1

Removal of CO2 from Natural Gas using

Membrane Process: Design, Fabrication and
Parametric study
Faizan Ahmad, Lau Lok Keong, Azmi Shariff, Department of Chemical Engineering, Universiti
Teknologi PETRONAS, Malaysia

necessary to develop efficient processes for the removal of

CO2 from NG [5, 6]
Abstract— Natural Gas (NG) processing is one of the largest CO2 can be removed by the number of processes
markets for the industrial separation processes. Membrane considering the factors of; capital and operating costs, gas
process, relatively new technology among other available specifications and environmental concerns. The major
techniques, can be used for the removal of contaminants like
processes can be grouped as
carbon dioxide from NG. In this work, an experimental pilot
test rig would be designed and fabricated for the removal of
• Absorption Processes (Chemical and Physical absorption)
Carbon dioxide (CO2) from NG. The parametric study would • Adsorption Process (Solid Surface)
be planned and designed using Design of Experiment (DOE) • Hybrid Solution (Mixed Physical and Chemical Solvent)
approach. Different gas specifications and operating conditions • Physical Separation (Membrane, Cryogenic Separation) [7,
would be investigated for optimal cleaning of NG. In addition, 8, 9]
the study on pretreatment conditions, multistage system and Membranes processes are commercially proven
energy recovery would also be carried out. Moreover, technology for natural gas processing application.
mathematical models would be developed for membrane, Separation of CO2 with common polymeric or inorganic
pretreatment system and process flow. Based on the
(e.g. zeolite, sol–gel silica or carbon molecular sieve)
mathematical modeling of individual unit operations, process
simulation would help for scale up purposes. Finally, an membranes is achieved by differences in terms of diffusion
optimum model would be proposed based on the correlations rates and/or adsorption strengths of mixture components in
suggested by DOE and the proposed optimal model would be the polymer matrix or the inorganic membrane pores. For a
validated by experimental results. gas to permeate through a membrane surface, the gas must
first dissolve in the high-pressure side of the membrane,
Index Terms—Gas separation, Membrane process, Natural diffuse across the membrane wall, and evaporate from the
gas processing, Process design low-pressure side. Gas separation therefore works on the
principle that some gases are more soluble in, and pass more
I. INTRODUCTION readily through polymeric membrane than other gases [7, 10,
The composition of natural gas varies from source to 11].
source. Methane is the major component (75%-90%), but In membrane process, feed gas is pretreated before
natural gas also contains significant amounts of ethane, entering the membrane system to ensure efficient operation.
propane, butane and 1%-3% of other higher hydrocarbons The conditioning of feed stream controls the fouling,
[1]. In some deposits, it may contain complex contaminants plasticization and condensation of hydrocarbons on the
such as CO2, H2S, CO which constitutes environmental membranes [1, 12]. The pre-treatment equipment depends
hazards and also hindered natural gas processes. The on the conditions and composition of feed gas. Usually in
upgrading of low quality crude natural gas is an attracting the case for natural gas, the feed gas is filtered, as a first
interest due to the high demand for pipeline-grade gas in step, for removal of entrained particulates or aerosols
recent years. CO2 needs to be removed in order to; increase including sand, pipe scale. The feed gas is then cooled in a
the heating value of the gas, prevent corrosion of pipeline cooler. Any condensed liquids are removed in the gas/liquid
and process equipments and crystallization during separator. After liquid removal, the feed gas enters the feed
liquefaction process. [2, 3] preheater. The temperature control system is provided to
CO2 contents in the NG can vary from 4 to 50% maintain the gas at the desired operating temperature of the
depending upon the source gas or oil well. Before the membrane fibers. Finally, the heated gas enters the
transportation of NG, It must be pre-processed so as to meet membrane gas separators where it is separated into two
the typical pipeline specification of 2-5% CO2 [4]. Most of streams; the permeate, a low pressure CO2-rich stream and
the natural gas, produced in the lower 48 states of USA, the non-permeate or residue, a high pressure hydrocarbon
contains more than 5% CO2. Currently, many natural gas rich stream [7, 12].
wells are shut due to their low production rate and low The design of CO2 removal system depends on the
quality (i.e., high CO2 and/or H2S content) [5]. In Malaysia, following factors:
NG from Tangga Barat Cluster fields of PETONAS contains i) Selectivity and Permeability of the membrane
relatively high level of CO2 [6]. Therefore, it has become ii) Concentration of CO2 and the separation required
iii) The value of the gas
 2

iv) The location of the plant [1] scale study, a membrane separation system that can cater
high flow rates and pressures is required for CO2 separation.
II. BACKGROUND
Gas separation by membrane technology has become a IV. SCIENTIFIC OBJECTIVE
major industrial application only during the last few decades
A. Main Objective
but the study of gas separation has a long history [10]
Graham measured the permeation rates of all the gases The main focus of the work would be on the development
then known through different diaphragms [10, 13]. In 1940s of a process for CO2 removal from NG using membrane
to 1950s, Barer, Amerongen and Stern played an important technology.
role in the development of solution diffusion model for the B. Measurable Objectives
explanation of gas permeation [14, 15, 16]. The first
Listed below are the proposed measurable objectives:
company to establish a Prism membrane was Monsanto for
1. To design and fabricate pilot experimental test rig
hydrogen separation. The success of Monsanto encouraged
2. To apply Design of Experiment (DOE) approach
other companies like Cynara, Separex and Grace Membrane
for experiment planning and design.
Systems to produce membrane plants to remove CO2 in
3. To conduct parametric study under different
natural gas [10, 17].
operating conditions.
Datta and Sen optimized gas processing cost of the
4. To develop mathematical models for unit
membrane unit. It was shown that there is no unique
operations involved and simulate the process.
configuration that will be always optimum irrespective of the
5. To Optimize/validate the proposed model by
values of carbon dioxide concentration and cost of natural
experimental results.
gas. However, within certain ranges of the carbon dioxide
concentration and the cost of natural gas, the optimum
V. METHODOLOGY
configuration may be unique and the minimum gas
processing cost can be achieved by adjusting only the A. Conceptual Design
number of modules in each stage and the compressor power
[4]. Feed gas is prepared by mixing methane, carbon dioxide,
Wang enhanced operational flexibility and adaptability of ethane and propane using mass flow meters with each gas
membrane process, using one optimal method in which auto- cylinder. The flow rate of 2-5 million standard cubic foot per
controlling of the permeate gas flux was applied for the first day (MSCF) is maintained and sent to compressor for
time. Moreover, it was shown that the permeate gas flux increasing pressure up to almost 80-100 bars.
could be auto-controlled through a control valve installed on The compressed gas is then sent into preheater where
the residue gas line [18] temperature is raised to approximately 60-800C.
Qi and Hensen developed the optimal design strategy for
spiral membrane networks for gas separations [19] whereas T, F, C, P
Lababidi developed the mathematical model to optimize
three configurations including single stage, the two stage, T, F, C, P
and the continuous membrane column (CMC) [20]. The
simple models for the permeability and selectivity variations
of the CO2 /CH4 system have been derived by Safari, Mixer
Ghanizadeh and Rehmat that included both temperature and
pressure effects simultaneously [21]. F
J. Hau worked on upgrading low quality natural gas with C Membrane
H2S and CO2 selective membranes and studied process C T, F, C, P
design, economics, and sensitivity of membrane stage with M
recycle streams [22].
Lee, Feldkirchner, and others made field tests of
membrane modules for the separation of carbon dioxide
from low-quality natural gas. In their study, they investigated
the effects of the operating variables of pressure, feed flow
rate, and the carbon dioxide concentration in the feed. In
addition, computer models were applied for the separation of Compressor Preheater
gases under perfect mixing and cross flow conditions to the
analysis of the field data [5]
Fresh feed
III. PROBLEM STATEMENT
Figure 1: Flow sheet for the removal of CO2 from natural gas using
The above cited works for CO2 separation using membrane
membrane process
technology are mostly limited to laboratory scale. It is found
that laboratory scale experimental data is non-reliable for
commercial scale design especially in the case of high feed
Finally, the gas is introduced into membrane test skid where
flow rates and pressures [5, 23]. Hence, there is a lot of
potential in the field of CO2 separation from NG by gases will be separated. The permeate stream would contain
membrane process on commercial scale. Under the pilot carbon dioxide rich stream whereas retentate side will
 3

contain hydrocarbon rich stream. The analysis for flow rate, [3] Mohammad Hosein Safari, Amin Ghanizadeh, and Mohammad
Mehdi Montazer-Rahmati, “Optimization of membrane-based CO2-
pressures, temperatures and concentration will be done removal from natural gas using simple models considering both
before and after the separation through the membrane. Both pressure and temperature effects” International Journal of
streams will be mixed together in order to recycle back for Greenhouse Gas Control, 2008
the next cycle. [4] Datta, A.K. and Sen, P.K., “Optimization of membrane unit for
removing carbon dioxide from natural gas,” J. Membr. Sci. 283, 28
B. Design of Experiment (DOE) june 2006, pp. 291–298.
[5] A.L. Lee, H.L. Feldkirchner, S. A. Stern, A.Y. Houde, J.P. Gomez,
Design of Experiment (DOE) approach would be applied and H.S. Meyer, “Field tests of membrane modules for the separation
to plan and design the number of experiments to be carried of carbon dioxide from low quality natural gas”, Gas sep. Purif., Vol.
9, 10 May 1994, pp 35-43.
out in order to determine the correlations for scale up and [6] PETRONAS media releases & news 2008.
modeling purposes. These correlations include Available:http://www.petronas.com.my/internet/corp/news.nsf/2b372
Flux = ƒ (T, P, C, pretreatment conditions) bb45ff1ab3a48256b42002b19a7/d9473b4fd966e901482574eb002b3
Recovery = ƒ (T, P, C, pretreatment conditions) fce?OpenDocument
[7] Salako Abiodun Ebenezer, “Removal of Carbon dioxide from natural
Methane loss = ƒ (T, P, C, pretreatment conditions) gas for LPG production”, Semester project work. Institute of
Moreover statistical analysis and ANOVA study would be Petroleum Technology, Norwegian University of Science &
conducted using different DOE methods such as Taguchi Technology, Trondheim, Norway. 2005
[8] Maddox, R.N. Gas and Liquid Sweetening, 2nd ed., Maddox, R.N.,
and response methods. John M. Campbell (1974)
[9] W.J. Koros and R.T. Chern, “Separation of gaseous mixtures using
C. Experimental Parametric Study polymer membranes, in: R.W. Rousseau (Ed.)”, Handbook of
Based on DOE, experimental investigation would be Separation Process Technology, Wiley, New York, 1987, pp. 863–
953
conducted on the process to find the suitable operating
[10] R.W. Baker, Membrane Technology and Application, 2nd ed., John
conditions such as pressure, temperature and concentration. Wiley & Sons, Chichester, UK, pp. 287-295
Different pretreatment conditions would also be studied. [11] George Xomeritakis, Chung-Yi Tsai and C. Jeffrey Brinker, “Anodic
Moreover energy recovery system and multistage study can alumina supported dual-layer microporous silica membranes”,
Journal of Membrane Science, Volume 287, Issue 2, 15 January
also be done in order to get the optimal and economical 2007, pp. 157-161.
configuration. The fouling/concentration study can lead to [12] Ahmad Fauzi Ismail, “Specialized workshop on membrane gas
find the ways to prolong the lifespan of the membrane. separation technology”, Advanced Membrane Technology Research
Centre, Universiti Teknologi Malaysia, 2009.
D. Mathematical Modeling and Process Simulation [13] T. Graham, “On the absorption and dialytic separation of gases by
colloid septa”, Philos, Mag, 32, 1866, p 401
The mathematical modeling would be done for membrane, [14] R.M. Barrer, Diffusion in and through solids. Cambridge University
pretreatment system and process flow. Based on the Press, London, 1951.
modeling of these unit operations, process would be [15] G.J. van Amerongen, “Influence of structure of Elastomers on their
permeability to gases,” J. Appl. Poly. Sci. 5. 1950, p 307.
simulated using MATLAB that would serve for scale up [16] S.A. Stern, “Industrial applications of membrane processes: The
purposes. Finally, optimal model would be proposed based separation of gas mixtures,” Proceedings of the symposium southern
on DOE and validated by experimental results. Research Institute, Brimingham, May 1966.
[17] J. M. S. Hennis and M.K. Tripodi, “ A novel approach togas
separations using composite hallow fibre membranes,” Sep. Sci. and
VI. CONCLUSION Tech. 15, 1980, p 1059
[18] Wang, L., Shao, C., Wang, H., “Operation optimization of a
Even though the conventional technologies are still in membrane separation process through auto-controlling the permeate
operation for the removal of the CO2 from the natural gas, gas flux.” Sep. Purif. Technol. 55, 15 May 2007, p 30
the membrane process has certain advantages over others [19] R. Qi, M.A. Hensen, “Opitmal design of spiral wound membrane
available technologies. Membrane has good weight and networks for gas separations”, Journal of membrane science, Vol,
148, 22 May 1998, pp. 71-89.
space efficiency, which make it more applicable for off [20] H. Lababidi, Ghazi A., Al-Enezi and Hisham M. Ettoney,
shore environment. Moreover its high adaptability to “Optimization of module configuration in membrane gas separation,”
variation of CO2 content in the feed gas makes it a suitable Journal of membrane Science, Vol 112, 1996, pp 185-197.
[21] M.H. Safari, A. Ghanizadeh. M.M. Montazer-Rahamti,
process. Finally, periodical removal and handling of spent “Optimization of membrane based CO2- removal from natural gas
solvent or adsorbent used by conventional system make this using simple models considering both pressure and temperature
system more environmental friendly. effects,” International Journal of Green House Control, Vol. 105,
May 2008.
Thus, the optimal design for the removal of carbon [22] J. Hao, P.A. Rice, S.A. Stern, “Upgrading low quality natural gas
dioxide from the natural gas using membrane technology can with H2S and CO2 selective polymer membranes Part II. Process
prove itself to be key stone in the development of separation design, economics, and sensitive study of membrane stages with
recycle streams”, Journal of Membrane Science, Vol. 320, 23 march
technology.
2008, pp. 108-122.
[23] T.E. Cooley and W.L. Detloff, “Field tests show membrane process
REFERENCES attractive”, Chem Eng Prog, Vol. 81, 1985, p 45.
[1] Richard W. Baker and Kaaeid Lokhandwala, “Natural gas processing
with membranes: an overview,” Industrial Engineering Chemistry
Research, Vol. 4, Nov 2008, pp. 2109-2021
[2] J. Hao, P.A. Rice and S.A. Stern, “Membrane processes for the
removal of acid gases from natural gas. II. Effects of operating
conditions, economic parameters, and membrane properties”, Journal
of Membrane Science, Volume 81, Issue 3, 29 June 1993, pp. 239-
252.