoires
abroad, or from public or private research centers. publics ou privés.
 Effect of impurities on thermophysical properties and phase
behaviour of a CCS stream

Antonin Chapoya, Mahmoud Nazeria, Mahdi Kapateha, Rod Burgassa,

Christophe Coqueletb, Bahman Tohidia

a
Hydrates, Flow Assurance & Phase Equilibria Group, Institute of Petroleum Engineering,
Heriot-Watt University, Edinburgh EH14 4AS, Scotland, UK
b
MINES ParisTech, CTP-Centre Thermodynamique des Procédés, 35, Rue Saint Honoré,
77305 Fontainebleau, France

Abstract
CO2 obtained by capture process is not 100% pure and may contain impurities such as O2, Ar,
N2 and water. The presence of such impurities in CO2 stream can lead to challenging flow
assurance and processing issues. The aim of this communication is to present experimental
results on the phase behaviour and thermo-physical properties of carbon dioxide in the
presence of O2, Ar, N2 and water. The effect of these impurities on density and viscosity
were experimentally and theoretically investigated over the range of temperature from 243.15
K to 423.15 K up to 150 MPa. A corresponding-state viscosity model was developed to
predict the viscosity of the stream and a volume corrected equation of state approach was
used to calculate densities. Saturation pressures and the hydrate stability (in water saturated
and under-saturated conditions) of the CCS stream were also experimentally determined and
modelled. It is demonstrated that the thermodynamic models and approaches were able to
satisfactorily describe the thermophysical properties and phase behaviour of this CO2-rich
stream.

Keywords: Carbon Dioxide; CCS; Phase Behaviour; Density; Viscosity; Hydrate;

Thermodynamic Modelling


Corresponding author. Tel.: + 44 131 451 3797
E-mail address: [email protected]
Highlights

 Density and viscosity of a CO2 rich stream at 243 to 423 K and up to 150 MPa
 Saturation pressures were measured
 Experimental water saturated and undersaturated hydrate data were measured
 Developed models were independently validated using the experimental data
 1. Introduction

Carbon Capture and Storage (CCS) is presently being studied as one of several options

against the increasing anthropogenic CO2 emissions from the burning of fossil fuels. A CCS

chain will consist of three main components: Capture, Transport and Storage. Depending on

the capture processes, the flue gases are likely to contain various contaminants/impurities and

compositions of these contaminants (de Visser et al., 2008). The presence of these impurities

could lead to challenging operating and engineering issues all over the CCS chain, such as

safety and toxicity, pressure and compression requirement, dehydration requirement,

etc...The properties of a CO2-rich captured stream may differ significantly from pure CO2. No

experimental data are currently available for such fluids in the open literature for validation

of existing models.

For an oxyfuel combustion power plant, the main contaminants would be argon (Ar),

nitrogen (N2), oxygen (O2) and water (H2O). These impurities were selected for studies in this

communication. To avoid two-phase flow and also to increase the density of the system,

CCS streams are planned to be transported as liquid or dense phase, therefore knowledge of

the phase envelope of CCS is essential. In this paper, saturation pressures of a multi-

component rich CO2 mixture were measured from 253 to 293 K. Water is also expected to be

present in the stream, knowledge of the maximum allowable water content in CO2-rich fluids

is critical for safe transport of CO2 to storage sites. The presence of water may result in

corrosion, ice and/or gas hydrate formation and pipeline blockage, so the fluid system should

meet certain dehydration requirements. Experimental measurements of the locus of incipient

hydrate-liquid water-vapour curve for the mixture in equilibrium with liquid water are
presented herein up to 36 MPa and new experimental data are reported for the water content

of in the liquid region in equilibrium with hydrates at 15 MPa and temperatures range from

233.15 to 283.15 K. Density and viscosity of the mixture were also measured from 243.15 to

423.15 K at pressure up to 150 MPa.

The Cubic-Plus-Association (CPA-EoS) or the Soave-Redlich-Kwong (SRK) equation of

state combined with the solid solution theory of van der Waals and Platteeuw (1959) as

developed by Parrish and Prausnitz (1972) was employed to model the fluid and hydrate

phase equilibria as previously described by Chapoy et al. (2012). The thermodynamic model

was used to predict and compared the properties experimentally obtained (saturation pressure,

hydrates). The thermodynamic model was further developed by introducing a volume

correction factor for density calculation. A corresponding state model was also developed to

predict viscosity of the CO2-rich stream.

2. Experimental section

2.1 Materials

Carbon dioxide (CO2) was purchased from BOC and has a certified purity higher than 99.995

vol%. Composition of the CO2-rich synthetic mixtures prepared by BOC is given in Table 1.

De-ionized water was used in all hydrate tests

2.2 Saturation Pressure Measurements and Procedures

The experimental set-up used consisted of an equilibrium cell, cryostat, rocking/pivot

mechanism, and temperature/pressure recording equipment controlled by a PC (Fig 1). The

equilibrium cell is a piston-type variable volume (maximum effective volume of 300 ml),

titanium cylindrical pressure vessel with mixing ball, mounted on a horizontal pivot with

associated stand for pneumatic controlled rocking through 180 degrees. Rocking of the cell,
and the subsequent movement of the mixing ball within it, ensures adequate mixing of the

cell fluids. For the tests reported here, the cell was rocked through 180 degrees at a rate of 7

times per minute. Cell volume, hence pressure, can be adjusted by injecting/withdrawal of

liquid behind the moving piston.

The rig has a working temperature range of 183 to 373 K, with a maximum operating

pressure of 70 MPa. System temperature is controlled by circulating coolant from a cryostat

within a jacket surrounding the cell. The cryostat is capable of maintaining the cell

temperature stability to within better than 0.05 K. To achieve good temperature stability, the

jacket is insulated with polystyrene boards, while the connecting pipe work is covered with

plastic insulationfoam. The temperature is measured and monitored by means of a PRT

(Platinum Resistance Thermometers) located within the cooling jacket of the cell, which is

calibrated regularly against a Prema 3040 precision thermometer. Cell temperature can be

measured with an accuracy of 0.05 K. A Quartzdyne pressure transducer with an accuracy

of ± 0.04 MPa was used to monitor pressure. Temperature and Pressure are monitored and

recorded by the PC through an RS 232 serial port.

A typical test to determine the bubble point of the mixture is as follows: The cell was charged

with the test sample and set to the desired temperature for the measurement. The sample

volume was then reduced by pumping liquid into the cell (behind the moving piston), at the

opposite end to the sample. By this means the sample pressure was increased such that the

sample was at a pressure significantly higher than the expected bubble point pressure. The

cell was then rocked to mix the contents and ensure equilibrium. The sample volume was

then increased step-wise by removing measured quantities of the pumped liquid behind the

piston. At each step mixing was continued until equilibrium was achieved, indicated by a

constant pressure. The stabilised equilibrium pressures and change in sample volumes were

plotted and the bubble point was indicated by a sharp change in the pressure versus volume
plot. An example is shown in Fig. 2. As can be seen from Fig.2 the bubble point is easy to

identify with an accuracy of ±0.03 MPa.

2.3 Hydrate Dissociation Measurements and Procedures

Dissociation point measurements were conducted using the isochoric step-heating method

developed in this laboratory, which has previously been demonstrated as being considerably

more reliable and repeatable than conventional continuous heating and/or visual techniques

(Tohidi et al., 2000). Figure 3 shows the apparatus used to determine the phase equilibrium

conditions. The phase equilibrium is achieved in a cylindrical cell made of Hastelloy. The

cell volume is about 80 cm3 and it can be operated up to 40 MPa between 243 K and 323 K.

A detailed description of the apparatus and test procedure can be found elsewhere (Haghighi

et al., 2009). The weight of fluids (i.e. water and the multicomponent CO2 fluid) injected are

recorded prior to any measurements and the overall feed can be calculated.

2.4 Water Content Measurements and Procedures

The core of the equipment for water content has been originally described by Chapoy et al.

(2012), the setup is comprised of an equilibrium cell and a set-up for measuring the water

content of equilibrated fluids passed from the cell. The equilibrium cell is similar to the one

described in the saturation pressure measurements (Fig. 1)

The moisture content set-up is comprised of a heated line, a Tuneable diode laser adsorption

Spectroscopy (TDLAS) and a flow meter. The tuneable diode laser absorption spectroscopy

(TDLAS) for accurate water content measurements has been purchased from Yokogawa.

The unit is constructed of polished Monel and thus can be used with corrosive gases such as

H2S. The set-up is such that a tuned infrared source is passed through the test sample to a

detector. The measurement of water content is based upon true peak area and therefore,

unlike other units which use peak height, it is not influenced by changes in background gas.
The unit has two measurement ranges 0-100 ppmV and 0-3000 ppmV, both having a stated

accuracy of ±1% of full scale.

At the start of the test around 10 ml of 0.1 mm glass beads are placed in a cup shaped

depression in the top of the piston. 5 ml of deionised water was then mixed with the glass

beads. The glass beads have been found to aid in formation and dissociation of hydrates in

previous work, helping to achieve equilibrium. The cell was then closed and the temperature

reduced to 263.15 K and evacuated before injecting pure CO2 or the multicomponent

mixture. The cell temperature and pressure were then adjusted to achieve the desired test

conditions. The cell temperature was then cycled to lower and higher temperatures than the

set point over at least 20 hours. This was confirmed as being sufficient time for equilibrium

to be achieved by conducting water content measurements over a number of days in one test.

It was found that the water content remained stable.

The key to getting a reliable and stable water content measurement was to achieve a steady

flow rate of equilibrated sample through the measurement set-up. The top cell valve was

sufficiently opened to achieve a flow rate of between 0.5 and 1 litre per minute through the

hygrometer at the same time nitrogen was introduced into the base of the cell in order to

maintain the pressure constant. The water content reading from the TDLAS was then

monitored until it was stable for at least 10 minutes. This was then taken as the moisture

content of the equilibrated fluid passing from the cell. During sampling the heated line was

maintained at a temperature of 368.15 K. The repeatability was found to be good, within the

estimated experimental accuracy of ±5 ppm mole.

2.5 Viscosity and Density Measurements

2.5.1 Experimental Apparatus

All viscosity measurements were conducted in an in-house designed and constructed set-up, a

schematic view is shown in Figure 4. This setup has been designed to have a maximum

working pressure of 200 MPa and a maximum working temperature of 523.15 K. The set-up

is located inside the chamber of an oven, manufactured by BINDER GmbH, capable of being

used at temperatures from 203.15 to 473.15 K.

The set-up is comprised of two small cylinders, with volumes of 25 cm3, connected to each

other through a capillary tube with measured length of 14.781 metres and a calculated

internal diameter of 0.29478 mm. An oscillating U tube densitometer Anton Paar DMA-HPM

is connected to the set-up. The measuring cell contains a U-shaped Hastelloy C-276 tube that

is excited to vibrate at its characteristic frequency electronically. The DMA-HPM is

connected to a mPDS 2000V3 evaluation unit which measures the period of oscillation with a

resolution of seven significant digits. The temperature of the vibrating tube cell is measured

by a built-in thermometer connected to the mPDS 2000V3 unit. Two three-way valves, one

on top of the cylinders connected to capillary tube and one on top of the densitometer, are

installed to inject the sample inside the cylinders, tube system and densitometer. The base

side of the two cylinders are connected to opposite sides of a push-pull, motor driven mercury

pump. This pump can move the sample fluid forwards and backwards between the two

cylinders. There is also a hand pump connected to the system to control the pressure of the

entire fluid system by injection and withdrawal of mercury. Both the opposed piston pump

and the hand pump are fitted with Mitutoyo linear transducers readable to 0.005 mm on

Mitutoyo SD-D1E readouts. As a 1 mm movement represents 0.151 cm3 displacement in both

pumps, the readability is 0.000755 cm3. The opposed piston pump has a variable control with

which the speed can be adjusted to a maximum of 5cm3/sec. The rate can be set with an error

margin of ±0.00003 cm3/sec.

 2.5.2 Procedures

The capillary tube viscosity measurement method has been employed to measure the

viscosity of CO2 systems with impurities. In each test, the set-up was loaded with the sample

mixtures through the injection point on top of the densitometer after vacuuming the entire

system. After disconnecting the sample cylinder from the system, the sample fluid was

pushed through the capillary tube into the other cylinder using the push-pull mercury pump.

The temperature of the system was set to the desired condition and the desired pressure was

set using the hand pump. Once conditions had stabilised after isolating the densitometer by

closing the related valve the sample was pumped through the capillary tube at a number of

flow rates. To ensure the consistency of the measurements, at each pressure, viscosities were

determined at two or three different flow rates and at each flow rate three readings were

logged, so the reported viscosity data in this study are an average of at least six or nine

separate readings.

Pumping the sample fluid through the capillary tube by the piston pump resulted in a

dynamic differential pressure that was monitored and recorded until stable. The pump was

then stopped to record the static differential pressure. The difference between the dynamic

and static differential pressure was used as the pressure drop across the tube. To ensure

laminar flow conditions, Reynolds numbers were checked for the flow rates in which the

measurements were performed. The Poiseuille equation, Eq (1), can relate the pressure drop

across the capillary tube to the viscosity, tube characteristics and also flow rate for laminar

flow:

(1)

Where, ΔP is differential pressure across the capillary tube viscometer in psi, Q represents

flow rate in cm3/sec, L is length of the capillary tube in cm, D refers to internal diameter of
the capillary tube (0.029478 cm), η is viscosity of the flown fluid in cP and C is the unit

conversion factor equal to 6894757 if the above units are used.

The internal diameter of the tube was calculated by knowing the length and volume of the

tube. The tube length changes with temperature but this had no noticeable influence on the

obtained viscosity. The set flow rate has no effect on the accuracy of the viscosity

measurement. Only differential pressure as a variable in the above formulation can cause

error in viscosity measurement. The usual variation in differential pressure measurement is

0.01 psi and this leads to ±1% of error in the calculated viscosity for those measured in this

study.

Density measurements were performed when the temperature of the vibrating tube became

stable. Once conditions had stabilised, the oscillation period of the U-tube was determined

from the interface mPDS 2000V3 evaluation unit. The measurement of density with a

vibrating tube densitometer is not absolute, thus, the raw data (period of oscillation) was be

further treated to obtain the densities. The relationship (Eq (2)) between them is:

(2)

where  (T , P) is the sample density at temperature T and pressure P,  (T , P) is the period of

oscillation at temperature and pressure, A(T , P) and B(T , P) are the apparatus parameters

depending on temperature and pressure, and they must be determined from calibration

measurements. In our calibration, CO2 density is used as a reference substance at two

different pressures (the lowest and the highest desired pressures in the system at the same

temperature) in gas, liquid and supercritical phases. The apparatus parameters were defined

as follows:

(3)

(4)
 3. Thermodynamic and viscosity modelling

3.1 Phase Equilibria Model

A detailed description of the original thermodynamic framework used in this work can be

found elsewhere. In summary, the thermodynamic model based on the uniformity of fugacity

of each component throughout all the phases is extended to model the equilibrium conditions.

The CPA-EoS or the SRK-EoS (if no water is present) are used to determine the component

fugacities in fluid phases. The hydrate phase is modelled using the solid solution theory of

van der Waals and Platteeuw (1954) as developed by Parrish and Prausnitz (1972). The

critical properties of the components used in this work and the CPA parameters for water are

given in Table 2 and Table 3, respectively. The non-water binary interactions parameters are

reported in Table 4. The CPA-EoS binary interaction parameters for O2 – water and Ar –

water are listed in Table 5, the parameters for N2 – water and CO2 –water were reported in

previous works (Chapoy et al., 2012; Haghighi et al., 2009).

3.2 Density Calculation

In this work, the volume for CO2 or a CO2-rich mixture calculated by the SRK-EoS is
corrected using the exact volume of CO2 at the given T and P.
(5)
Where V EoS is the molar volume obtained from the equation of state. The correction of molar
volume in the Eq. (5), V c , is defined as:

(6)

xi is the composition of component i in the phase in which the volume is calculated. For CO2,
Vi c is defined by

(7)
For the other components, Vi c is set to 0. The carbon dioxide density is computed from the
MBWR equation in the form suggested by Ely et al. (1987):

9 15
P   an (T )  n   an (T )  2 n 17 e 
2

n 1 n 10 (8)

3.3 Viscosity Model

The proposed model is a modification of the corresponding state viscosity model described in
Pedersen and Christensen (2007) . According to the corresponding states principles applied
to viscosity, the reduced viscosity, , for two components at the same reduced

pressure, , and reduced temperature, , will be same.

(9)

Based on the dilute gases considerations, viscosity at critical point can be approximated as:

(10)

Where, M denotes the Molecular weight. Thus, the reduced viscosity can be expressed as:

(11)

For one component as a reference component if the function f in Eq. (9) is known, it is
possible to calculate the viscosity of any other components, such as component x, at any
pressure and temperature. Thus,

(12)
Where, 0 refers to the reference component. Methane with the viscosity data published by
Hanley et al. (1975) was selected as the reference fluid in the original Pedersen Model.

In this work, CO2 with the viscosity data published by Fenghour et al. (1998) has been
selected as the reference fluid. The viscosity of CO2 as a function of density and temperature
can be calculated from the following equation:

(13)

Where, is the zero-density viscosity and can be obtained from the following equation:

(14)

In this equation, the zero-density viscosity is in units of Pa.S and temperature, T, in K. The
reduced effective cross section, , is represented by the empirical equation:

(15)

Where the reduced temperature, , is given by

(16)

And the energy scaling parameter, . The coefficients ai are listed in Table 6.

The second contribution in Eq. (13) is the excess viscosity, , which describes how
the viscosity can change as a function of density outside of the critical region. The excess
viscosity correlation can be correlated as follows:

(17)

Where, the temperature is in Kelvin, the density in kg/m3 and the excess viscosity in Pa.s.
The coefficients dij are shown in Table 7. The corresponding states principle expressed in Eq.
(12) for the viscosity of pure components works well for mixtures. Pedersen et al. (1984)
have expressed the following expression to calculate the viscosity of mixtures at any pressure
and temperature.
 (18)

Where
(19)

The critical temperature and pressure for mixtures, according to recommended mixing rules
by Murad and Gubbins (1977)Erreur ! Source du renvoi introuvable., can be found from:

(20)

(21)

The mixture molecular weight is found from

(22)

where and are the weight average and number average molecular weights,
respectively.

(23)

(24)
The parameter  for mixtures in Eq. (19) can be found from

(25)

Also,  for the reference fluid can be found from the Eq. (25) by replacing the molecular
weight of the mixture with that of the reference fluid, carbon dioxide. The reduced density,
, is defined as
 (26)

The critical density of carbon dioxide, , is equal to 467.69 kg/m3. The Modified Benedict–
Webb–Rubin (MBWR) equation of state would be applied for computing the reference fluid
density, , at the desired pressure and temperature of . The mathematical

equation of the MBWR has been presented by Ely et al. (1987).

The procedure below should be followed to calculate the viscosity of CO2 systems with
impurities by the proposed corresponding state principle model:
1. Calculate the Tcmix, Pcmix and Mmix from Eq. (20), (21) and (22) respectively.
2. Obtain the CO2 density at from the MBWR EOS and calculate the

reduced density from Equation (26)

3. The mixture parameter, mix, and 0 should be calculated from Eq. (25)
4. The reference pressure and temperature, P0 and T0, should be calculated from
Equation (19)
5. Calculate the CO2 reference fluid, , in Eq. (18) from Eq. (13)
6. Calculate the mixture viscosity from Eq. (18)

4. Results and discussions

4.1 Phase Envelope

Bubble point measurements were carried out for the multicomponent mixture from about 253

to about 283 K. The results are listed in Table 8 and plotted in Figure 5. As seen in Figure 5,

the SRK-EoS model can predict the phase envelope of the multi component systems with

good accuracy, the predictions are in better agreement at higher temperature. The predicted

critical temperature and pressure are 296.9 K and 8.79 MPa, respectively.

4.2 Hydrate Stability

The experimental hydrate dissociations for the mixture in equilibrium with water are reported

in Table 9 and plotted in Figure 6 along with the prediction of the dry multicomponent

mixture phase envelope and the pure CO2 hydrate stability zone. As seen in the figure, we
have first a vapour + hydrate + liquid water line, then a vapour +liquid rich CO2 + hydrate +

liquid water line and finally a liquid rich CO2 + hydrate + liquid water line. The bubble point

of the mixture is higher than pure CO2, hence hydrates would be more stable in the liquid

region, as the vapour +liquid rich CO2 + hydrate + liquid water line is intersecting the bubble

line at higher temperature.

For this system the thermodynamic model is in excellent agreement with the new

experimental data in the vapour phase region and within 0.5-0.8 K above the dew line. The

ratio between the water mole fraction and the mixture fraction has a limited effect in this

range of aqueous fraction as nitrogen, oxygen and argon are weak hydrate former. It is also

interesting to note that the hydrate stability zone of the multicomponent mixture is more

stable than any of the pure constituent of the mixture above the bubble line.

4.3 Water content in equilibrium with hydrates

The experimental data for the pure CO2 and the multicomponent systems are presented in

Table 10 and plotted along with predictions of the thermodynamic model (descriptions given

in next section) in Figure 7. As can be seen from the figure the experimental and predicted

data are in good agreement with some deviation (AAD≈ 5%). As expected it can seen that

for the multicomponent system less water can be dissolved than in pure CO2, because the

amount of water that can be dissolved in nitrogen, oxygen or argon is far lower at same

temperature and pressure than in liquid CO2. The water content is in average about 10-30%

lower with these impurities than in pure CO2.

4.4 Viscosity

The experimental and modelling results for the viscosity of the stream are given in Table 11,

Figure 8 and Figure 9. All experiments for the stream were conducted at pressures above

saturation or in the supercritical region and then at low pressures, i.e., in the single gas phase

region. Then, the viscosity of each conducted test was calculated using the modified Pedersen
model. The results are also shown in Tables 11. As can be seen from the table, the model

predictions and experimental data are in good agreement. The Absolute Average Deviation

(AAD) is 1.7%.

4.5 Density

Densities of this multi component mixture with 10% impurity were measured at different

pressures and temperatures in gas, liquid and supercritical regions. Both experimental and

modelling results with and without density correction are shown in Table 12, Figure 10 and

Figure 11. In addition the Absolute Average Deviations (AADs) for all data are listed in the

table. As can be seen, by employing density correction using SRK-EoS the absolute average

deviation reduces from 5.4% to 1.7%.

It is interesting to note the peculiar behaviour of the density at temperature above the critical

temperatures compared to the density of pure CO2 as seen in Figure 12 in which the

difference between the density of the multicomponent mixture and pure CO2 peaks is plotted

at 323.48 K. A maximum reduction of the CO2 capacity at a certain pressure under the given

temperature is observed for the CO2 mixtures. The maximum reduction is 180 kg/m3 at a

pressure around 12 MPa.

5. Conclusion

Knowledge on how impurities/contaminants impacts the phase behaviour and the

thermophysical properties of CO2 rich systems is currently of great importance for the

deployment of CCS. As discussed in this work, no data sets are available for such systems.

In this communication impact of impurities on the phase behaviour and thermophysical

properties of a CO2-rich stream have been studied on

(i) The phase envelope of the stream from 253 to 293 K

(ii) The hydrate stability of a CCS stream up to about 35 MPa

(iii) The dehydration requirement

(iv) The viscosity and density of the mixture

(v) Models have been developed to calculate and predict these properties
Acknowledgements
This work is part of an ongoing Joint Industrial Project (JIP) conducted jointly at the Institute
of Petroleum Engineering, Heriot-Watt University and the CTP laboratory of MINES
ParisTech. The JIP is supported by Chevron, GALP Energia, Linde AG, OMV, Petroleum
Expert, Statoil, TOTAL and the UK National Grid, which is gratefully acknowledged. The
authors would also like to thank the members of the steering committee for their fruitful
comments and discussions.
References

Chapoy, A., Burgass, R., Tohidi, B., Austell, J.M., Eickhoff, C., 2011. Effect of Common
Impurities on the Phase Behavior of Carbon-Dioxide-Rich Systems: Minimizing the
Risk of Hydrate Formation and Two-Phase Flow. SPE J. 16, 921-930 (document ID
SPE-123778-PA)

Chapoy, A., Haghighi, H., Burgess, R., Tohidi, B. 2012. On the Phase Behaviour of the
Carbon Dioxide - Water Systems at Low Temperatures: Experimental and Modelling.
J. Chem. Therm. 47, 6-12.

De Visser, E., Hendricks, C., Barrio, M., Molnvik, M. J., de Koeijer, G., Liljemark, S., Le
Gallo, Y., 2008. Dynamis CO2 quality recommendations. International Journal of
Greenhouse Gas Control. 2, 478 – 484.

Ely, J.F., Magee, J.W., Haynes, W.M., 1987. Thermophysical properties for special high CO2
content mixtures. Research Report RR-110, Gas Processors Association, Tulsa, OK.

Fenghour, A., Wakeham, W.A., Vesovic, V., 1998. The Viscosity of Carbon Dioxide. J.
Phys. Chem. Ref. Data. 27, 31-44.

Haghighi, H., Chapoy, A., Burgess, R., Tohidi, B., 2009. Experimental and thermodynamic
modelling of systems containing water and ethylene glycol: Application to flow
assurance and gas processing. Fluid Phase Equilib. 276, 24-30.

Hanley, H.J.M., McCarty, R.D., Haynes, W.M., 1975. Equation for the viscosity and thermal
conductivity coefficients of methane. Cryogenics 15, 413–417.

Kontogeorgis, G.M., Yakoumis, I.V., Meijer, H., Hendriks, E.M., Moorwood, T., 1999.
Multicomponent phase equilibrium calculations for water–methanol–alkane mixtures.
Fluid Phase Equilibr. 158, 201-209.

Marshall, D. R., Saito, S., Kobayashi, R. 1964. Hydrates at high pressures: I methane +
water, argon + water, and nitrogen + water systems. AIChE J. 10, 202-205.

Murad, S., Gubbins, K.E., 1977. Corresponding states correlation for thermal conductivity of
dense fluids. Chem. Eng. Sci., 32, 499–505.

Parrish, W.R., Prausnitz, J.M. 1972. Dissociation pressures of gas hydrates formed by gas
mixtures. Ind. Eng. Chem. Process. Des. Develop. 11, 26-34.

Pedersen, K.S., Christensen, P.L., 2007. Phase behaviour of petroleum reservoir fluids. CRC
Press, Taylor & Francis Group.
Pedersen, K.S., Fredenslund, Aa., Christensen, P.L., Thomassen, P.,1984. Viscosity of crude
oils. Chem. Eng. Sci. 39, 1011–1016.

Poling, B.E., Prausnitz, J. M., O’Connell, J. P. 2000. The properties of gases and liquids, 5th
ed.; Mc Graw Hill: New York.

Tohidi, B., Burgass, R.W., Danesh, A., Todd, A.C., Østergaard, K.K., 2000. Improving the
accuracy of gas hydrate dissociation point measurements, Ann. N.Y. Acad. Sci. 912,
924-931

Van Cleeff, A., Diepen, G. A. M., 1960. Gas hydrate of nitrogen and oxygen, Rec. Trav.
Chim. 79, 83-86.

Van der Waals, J.H. Platteeuw, J.C., 1959. Clathrate solutions. Adv. Chem. Phys. 2, 2-57.

Younglove, B.A., Ely, J.F., 1987. Thermophysical Properties of Fluids. II. Methane, Ethane,
Propane, Isobutane, and Normal Butane. J. Phys. Chem. Ref. Data, 16, 577-798.
 Cooling Fluid in/out

P Transducer
2-way valve
(i.e. gas injection)

Cooling Jacket

Equilibrium Cell
Pivot

Mixing Ball

Piston

T Probe

2-way valve

Pressure
Transducer

Pressure Buffer

Fig. 1. Schematic illustration of equilibrium rig used for saturation pressure measurements
 12

11

10
P /MPa

6
0 5 10 15 20 25 30
V / cm3
Fig. 2. Plot showing an example of bubble point determination from plot of change in cell
pressure, P, versus volume, V, from test at 283.25 K. As shown in the figure, the measured
bubble point pressure at 283.25 K is 8.48 MPa.
Fig. 3. Schematic illustration of high-pressure hydrate rig
Fig. 4. Schematic drawing of the viscosity - density set-up
 10

6
P /MPa

0
220 240 260 280 300
T/ K
Fig. 5. Experimental and predicted phase envelope of the synthetic mixture. (), this work.
Black lines: bubble lines predictions using the SRK-EoS; Dotted black lines: dew lines
predictions using the SRK-EoS; Grey broken lines: predicted vapour pressure of pure CO2
using the SRK-EoS.
 40

35

30

25
P /MPa

20

15

10

0
260 270 280 290 300
T/ K
Fig. 6. Predicted and experimental hydrate stability of the CO2-rich stream in presence of
distilled water. (), this work; Black lines: hydrate stability zone predicted using the CPA-
EoS model using an aqueous mole fraction of 0.8; (), dry system saturation points, this
work; Dotted lines: phase envelope of the dry system (no water) using the SRK-EoS.
Broken lines: pure CO2 hydrate stability zone predicted using the CPA-EoS model. (), pure
N2 hydrate stability zone Erreur ! Source du renvoi introuvable.; (), pure O2 hydrate
stability zone Erreur ! Source du renvoi introuvable.; (), pure Ar hydrate stability zone
(Marshall et al., 1964) (), pure CO2 hydrate stability zone (Chapoy et al. 2011).
 2000

1800

1600
yw (106) / mole fraction

1400

1200

1000

800

600

400

200

0
223.15 233.15 243.15 253.15 263.15 273.15 283.15
T/ K
Fig. 7. Plot showing experimental water content data and predictions for the water content of
pure CO2 and the CO2-rich stream in equilibrium with hydrates at 15 MPa and different
temperatures. (), pure CO2, this work; Dotted lines: water content predictions using the
CPA-EoS model for pure CO2; (), multicomponent system, this work; Black Dotted lines:
predictions using the CPA-EoS model for pure the multicomponent system. (Error bars:
5%).
 350

300
Viscosity, η / Pa.s

250

200

150

100

50

0
0 25 50 75 100 125 150
P / MPa
Fig. 8. Predicted and experimental viscosity of the multicomponent CO2-rich stream. Black
lines: Predictions using CSP model. Black Dotted lines: Predictions using CSP model at the
bubble and dew pressures of the system. Data inside the grey box are plotted in Fig. 9. This
work: (), T = 243.15 K (), T = 253.15 K (), T = 273.15 K (), T = 283.15 K (), T =
298.15 K (), T = 323.15 K (), T = 373.15 K (), T = 423.15 K.
 35

30
Viscosity, η / Pa.s

25

20

15

10

0
0 2 4 6 8 10
P / MPa
Fig. 9. Predicted and experimental viscosity of the multicomponent CO2-rich stream. Black
lines: Predictions using CSP model. Black Dotted lines: Predictions using CSP model at the
bubble and dew pressures of the system. This work: (), T = 273.15 K (), T = 283.15 K
(), T = 298.15 K (), T = 323.15 K (), T = 373.15 K (), T = 423.15 K.
 1250

1000
Density, ρ / kg.m-3

750

500

250

0
0 25 50 75 100 125 150
P / MPa
Fig. 10. Predicted and experimental density of the multicomponent CO2-rich stream. Black
lines: Predictions using the corrected SRK-EoS model. Black Dotted lines: Predictions using
the corrected SRK-EoS model at the bubble and dew pressures of the system. Data inside the
grey box are plotted in Fig. 11. This work: (), T = 273.26 K (), T = 283.31 K (), T =
298.39 K (), T = 323.48 K (), T = 373.54 K (), T = 423.43 K.
 110
100
90
80
Density , ρ / kg.m-3

70
60
50
40
30
20
10
0
1 2 3 4
P / MPa
Fig. 11. Predicted and experimental density of the multicomponent CO2-rich stream. Black
lines: Predictions using the corrected SRK-EoS model. This work: (), T = 273.26 K (), T
= 283.31 K (), T = 298.39 K (), T = 323.48 K (), T = 373.54 K (), T = 423.43 K.
 0

-50
ρ / kg.m-3

-100

-150

-200
0 25 50 75 100 125 150
P / MPa
Fig. 12. Predicted and experimental density difference ρ= ρMIX –ρCO2, between the
multicomponent CO2-rich stream and pure CO2 density at 323.48 K. Black lines: Predictions
using the corrected SRK-EoS model. This work: (), T = 323.48 K.
Table 1
Composition, mole% each component, of the synthetic multicomponent mixture used in this
study as prepared and certified by BOC.
Composition,
Components
mole%
CO2 Balance
O2 5.05 ±0.1%
Ar 2.05 ±0.06%
N2 3.07 ±0.04%
Table 2
Critical Properties of Pure Compounds (Poiling et al., 2000)Erreur ! Source du renvoi
introuvable.Erreur ! Source du renvoi introuvable..

Compound Pc / MPa Tc / K ω
CO2 7.386 304.21 0.2236
O2 5.043 154.58 0.0222
Ar 4.898 150.86 -0.0040
N2 3.394 126.05 0.0403
Table 3
CPA Parameters for Water (Kontogeorgis et al. 1999).
β
a 0 /MPa L2 mol-2 b /L mol-1 c1 ε /MPa L mol-1
(x103)
0.1228 0.01452 0.6736 16.655 69.2
Table 4
Binary interaction parameters between O2, Ar, N2 and CO2 for the SRK-EoS and CPA-EoS
used in this work.
CO2 O2 Ar N2
CO2 0 0.106 0.123 -0.03
O2 0 0 -0.014
Ar 0 -0.008
N2 0
Table 5
Binary interaction parameters between O2, Ar, and water for the CPA-EoS .
System BIPs
O2 - Water 0.005
Ar - Water -0.038
Table 6

Values of Coefficients ai for CO2 in Eq. (15).

i ai
0 0.235156
1 -0.491266
-2
2 5.211155 x 10
-2
3 5.347906 x 10
-2
4 -1.537102 x 10
Table 7

Values of dij coefficients in Eq. (17)

dij Value
-2
d11 0.4071119 x 10
-4
d21 0.7198037 x 10
-16
d64 0.2411697 x 10
-22
d81 0.2971072 x 10
-22
d82 -0.1627888 x 10
Table 8
Saturation pressure of the multicomponent mixture.

b
T/ Ka P / MPa

252.65 7.10
263.05 7.54
272.55 8.03
283.25 8.48
293.35 8.96
a
Uncertainty on temperature u(T) = ±0.1 K
b
Uncertainty on pressure u(P) = ±0.03 MPa
Table 9
Experimental hydrate dissociation conditions in the presence of distilled water for the
multicomponent system (AqFr: aqueous mole fraction in the system).

a b c
AqFr T/ K P / MPa

0.976 273.65 1.714

0.963 275.35 2.137
0.948 280.30 3.768
0.945 282.20 5.049
0.866 283.65 6.805
0.730 285.25 14.279
0.740 287.75 36.094
a
Uncertainty on the aqueous mole fraction u(AqFr) = ±0.002
b
Uncertainty on temperature u(T) = ±0.1 K
c
Uncertainty on pressure u(P) = ±0.03 MPa
Table 10
Water contents (106 mole fraction) for pure CO2 and the multicomponent system in
equilibrium with hydrates at 15 MPa.

yw, Water
a
T/ K 1) Contentb
CO2 MIX
288.15 2391 2010
278.15 1711 1375
273.15 1391 1050
268.15 1130 885
263.15 913 690
253.15 581 477
233.15 238 220

a
Uncertainty on temperature u(T) = ±0.1 K
b
Uncertainty on water content u(yw) = ±5%
Table 11
Experimental and predicted viscosity, η of the multicomponent mixture.
η /Pa.s
T/Ka P/MPaa c
Exp Predicted AD
243.15 21.15 160.5 155.6 3.0%
243.15 51.35 205.3 199.8 2.7%
243.15 102.10 258.0 263.4 2.1%
253.15 52.27 177.7 178.9 0.7%
253.15 102.86 236.8 239.3 1.0%
253.15 150.89 286.6 289.2 0.9%
273.15 1.77 14.4 14.2 1.0%
273.15 2.23 14.0 14.4 2.9%
273.15 20.84 106.3 103.6 2.5%
273.15 51.71 142.0 144.1 1.5%
273.15 102.20 194.3 198.2 2.0%
273.15 149.35 240.1 244.0 1.6%
283.15 1.72 14.6 14.7 1.0%
283.15 2.25 14.6 14.8 1.3%
283.15 20.05 89.0 89.7 0.8%
283.15 51.54 124.4 131.1 5.4%
283.15 102.34 186.1 182.4 2.0%
283.15 149.80 229.6 226.3 1.4%
298.15 2.05 15.3 15.5 1.4%
298.15 3.49 15.2 15.9 4.7%
298.15 10.30 42.9 45.3 5.4%
298.15 20.33 73.5 74.1 0.8%
298.15 50.83 114.5 113.2 1.1%
298.15 102.77 166.9 163.1 2.2%
298.15 148.72 205.7 202.4 1.6%
323.15 2.54 16.7 16.7 0.2%
323.15 3.66 17.0 17.0 0.3%
323.15 12.62 29.7 30.7 3.4%
323.15 20.53 53.4 54.1 1.3%
323.15 51.82 95.2 93.4 1.9%
323.15 102.27 138.8 136.8 1.5%
323.15 149.50 177.1 173.2 2.2%
373.15 1.51 19.0 18.9 0.4%
373.15 2.54 19.4 19.0 2.1%
373.15 20.07 33.0 32.8 0.4%
373.15 51.26 68.8 67.7 1.6%
373.15 100.03 104.7 103.0 1.7%
373.15 150.53 137.1 135.3 1.3%
423.15 2.39 21.2 21.2 0.1%
423.15 3.53 21.4 21.4 0.4%
423.15 27.03 35.4 34.7 1.9%
423.15 51.21 54.9 54.6 0.5%
423.15 102.25 86.8 85.9 1.0%
423.15 149.96 112.6 111.9 0.6%
AAD 1.7%
a
Uncertainty on temperature u(T) = ±0.1 K
b
Uncertainty on pressure u(P) = ±0.005 MPa cUncertainty on viscosity u(ρ) =±1%

Table 12
Experimental and predicted density of the multicomponent mixture.
Density /kg/m3 AD, %
a b
T/K P/MPa Model with Without With Without
Exp. c
Correction Correction Corr. Corr.
273.18 1.79 38.64 38.2 37.9 1.2 1.8
273.18 2.25 49.98 49.7 49.3 0.5 1.4
273.27 8.81 841.98 826.8 755.7 1.8 10.2
273.28 10.94 867.96 855.2 785.3 1.5 9.5
273.28 21.04 940.91 937.0 874.5 0.4 7.1
273.29 52.14 1,048.27 1,053.8 1,005.0 0.5 4.1
273.29 104.35 1,144.49 1,152.7 1,108.7 0.7 3.1
273.29 125.93 1,173.60 1,180.8 1,136.5 0.6 3.2
283.31 1.74 36.44 35.0 34.9 4.0 4.1
283.30 2.28 46.40 47.4 47.4 2.1 2.1
283.31 10.69 804.24 774.9 705.4 3.6 12.3
283.31 20.88 894.93 889.6 826.3 0.6 7.7
283.32 52.55 1,011.85 1,026.0 978.8 1.4 3.3
283.32 104.30 1,113.02 1,131.0 1,090.3 1.6 2.0
283.31 125.48 1,143.66 1,160.5 1,119.7 1.5 2.1
298.38 2.08 38.48 39.6 39.6 3.0 2.9
298.38 3.54 74.81 73.4 73.2 1.8 2.1
298.38 12.60 701.23 677.6 612.4 3.4 12.7
298.39 20.84 825.87 812.1 748.3 1.7 9.4
298.40 51.43 978.52 978.5 932.4 0.0 4.7
298.40 104.29 1,095.47 1,098.8 1,062.4 0.3 3.0
298.38 126.11 1,128.77 1,131.6 1,095.7 0.3 2.9
323.49 2.58 45.26 44.9 44.8 0.8 0.9
323.50 3.71 67.48 67.6 67.5 0.2 0.0
323.45 12.29 408.79 398.6 358.1 2.5 12.4
323.45 20.95 686.84 668.8 607.9 2.6 11.5
323.48 51.65 908.44 904.3 859.7 0.5 5.4
323.50 105.25 1,046.64 1,048.1 1,016.9 0.1 2.8
323.49 125.71 1,081.30 1,082.7 1,053.1 0.1 2.6
373.49 17.25 352.26 339.7 315.9 3.6 10.3
373.50 21.04 438.04 427.0 390.1 2.5 11.0
373.53 52.66 776.02 765.9 724.0 1.3 6.7
373.55 104.12 950.13 947.8 921.5 0.2 3.0
373.55 125.37 994.44 992.6 969.4 0.2 2.5
423.48 2.40 28.22 29.9 29.8 6.1 5.8
423.49 3.42 41.42 43.2 43.0 4.3 3.7
423.38 18.51 282.42 269.7 255.5 4.5 9.5
423.38 21.79 334.13 321.5 301.3 3.8 9.8
423.41 52.88 658.74 647.0 608.6 1.8 7.6
423.42 102.96 860.27 856.5 831.6 0.4 3.3
423.42 124.87 913.59 911.5 890.1 0.2 2.6
1.7 5.4
a
Uncertainty on temperature u(T) = ±0.01 K
b
Uncertainty on pressure u(P) = ±0.005 MPa
b
Uncertainty on density u(ρ) = ±0.74%