УДК 622.276

Design and potential applications of polymeric membranes for Oil & Gas

Абдуллаев Эльдар Фазахирович – магистр наук в области нефтегазовой инженерии Абердинского университета (Шотландия).

Abstract: In the modern world energy demand is increasing day by day. Natural gas is one of the power sources that could meet the market requirements. However, one-third of gas resources have a high concentration of acid gases, such us CO₂ and H₂S. To meet selling specifications such hydrocarbon mixtures should be processed.

Membrane separation – is a relatively new method for CO₂ and H₂S removal, but it was proven to be an efficient technique. In fact, the processing cost depends on the membrane material nature and operating conditions, which is why some test tool is required for best process optimization. This paper gives a description of the main membrane properties (permeability and selectivity) and suggests a laboratory module design for their determination. The assembly suggests a wide range of operating conditions and membrane types to be applied.

Аннотация: В настоящее время потребление энергии увеличивается с каждым днем. Одним из источников энергии, который может удовлетворить мировые потребности, является природный газ. Однако треть газовых ресурсов имеет высокую концентрацию кислых газов, таких как CO2 и H2S. Для соответствия нормам сертификации такие смеси углеводородов должны быть переработаны.

Мембранное разделение смеси углеводородов – относительно новый метод удаления CO2 и H2S, но уже зарекомендовавший себя как эффективный. Стоимость обработки смеси газов зависит от природы материала мембраны и условий эксплуатации, поэтому для наилучшей оптимизации процесса требуется некоторый тестовый инструмент. В данной статье представлено описание основных свойств мембран (проницаемость и селективность) и предлагается конструкция лабораторного модуля для определения этих параметров.

Keywords: polymeric membranes, sour gas, separation, diffusion, sorption, gas transport.

Ключевые слова: полимерные мембраны, кислый газ, разделение, диффузия, сорбция, транспортировка.

1. Introduction

1.1 Role of natural gas

Economic growth and increase in population worldwide implies a significant increase in world energy demand. To meet this growth humanity should enhance hydrocarbons’ effectivity as an energy source and use renewables. Nowadays, the natural gas accounts for the largest increase in world energy sector. Total worldwide recoverable gas resources are estimated to be equal to 250 years of current gas production. Consumption of natural gas is projected to increase from 120 Tcf in 2012 (22% of total consumption) to 203 Tcf in 2040 (28% of total consumption) [1]. Figure 1.1 shows the growing in gas consumption till 2040.

Figure 1.1 Predicted natural gas consumption.

Natural gas remains a significant energy source for electric power sector and industrial sector. One of the ways of increasing efficiency of natural gas as an energy source is modifying processing process.

Gas composition varies, in addition to higher hydrocarbons and CH₄, it also contains CO₂ and N₂, as well as small amounts of the following gases: H₂S, He, O₂, Ar, H₂, and H₂O vapour. A significant part (up to 40%) of gas reserves contain more than 10% CO₂ and/or H₂S (sour gases) [2]. Several intensification methods that usually in use during hydrocarbons production, can negatively affect the gas composition. For example, some of the production intensification methods are injecting the wells with a high-pressure carbon dioxide (CD) mixture, and as a consequence, the produced gas has high amounts of CD [16].

Removing CO₂ and H₂S from the crude natural gas before combustion is necessary. It helps to increase the heating value of gas, decrease atmospheric pollution, and multiply the product price. There are different methods of sour gas processing, but some relatively new method with higher efficiency was introduced in last decades. The method implies using membranes as a separation tool.

1.2 Conventional methods of sour gas processing

Hydrogen sulfide removal

Nowadays, there are only two main methods used in the removal of hydrogen sulfide from natural gas. These are: physical/chemical absorption and adsorption. Speaking about chemical absorption with application of amine scrubbing, it is necessary to note, that it is the leading process (way) of hydrogen sulfide removal from sour natural gas [3]. In addition, the process under consideration is also capable of removing CD, and hydrogen sulfide content in the purified stream can be less than 4 ppm. It is emphasized that the purification process is based on absorption of hydrogen sulfide and further reaction with amine [4]:

2RNH₂ + H₂S ↔ (RNH3)₂S (1.1)

(RNH₃)₂S + H₂S ↔ 2RNH₃HS (1.2)

Carbon Dioxide Capture

CD removal plays a rather important and significant role. This is due to the fact that the gas is considered corrosive. According to the available specifications, the limit content of CD in gas can never be higher than 2 mol [17]. Amine scrubbing is the main method of removing CO₂ from natural gas [18]. Physical absorption with polyethylene glycol, methanol, potassium carbonate and water provides an opportunity to separate CD.

The description of the water absorption process is as follows: The acid gas enters the packed column, where the CD is dissolved in water. After that the concentrated stream is separated by means of air to obtain CD, while the water is recycled. It is emphasized that the efficiency of the process under consideration is guaranteed due to the fact that water is quite readily available [19].

Considering on the other hand, it should be noted that polyethylene glycol (PEG) has the best selectivity in contrast to water, and in addition is considered a non-corrosive solvent. The main disadvantage of using PEG is the extremely low regeneration rate [20].

The main advantage of the hot potassium carbonate absorption process is the possibility of removing quite a large amount of CD, and at the same time a small amount of hydrogen sulfide. The considered mechanism is based on reaction (1.3) of potassium carbonate solution with CD [5].

1.3 Membrane separation

The membrane is considered as a perm-selective barrier or interface between two phases and is the heart of every membrane process [29]. The gas separation mechanism is based on the components’ permeation rate through the membrane. Pressure differential across the membrane maintains the driving force. The stream entering the pressure cell is called feed-stream, the fluid that passes through the membrane is known as the permeate while the fluid that contains the retained components is called retentate. The driving force may derive from compression of feed gas to a high pressure; and/or the downstream side evacuation for permeant. Faster components more rapidly permeate through the membrane and become enriched on the permeate side, while the slower particles are concentrated in the retentate (residue) side. Permeability and selectivity are the only two criteria that must be met to produce a useful membrane. The technical success of membrane gas separation is mainly attributed to the engineering approach to reduce the membrane effective-thickness and to increase the packing density of the membrane module [30].

Thomas Graham first suggested the definition of the solution-diffusion mechanism in 1866. By exploiting this, isotope separation was achieved by a microporous membrane at large scale in 1945. Then, Van Amerongen (1950), Barrer (1951), Mears (1954), Stern (1966) and others stated the modern theories of gas permeation by developing solution-diffusion model [31]. In 1980, Monsanto opened the hydrogen-separating Prism membrane [32]. After Monsanto's success, other people were encouraged to go ahead with their membrane technologies.

Membrane advantages in hydrogen sulfide removal

The membrane technology can provide an alternative way for hydrogen sulphide removal. The key benefit is that the membrane does not require a solvent to operate (unlike amine scrubbing or methanol absorption) and this fact cuts down the cost of purchasing and supplying the solvent.

Another membrane’s advantage over CMS is that it could be operated with gases containing up to 16 mol of hydrogen sulphide [17]. Bhide and Stern have done an economical study for natural gas treatment using membranes and amine scrubbing. Some interesting facts were found during the research. Firstly, the processing cost in a membrane system (MS) is proven to be a function of the concentration of CD and hydrogen sulphide i.e. higher the content - the higher the price.

Membrane advantages in carbon dioxide removal

Additionally, membranes have advantages in the CD capture process. Comparing with the conventional methods, the membranes can remove CD along hydrogen sulphide and water with one step [26, 27]. In addition to low operating energy, membrane operating life is at least five years [33]. Unfortunately, the membrane lifetime depends on particulates were presented in the sour gas; therefore, removing of these particles recommended.

An economical research was published by Peters et al. to compare the amine scrubbing with MS for natural gas processing. The process was run on 60^∘C and 90 bars. The composition of feed gas was 72.4 mol CH₄ 20 ppm H₂S, 9.5 mol CO₂, 10 ppm H₂O and the remaining part for C₂ to C₆. Results showed that both technologies reached the sale gas specification of 4 ppm H₂S and 2 mol CO₂. The treated gas by amine scrubbing has better CD purity compared to the membrane separation, but it requires higher capital investment.

Features of membrane gas separation

The basic properties of membrane operations make them ideal for industrial operations. The most attractive features of membrane gas separation compared with other separation methods are the following:

• The absence of phase and temperature change phenomena, leading to lower energy requirement.
• Low maintenance costs because of the absence of moving parts.
• Easy plant operation due to steady continuous process.
• Due to small foot print and light weight, membranes are ideal for use on offshore platforms, in aboard aircraft, etc. where space and portability are significant factors.
• Easy to scale up based on laboratory or pilot-scale data due to the modular design of membrane.
• Can be combined with other separation processes in the same facility (hybrid processing).
• Small environmental impacts due to the absence of chemical additives etc., and usually high quality of final products.

However, it also has some disadvantages, which constraint its application in different systems. The problems of membrane process include fouling due to contaminated feed, expensive fabrication method, and incapability to handle corrosive substances. Also, polymer membrane process cannot sustain high-temperature condition.

2. Permeability measurement equipment design

Background

Methods for measuring gas permeability can be divided into two main types: the constant-volume variable-pressure method (also is known as time-lag method), and the gas concentration method.

I used the second method as it provides highly reproducible measurements, and allows to measure the diffusion of multiple components in a mixture, for each component individually, rather than only the total gas diffusion. So that the real gas separation process can be simulated [34].

Theoretically, it is required to split feed flow into permeate and renetante. To achieve that a membrane should be sealed in specialised pressure cell which isolates the feed (upstream side) from permeate (downstream) side, allowing gas exchange between the two parties to happen only by diffusion through the membrane. For a mixed-gas permeation measurement, it is important to minimise gas composition change over time on both sides that result from diffusion through the membrane.

This is achieved by continuously flowing the feed gas at a particular flow rate and pressure on the upstream side to sweep the entire membrane face, and to do the same on the downstream side using a sweep gas. The experiment is typically set up such that the feed flow is much greater than the amount of gas permeating through the membrane to minimise compositional changes on both feed and sweep streams: the stage cut (i.e. ratio of transmembrane gas flow to the feed flow rate) is typically less than 1% [35]. The permeances of the individual components can be calculated by measuring the permeate stream (i.e. the stream consisting of sweep gas plus permeating components) flow rate and its composition.

Scale of the process

The next module was designed to test membranes to determine their permeability in wide range of operating conditions (Figure 2.1)

Figure 2.1. Principal drawing of permeation equipment.

where comp. - compressor, ST1, ST2- storage tanks, PC-pressure controller, GC- gas chromatograph, MFC- mass flow controller, Software- self-design software to data analysis, WB – water bath.

Permeation cell

The permeation cell is modified Wicke Kallenbach design (Figure 2.2). It allows gas flow in and out, on both sides of the membrane, making it possible to have a flow of gas along the membrane. The module is custom cell made from stainless steel (Figure 2.3).

Figure 2.2.

The membrane to be tested is sealed within this cell supported on a porous metal disk and sealed to the upstream face of the cell using the O-rings. Of particular note, the use of a metal disk as a diaphragm support significantly reduces the dead volume in the diaphragm cell. This results in the most efficient passage of gas through the lower membrane surface even at low flow velocities. It should be emphasized that the upstream surface was purged with feed gas and the downstream surface was purged with argon, using immersion tubes to bring the flows as close as possible to the membrane surface. In all this, the flow rate was quite sufficient to ensure good and proper mixing of the gases at the membrane surface. The presented design was made with a program such as AutoCAD.

Figure 2.3 Pressure cell drawing.

Software

 The next stage is the formation of the appropriate software required to analyze the input data of the chromatograph, and in addition, to calculate the permeability of the membrane. The control software was written using MATLAB to create a static program of result analysis and results interpretation. 

Figure 2.4. Program visual window.

The user interface of this software is shown in (Figure 2.4). For example: Given that membrane thickness is 0.01 cm; Feed Pressure 75 cmHg; Permeate Flow pressure 75 cmHg; Permeate flow rate 10 cm³/sec, Membrane area 0.13 cm². Imagine that initial composition of threating components is 0.3 and 0.1 of total gas. After passing a membrane composition changes to 0.155 and 0.080 respectively. The permeability of the first component was calculated as 110 Barrer and 795 Barrer for the second one.

Test running and properties calculation

Recommended test time is in the range of 7-20 hours for getting more accurate results. Several steps have to be done to determine membrane gas permeability. Firstly, required gas mixture should be prepared and analysed using a gas chromatograph. Secondly, a constant pressure regime, flow rate and temperature must be chosen and set up. After measuring mole friction of component on permeate stream and permeate flow rate, the component permeability can be calculated. It is necessary to measure and calculate the gas permeabilities for interested gas every half an hour. As a last step, the graph test time versus permeance in Barrer should be plotted, trending the given line will give an average amount of permeability. As we noted earlier, the ratio of the permeability of two gases, such as i and j, is called the selectivity of membrane ij, and at the same time it can be determined by applying the formula.

3. Diffusion and sorption coefficients as factors that affect the permeability

It is emphasized that the process of gas separation by means of polymeric membranes arrives in direct dependence on gas concentration in polymeric matrix. As it was noted earlier, namely in section 2.1, gas permeability (P) arrives as a function of both diffusivity (D) and solubility (S) (Equation 3.1).

P = D * S (3.1)

where S represents the thermodynamic term, which characterizes the number of gas molecules that are sorbed in the polymer, D represents the kinetic term, which characterizes the mobility of gas molecules when they diffuse through the polymer. In other words, the permeability comes depending on the number of gas molecules that are dissolved in the polymer and, in addition, on the rate of their migration through the polymer matrix.

Solubility determination

For rubbery polymers, the solubility coefficient (S) equal to the Henry’s law sorption coefficient (K)(Equation 3.2)

S = K           (3.2)

where, K can be found using Flory-Huggins theory (Equation 3.3).

           (3.3)

where φ_p is the volume fraction of the polymer, V_R is the partial molar volume of the gas, V_S the molar volume of ideal gas at STP and χ is a Flory-Huggins interaction parameter between gas and polymer.

For glassy polymers, an additional gas sorption takes place. It can be described by the free volume between polymeric chains. The solubility equations changes to (Equation 3.4).

           (3.4)

where b is the Langmuir affinity constant, C_p the maximum free volume capacity. The equation was modelled through a Langmuir isotherm.

Diffusion coefficient determination

The diffusion coefficient can be determined experimentally. It is worth noting that the method is based on the construction of the sorption isotherm, and in addition and forward data interpretation (Equation 3.5) [37]. Sorption isotherms usually consist of huge number of measurements and it is difficult to get accurate results with such degree of complexity. The main purpose of the method in question is to determine the tangent slope of the first linear part of the sorption isotherm curve that allows to get relatively accurate calculations.

           (3.5)

where D is the diffusion coefficient (length²/time); r_s is the membrane thickness, b  represents th.e tangent slope of first linear part of sorption data, V_i is the gas content at the end of step I (mol/mass) and V_i-1 represents the gas content at the end of step I-1 (mol/mass).

As an example to establish the diffusion coefficient, a laboratory test was carried out for activated carbon and CO2 acting as a diffusion gas. The experimental setup used to plot the adsorption isotherm is based on Boyle's law. Figure 3.1 illustrates the process is shown.

 Figure 3.1. Laboratory module for adsorption isotherm determination.

It consists of two cells, a cylindrical sample cell in which mushed coal sample is placed and reference cell. The expansion valve connects two cells and the pressure and temperature is measured using transducer. The module is place in oven, so the temperature could be changed.

Methodology

1. After the sample is placed into pressure cell and desired pressure is set, the cells volume calibration stage should be done, using helium. This is due to the aspect that it is a rather small molecule and at the same time is not sorbed by carbon. Initially, the expansion valve is closed, so the reference pressure is measured. After opening the valve, expansion pressure recorded.
2. The feed is changed from helium to CD and pressure measurements are repeated. It should be mentioned that expansion pressure was measured every minute till the number is stabilized.
3. The p.2 was repeated for 3 different pressures.

After calculations are done, the adsorption isotherm was plotted. It should be emphasized that the results obtained were compared with reference experiments done by another research group [36]. A small difference could be explained by properties of sample materials. The isotherm is shown in Figure 3.2.

Figure 3.2 Measured adsorption and fitted reference measurements.

After adsorption isotherm is plotted, the is the diffusion coefficient (D) could be find. But firstly, the b coefficient should be found. The D was found to be equal to 4.34484 * 10^-8cm²/sec, which is comparable with similar research done by C.F. Rodrigues in 2016 [37].

The sensitivity of the sorption coefficient was demonstrated by Sina Nabati Shogle and Ahmadreza Raisia [38]. The most condensed gases had a higher level of gas sorption. The trend presented is due to the critical temperature acting as a measure of gas condensability. Consequently, gas sorption on the polymeric membrane increased as the critical gas temperature increased [38].

As various studies on the subject show, increasing the pressure has different effects on the sorption coefficient and, consequently, on the permeability for all kinds of polymeric membranes. In this regard, the CO₂ gas solubility coefficient and, in addition, the permeability decreases with increasing pressure for the glassy membrane. But on the other hand, the sorption diffusion coefficients increase simultaneously with increasing pressure for rubber membranes, which leads to an increase in permeability. At the same time for other penetrants gas solubility and, as a consequence, permeability significantly less arrives depending on the supply pressure.

Changing of temperature also affects membrane properties. Consequently, we can say that gas solubility decreases as temperature increases, but permeability herewith increases.

Such behavior is due to the aspect that during the gas permeation in polymeric membranes, the gaseous penetrants diffuse through free volumes in the polymeric matrix. In doing so, the thermal movements of the polymer chains in the amorphous regions form the free volumes in the membrane matrix through an arbitrary method. It is worth noting that the frequency and amplitude of the jumps of the mentioned chains increase directly with increasing temperature, resulting in an increase in the free volumes in the membrane. In this connection, diffusion of gas molecules through the membrane is facilitated at the highest operating temperatures, which, as a consequence, leads to an increase in the gas permeability of the membrane.

Conclusions

Membrane systems are quite suitable for separating of CO₂ and H₂S from the hydrocarbon mixture. It has been found that in the case of hydrogen sulfide separation, the cost of natural gas becomes somewhat less compared to amine scrubbing when treating natural gas with 1% hydrogen sulfide. It is worth noting that the polyimide membrane has many more advantages over other modern technologies used. This is due to the aspect that it is able to remove hydrogen sulfide and water in just one step. In addition, the membrane requires significantly less capital expenditure than traditional methods.

Continuing enhancements in MS technology is the most correct and natural choice for the future. It is emphasized that the proposed system can be used in a fairly wide range of operating pressures (0-69 bar), temperatures (1-98 C⁰) (water bath limit) and investigated gas concentrations. It gives an opportunity to investigate membrane properties in laboratory conditions and determine best option for industry application. Moreover, it becomes possible to make a sensitivity analyse of membrane separation’s cost from temperature, flow rate, pressure and permeate concentration.

Additionally, it becomes possible to improve current properties of exact polymer as well as test new materials. The synthesis of polymers with significantly higher selectivity’s and permeabilities for CO₂ and H₂S will provide an opportunity to significantly reduce the current cost of separation.

Recommendations and suggestions for future work

The applicable list of operating conditions can be improved to perform various tests in a wider range of parameters. For example, a water bath can be replaced by furnace. As a result of such a substitution, the range of test temperatures will significantly increase. This is due to the fact that membranes are more sensitive to temperature changes a programmable temperature control unit is desired.

Additionally, the created software could be improved by changing it from static operating system to dynamic. The changing will let to perform a completely automated long-term experiment which does not require operator interaction after the initial setup. This can be achieved by using control modules coupled with a rugged computer system.

As was mentioned earlier, a three-stage membrane module configuration was proven to be most effective. Therefore, it recommended to modify current system by adding two more pressure cells to perform the test in the more effective way.

References

1. International Energy Outlook 2016, Derived from Oxford Economic Model, June 2015. Available at:http://www.oxfordeconomics.com
2. Lallemand F., Rocher A., Aimard N., (2006). Sour gas production: Moving from conventional to advanced environmentally friendly schemes. SPE paper 103802,
3. McKetta J., (1988). Lactic Acid to Magnesium Supply-Demand Relationships, Taylor & Francis Encyclopedia of chemical processing and design, Volume 28.
4. Mork M. and Gudmundsson J., (2011). “Natural gas cleaning,” Natural Gas journal, 2011 3, p.4.4.
5. Mork M. and Gudmundsson J. (2011) Natural gas cleaning. Natural Gas, vol. 3, p. 4.4.
6. Islamiyah M., Soehartanto T., Hantoro R. and Abdurrahman A., (2015) Water scrubbing for removal of CO₂ (carbon dioxide) and H₂S (hydrogen sulphide) in biogas from manure, KnE Energy, vol. 2, no. 2, pp. 126–131.
7. Bondar V. I., Freeman B. D. and Pinnau I., (2000) Gas transport properties of poly(ether-b-amide) segmented block copolymers, International Journal of Polymer Science Journal of Polymer Science, Part B: Polymer Physics, vol. 38, pp. 2051–2062.
8. Robb W. L., Thin (1968) silicone membranes-their permeation properties and some applications, Annals of the New York Academy of Sciences, vol. 146, no. 1, pp. 119–137.
9. Bernardo P., Drioli E. and Golemme G., (2009), Membrane gas separation: a review/state of the art, Industrial and Engineering Chemistry Research, vol. 48, no. 10, pp. 4638–4663.
10. Kemmish D., (1995) High-Performance Engineering Plastics, Rapra Technology Limited.
11. Baker R., (2012) Membrane Technology and Applications, JohnWiley & Sons.
12. David O. C., Gorri D., Nijmeijer K., Ortiz I. and Urtiaga A., (2012) Hydrogen separation from multicomponent gas mixtures containing CO, N₂ and CO₂ using Matrimid asymmetric hollow fiber membranes, Journal of Membrane Science, vol. 419-420, pp. 49–56.
13. Huang Y. and Paul D. R., (2007), Effect of film thickness on the gas-permeation characteristics of glassy polymer membranes, Industrial and Engineering Chemistry Research, vol. 46, no. 8, pp. 2342–2347.
14. . Liu S. L, Wang R., Liu Y., Chng M. L. and Chung T. S., (2001) The physical and gas permeation properties of 6FDA-durene/2, diamino-toluene copolyimides, Polymer, vol. 42, no. 21, pp. 8847–8855.
15. Sanders D. F., Smith Z. P., Guo R. (2013) et al., Energy-efficient polymeric gas separation membranes for a sustainable future: a review, Polymer, vol. 54, no. 18, pp. 4729–4761.
16. Al-Megren H. (2012), Advances in Natural Gas Technology, InTech.
17. Thomas D. and Benson S., (2005), Carbon Dioxide Capture for Storage in Deep Geologic Formations—Results from the CO₂ Capture Project, Elsevier Science.
18. Merkel T.C., Gupta R.P., Turk B.S., Freeman B.D., (2001), Mixed-gas permeation of syngas components in poly (dimethylsiloxane) and poly (1-trimethylsilyl-1-propyne) at elevated temperatures, Journal of Membrane Science. 191 85–94.
19. Ghosal, K.; Freeman, B. D. (1994) Gas separation using polymer membranes: an overview, p.673–697.
20. Scholes C. and Ujjal K., (2017), Review of Membranes for Helium Separation and Purification .
21. Flory, P. (1953) Principles of Polymer Chemistry; Cornell University Press: Ithaca, NY, USA.
22. Petropoulos, J.H. (1994), Mechanisms and Theories for Sorption and Diffusion of Gases in Polymeric Gas Separation; CRC Press: Boca Raton, FL, USA, pp. 17–81.
23. Syed A., Durucan S., Ji-Quan Shi, Korre A., (2013)., Flue gas injection for CO₂ storage and enhanced coalbed methane recovery: Mixed gas sorption and swelling characteristics of coals,
24. Rodrigues C. F., Dinisa A. P. and Lemos M. J., (2016), Gas Diffusion Coefficient in Coal: Calculation of Tangent Slope Accuracy through the Inflection Point Determination, Journal of Mining Science, vol.52.
25. Nabati S., Ahmadreza R., Abdolreza A., (2017), Modeling of gas solubility and permeability in glassy and rubbery membranes using lattice fluid theory, Department of Chemical Engineering, Amirkabir University of Technology.
26. Hellums M., Koros W., Husk G., Paul D., (1991), Gas transport in halogen-containing aromatic polycarbonates, International Journal of Polymer Science 43.
27. Lin H., Freeman B.D., (2004), Gas solubility, diffusivity and permeability in poly (ethylene oxide), International Journal of Polymer Science105-117.
28. Barrie J.A., Williams, M., Munday K., (1980), Sorption and diffusion of hydrocarbon vapors in glassy polymers, Journal Polymer Engineering and Science. vol. 20, pp.20-29.