Preliminary Studies of CO2 Removal from Precombustion Syngas

Article pubs.acs.org/IECR

Preliminary Studies of CO2 Removal from Precombustion Syngas through Pressure Swing Membrane Absorption Process with Ionic Liquid as Absorbent Xingming Jie, John Chau, Gordana Obuskovic, and Kamalesh K. Sirkar* Otto H. York Department of Chemical, Biological and Pharmaceutical Engineering, Center for Membrane Technologies, New Jersey Institute of Technology, Newark, New Jersey 07102, United States ABSTRACT: Using 1-butyl-3-methylimidazolium dicyanamide ([bmim][DCA]) as the absorbent and either hydrophobized porous ceramic tubule or hydrophobized polyether ether ketone (PEEK)-based porous hollow fiber membranes, we have carried out preliminary studies for CO2 removal from simulated precombustion syngas by a pressure swing membrane absorption process. We have used He as a surrogate for H2 (40% CO2 rest He) up to a temperature of 100 °C and a pressure of 250 psig. A novel 5-valve system was designed to improve product quality. The optimal duration for the key absorption step was determined, and the influence of helium-rich product withdrawal time was investigated. Temperature increase will degrade product quality, so adding polyamidoamine dendrimer of generation 0 to [bmim][DCA] will mitigate the temperature effect, and 15 wt % dendrimer in [bmim][DCA] showed the best performance. Feed pressure increase will degrade helium product quality, while it is beneficial for CO2 product quality, since more CO2 will be absorbed with a higher feed pressure. When feed gas pressure was 250 psig at CO2 side ∼88.2% CO2-enriched gas product was produced. Using a simulated two-stage large PEEK module system, with a 14.0% CO2 balance helium gas mixture, helium product with CO2 concentration as low as 4.2−5.8% could be achieved at different temperatures. A brief analysis has been provided to explain why a PEEK membrane module shows much better performance than the ceramic tubule.

1. INTRODUCTION Continuous emission of large amount of greenhouse gases to the atmosphere has brought obvious changes, such as higher earth surface temperature and more frequent and severe climatic disturbances. Among all greenhouse gases, CO2 is believed to make the largest contribution of ∼80%.1 Scientists have confirmed that the atmospheric concentration of CO2 has increased globally by ∼100 ppm (36%) over the last 250 years mostly due to human activities.2 CO2 capture and storage (CCS) is widely thought to be the most important technique to deal with global climate change in the short-term.3,4 Usually for coal-based power plants, three fields widely investigated for CO2 removal are oxygen-enriched combustion, postcombustion, and precombustion.5 Oxygenenriched combustion has the advantage that CO2 concentration in flue gas can be as high as 90% and is very easily concentrated further; however, it faces the difficulty of oxygen-enriched gas production and high energy cost.6 For postcombustion, due to the low flue gas pressure and much lower CO2 concentration, the capture of CO2 is difficult and requires high capital investment and high energy cost.7 Compared to this, precombustion CO2 capture from postshift reactor synthesis gas obtained via integrated gasification combined cycle for coal as a solid fuel is of significant interest since CO2 is present at a much higher partial pressure. From this point of view, CO2 removal from precombustion syngas can play a useful role in global CCS. However, the temperature of this high-pressure gas is also very high, which is around 150−200 °C+. This high temperature reduces CO2 solubility in various solvents which may be used for absorption-based separation, and further it imposes demanding conditions on any © 2013 American Chemical Society

membrane material as well as module potting materials if membranes are to be used. At present, solvent absorption is still the most successful and widely applied method for CO2 removal. For a precombustion process, low-temperature (L-T) water−gas shift reactor product stream is likely to be available at ∼20 atm and around 150−200 °C. If CO2 from this stream is to be absorbed in a liquid absorbent, the liquid absorbent must be thermally stable and essentially nonvolatile. Further it must be highly CO2 selective over H2 and any other impurities present (such as CO, etc.). Ionic liquids (ILs) have high thermal stability and essentially no vapor pressure. Yokozeki et al. have tested CO2 solubilities in different ILs and demonstrated solubility selectivity of CO2 over H2 of as much as 30−300 at H2 partial pressures of 0.5−3 MPa at around room and lower temperatures for the IL of 1butyl-3-methylimidazolium hexafluorophosphate ([bmim][PF6]).8,9 Solubility selectivity tests for CO2 over other gases also have been carried out by other researchers with different ILs.10−12 Aki et al. investigated the high-pressure phase behavior of carbon dioxide with imidazolium-based ILs and found that solubility is strongly dependent on the choice of the anion.13 Task-specific ILs (TSILs) having functional groups which can form complexes with CO2 have been synthesized and used as facilitated supported liquid membrane (FSLM). Goodrich et al. Special Issue: Giulio Sarti Festschrift Received: Revised: Accepted: Published: 8783

August 7, 2012 February 25, 2013 February 28, 2013 March 14, 2013 dx.doi.org/10.1021/ie302122s | Ind. Eng. Chem. Res. 2013, 52, 8783−8799

Industrial & Engineering Chemistry Research

Article

Table 1. Dimensional Characteristics of the Membrane Absorption Modules modulea

OD, cm

ID, cm

Lb, cm

pore size, Å

VVF

fiber no.

surface areab, cm2

ceramic PEEK-S PEEK-Lc

0.57 0.0452 0.0452

0.37 0.0290 0.0290

44.0 34.3 41.0

∼50 ∼20 ∼20

0.35−0.4 ∼0.4 ∼0.4

1 240 568

78.75 1168 3420

OD and ID are the outer and inner diameters, respectively; L is the effective fiber length; and VVF is the void volume fraction. bBased on outer diameter of fibers. cPEEK-L module has a packing density around 21.8% that was defined as the ratio between the total fiber volume and the real volume they occupied (total fiber volume plus space among fibers). a

porous tubules/fibers providing per unit device volume a very large surface area of nondispersive contact between the postshift reactor synthesis gas stream flowing through the fiber lumen and the liquid absorbent present as a thin stagnant absorbent liquid layer in between the microporous tubules/ hollow fibers on the shell side of the separation device. The principal objective of the work is to develop an advanced PSAB-based process to produce purified hydrogen at high pressure and simultaneously obtain a highly purified CO2 stream containing at least 90% of the CO2 that will be suitable for subsequent sequestration. In this study with porous ceramic tubules (can endure high pressure and temperature) or PEEK (organic material, easy to be processed with inorganic-like robust characteristics) hollow fiber membrane modules and IL [bmim][DCA] with or without PAMAM dendrimer of generation 0 as absorbent, a novel pressure swing membrane absorption (PSMAB) process was designed for separating a He−CO2 mixture. The feed gas mixture contained He as a surrogate for H2. Both are quite small molecules not widely apart from our process perspective; further both are inert for our system. In addition for preliminary explorations, He is ideal since we can avoid any leakage-related problems at higher temperatures and pressures. Based on the determination of optimal step duration time in each cycle, the effects of heliumrich product withdrawal quantity, feed pressure and test temperature were systematically investigated. A brief dimensional comparison between two types of membrane modules was also carried out to explain their considerable performance difference.

synthesized a series of six amine-functionalized anion-tethered ILs with tetraalkylphosphonium cations and investigated their suitability as potential absorbents for CO2 capture from postcombustion flue gas.14 Meindersma et al. reviewed the application of task-specific ILs for intensified separation and suggested that 1-butyl-3-methylimidazolium dicyanamide ([bmim][DCA]) could be a good choice as CO2 absorption solvent.15 The CO2/H2 selectivity achieved by such a FSLM reached up to 10−20 at a temperature of ∼85 °C.16 The major advantage of such ILs is that they can operate at a high temperature and separate in the absence of water, which is not possible in highly CO2-selective polymeric membranes containing amines in cross-linked poly(vinyl alcohol) and studied in the temperature range of 100−170 °C.17 A similar study indicated a precipitous drop in CO2/H2 selectivity of the polymeric membrane containing amine moieties as the feed gas moisture content decreased.18 A major weakness of the supported liquid membranes (SLMs) using ILs is that their overall CO2/H2 selectivity is low. It is known that a membrane’s selectivity is determined by the product of the solubility selectivity and diffusivity selectivity. Although their solubility selectivity for CO2 over H2 is quite high, the diffusivity selectivity favors H2 or He over CO2 due to their much smaller size. Therefore unless reversible chemical complexation is introduced, the overall SLM selectivity will be low. Unlike the permeation-based behavior in a separation membrane, an absorption-based process is unlikely to suffer from such a low CO2−H2 selectivity since solubility selectivity will be controlling. Actually the postshift reactor gas stream for coal-based gasification plants can also be purified to recover hydrogen with high purity via the pressure swing adsorption (PSA) process: H2 is purified at near ambient temperature and high pressure using adsorbents, such as molecular sieves, and silica gel to remove impurities, such as CO2.19 The desorption stream produces purified CO2. However there is a significant loss of H2 as much as 10% in such regeneration processes along with substantial thermal/cooling needs.20 Ionic liquids are very viscous. Ionic liquids containing polyamidoamine (PAMAM) dendrimers of generation 021 are even more viscous. Conventional absorption processes where such liquids are mobile face considerable pressure drops. This problem can be avoided if we adopt the novel concept of PSAB originally proposed in Bhaumik et al.22 They carried out a lower-temperature lower-pressure analog of such a concept (rapid pressure swing absorption, RAPSAB). A related technique was pursued in Obuskovic et al.23 This process combines the specific advantages of a number of basic separation techniques: selective absorption of CO2 in a nonvolatile liquid/oligomeric absorbent at temperatures and pressures characteristic of the feed stream under consideration; pressure swing absorption (PSAB) process simulating a PSA process (which however uses adsorbent particles); and hollow micro-

2. EXPERIMENTAL PROCEDURE 2.1. Materials. Ionic liquids, [bmim][DCA] and [emim][Tf2N], were purchased from EMD Chemicals (Philadelphia, PA) and used as received for membrane module breakthrough pressure tests. [bmim][DCA] was selected as absorbent because of its excellent CO2 absorption, and further, it is stable in the presence of moisture at a high temperature. It is also highly miscible with the PAMAM dendrimer, which will enhance the absorption process and improve the final separation results. Polyethylene glycol 400 (PEG-400) was purchased from Chemicals Direct (Roswell, GA) and used as received for membrane module breakthrough pressure tests. PAMAM dendrimer (generation 0) was purchased from Dendritech (Midland, MI). It was received as a 62.35 wt % dendrimer solution in methanol. To get pure dendrimer, the solution was vacuumed for several days under relatively high temperature (∼60 °C) to remove the methanol. Simulated precombustion syngas gas containing a 40.67% CO2/helium balance as a surrogate for H2 was purchased from Air Gas (Salem, NH). Ceramic membrane modules were obtained from Media and Process Technology (Philadelphia, PA). One module contains a single ceramic tubule in one stainless steel housing. 8784

dx.doi.org/10.1021/ie302122s | Ind. Eng. Chem. Res. 2013, 52, 8783−8799


Article

Figure 1. (a) Gas−liquid membrane contacting in a membrane contactor. (b) Membrane module of ceramic tubules or hollow fibers. (c) Schematic diagram of PSMAB apparatus.

The surface of the ceramic tube was completely hydrophobized by silane coating. During all tests a solid 1/8 in. diameter Teflon rod was inserted into the tube side to reduce tube side volume. Therefore in every cycle much less feed gas will enter the tube side to be treated, expecting better separation results. Two types of polyether ether ketone (PEEK) membrane modules were obtained from Porogen (Woburn, MA). One type is a small PEEK module (identified as PEEK-S) with straight fibers sealed inside. The other type (identified as PEEK-L) is exactly the same type but has much longer fibers in

the stainless steel housing. The main difference of PEEK-L from PEEK-S is that all the fibers are helically wound. Further the fibers are tightly bunched together as if they are in a strap. Details of all membrane modules are listed in Table 1. 2.2. Breakthrough Pressure Test for Membrane Modules. Breakthrough pressure test will determine how high a feed gas pressure may be used for PSMAB. It is determined mainly by two factors: pore size of membrane at the liquid−membrane interface on the shell side and the surface tension of the IL. During the test, one port of the membrane 8785



Article

Figure 2. Schematic diagrams of (a) 3- and (b) 5-valve PSMAB processes.

pneumatic valves were used to control exactly the time period for different steps in one absorption cycle (3-valve system). This valve control system was realized via a Programmable Logic Controller (PLC) scheme installed by PneuMagnetic (Quakertown, PA). A similar arrangement was also developed for a 5-valve system described later. Both helium- and CO2-rich product streams were analyzed by a Quantek Model 906 CO2 Analyzer (Grafton, MA) that will allow recording real time CO2 concentration fluctuations in different product gas streams. A pressure transducer installed inside the oven and directly connected to the tube side of membrane module reveals detailed pressure changes during gas absorption. This allowed a better understanding of the PSMAB process. For example, from the pressure drop in the absorption step, we could estimate how much gas was absorbed and find out which absorbent was better. Usually the PSMAB system will be able to generate stable separation results after 5−10 cycles, since during this period a stable pressure change procedure could be built. All data reported in this paper were taken after at least 30 cycles running. Also all tests were repeated three times to confirm their accuracy. 2.4. 3- and 5-Valve Systems. In a typical PSMAB process22 usually a 3-valve control system is applied, as shown in Figure 2a. There will be four steps in each absorption cycle. (1) Valve 1 is open, and valves 2 and 3 are closed. Feed gas was introduced into tube side of the membrane module for a certain time to develop the desired pressure (a sharp pressure increase in tube side).

module shell side was connected to a small cylinder containing [bmim][DCA] (the other port kept closed), and the module shell side was filled with IL. The IL cylinder was also connected to a N2 cylinder to develop the desired pressure. The membrane module tube side had a low N2 flow to bring any possible breakthrough of IL out when the pressure was gradually increased. The leaked IL will fall on a piece of paper next to the outlet. When leaked IL could be detected from tube side, the test pressure was defined as the breakthrough pressure. Liquids other than ILs were also tested to explore the capabilities of the membranes as well as compare ILs against other potential liquids. 2.3. PSMAB Process. Figure 1a shows a microporous ceramic tubule or PEEK hollow fiber. In the separator device, there will be many such tubules or hollow fibers, as in Figure 1b. Surrounding the tubule/fiber is the absorbent (e.g., IL) filling the shell side of the separator. The pores in the wall of the ceramic tubule/hollow fiber are gas-filled. In the test apparatus (Figure 1c), the membrane contactor module was put inside a PV-222 oven (Espec Corp, Hudsonville, MI) so that the exact temperature could be set and controlled. The shell side of the module was filled with a certain absorbent, such as IL, supplied from the absorbent container connected to a N2 cylinder to maintain the desired pressure. Feed gas was introduced into the membrane tube side where it contacted the absorbent through the micropores to be absorbed. The absorbent pressure in the shell side was kept ∼20 psig higher than the feed gas pressure in tube side to avoid any possible gas bubbling. CO2 product side was connected to a vacuum pump to supply driving force for product withdrawal. Usually three 8786



Article

membrane module, the feed pressure can go up to 300 psig. For PEEK-S membrane modules, the highest test pressure was 140 psig. Some minor liquid leakage could be found when liquid on the shell side was pressurized to 160 psig. However PEEK-L module could easily withhold liquids up to 300 psig. When we provide the symbol >300 psig, it means no leakage up to 300 psig, but tests at higher pressure were not done. 3.2. Performance Comparison between 3- and 5-Valve Systems. As described in Section 2.4, the advantage of the 5-valve system is that we can have a similar helium-rich product and an improved CO2-rich product at the same time. Experimental data listed in Table 3 clearly reveal the advantages

(2) All 3 valves closed, and absorption between feed gas in tube side and absorbent in shell side takes place mainly at the interface of micropores during this period (pressure in tube side will decrease gradually in this step due to gas absorption). (3) Valves 1 and 2 are closed, and valve 3 is open for certain time to withdraw helium-rich product (sharp pressure decrease in tube side because of product withdrawal). (4) Valves 1 and 3 are closed, and valve 2 is open for certain time to withdraw CO2-rich product (sharp pressure decrease in tube side because of product withdrawal). In the absorption process described in Figure 2a, feed gas in the tube side was finally separated into two parts: helium- and CO2-rich products. So the 3-valve system will have to face a trade-off situation. If a better quality helium-rich product is preferred (only take a very limited amount of gas in the tube side as the helium-rich product), then a low-quality CO2-rich product will have to be accepted and vice versa, if we want a better CO2-rich product, then we should take gas in the tube side as much as possible as the helium-rich product which will definitely generate poor quality helium product. To overcome this trade-off problem, we have designed a new 5-valve system as shown in Figure 2b. The main difference between the two arrangements is as follows: With a 5-valve system, after the gas absorption step we divide the gas mixture in the tube side into three parts: heliumrich product, middle part gas, and CO2-rich product. In other words, an extra step (middle part gas withdrawal step) was added between steps 3 and 4 of the 3-valve system. We have two options for the middle part gas: it could be directly released and collected for later treatment or recycled into the system before the feed gas-introducing step. In the latter case, there will be 6 steps in each cycle for a 5-valve system as shown in Figure 2b. In the new 5-valve system, the withdrawal of the middle part gas, containing still a significantly higher He-concentration gas, will leave much less He in the tube side as the CO2-rich product is desorbed at the end of the cycle; thus the CO2-rich product quality will be definitely improved. To enhance the helium-rich product, the recycled middle part gas, which will have a lower CO2 concentration than the feed gas, will be introduced to the tube side of the feed gas location at the beginning of next cycle. This will lower the effective CO2 level of the feed gas for the next cycle so that the quality of the helium-rich product could be significantly improved.

Table 3. Comparison between the Performances of 3- and 5Valve Systems feed pressure, CO2 pressure in tube He psig system cycle time,a s side,b psig productc productc 100 100 100

ceramic PEEK-S PEEK-L

[emim][Tf2N], psig

PEG 400, psig

300 >200 >300

>300 160 >300

180 80 N/A

300 140 N/A

70.10% 38.40% 70.95%

In 5-valve system, the middle part gas will be collected without recycling. bPressures in tube side at the end of each step. cAll are in terms of CO2 concentration.

of the 5-valve system. With one PEEK-S membrane module and pure [bmim][DCA], tests were carried out at room temperature with a feed pressure of 100 psig. As expected the newly designed 5-valve system showed a better performance. With feed gas pressure at 100 psig, using the 5-valve system we can have a helium-rich product with a CO2 concentration of 8.31% and a CO2-rich product with CO2 concentration of 70.10%. If using a 3-valve system, a CO2 concentration of 8.00% in helium side will bring a product with a CO2 concentration as low as 38.40% in the CO2 side. On the other hand, if a CO2-rich product with a CO2 concentration of 70.95% was achieved, a helium-rich product with CO2 concentration as high as 30.90% will result on the other side. 3.3. Determination of Optimal Absorption Duration for PSMAB Cycle. In a typical PSMAB process, usually there are four essential steps described as feed in, absorption, heliumrich product withdrawal, and CO2-product withdrawal. During our tests, for all applied various membrane module combinations, feed in duration was determined to be long enough to establish the desired feed pressure in tube side. Usually in order to make sure that most of the CO2 absorbed in IL will be desorbed, CO2-product withdrawal will be carried out from 30 to 60 s with different module combinations. Now we focus on the most important absorption step. It will directly determine how long the feed gas will be in contact with the IL in the shell side and undergo gas absorption. Next we will show how to determine optimal absorption step duration for a system. To find out the optimal absorption time, at first we set the absorption time in 1 cycle as long as 900 s to examine the pressure drop caused by gas absorption during this step. 3.3.1. Optimal Absorption Step Duration Time for Ceramic Module System. Here a combination of three ceramic membrane modules in series mode was employed, and pure IL [bmim][DCA] was used as the absorbent. The absolute pressure change in the tube side tests during the long time absorption step was carried out at different temperatures with a fixed feed pressure of 150 psig.

Table 2. Breakthrough Pressure Test Results [bmim][DCA], psig

5; 30; 1; 4; 95.0; 75.0; 0; 8.31% 30 −14.7 3-valve 5; 30; 1; 30 95.0; 75.0; −14.7 8.00% 3-valve 5; 30; 30; 30 95.0; 0; −14.7 30.90%

a

3. RESULTS AND DISCUSSION 3.1. Breakthrough Pressure. All breakthrough pressure test results for both ceramic and PEEK membrane modules with different liquids at room temperature are listed in Table 2.

water, psig

5-valve

Four different liquids were used for the testing, while only [bmim][DCA] was used as the absorbent. As seen from Table 2, when [bmim][DCA] was used as a solvent for the ceramic 8787



Article

Figure 3. Pressure drops in the tube side (a) and pressure drop percentage (b) during extended time absorption step in three ceramic membrane modules connected in series.

system. We will have to find a balance between absorption time and treatment capacity. From Figure 3b it is seen that at 120 s of absorption step time, i.e., ∼13.3% of total 900 s duration, the system has achieved about 55−65% CO 2 absorption for different temperatures. Thus the optimal absorption step duration for this system was chosen as 120 s. Further higher temperatures lead to a faster absorption at the beginning.

As shown in Figure 3a, during the long duration absorption step, pressure in the tube side gradually decreased due to absorption of CO2 and helium into the shell-side absorbent. From the pressure drop during 900 s absorption time, we can determine the optimal duration time for absorption step. It is obvious that longer absorption time will bring more CO2 absorption and therefore a better quality product. But a too long absorption step will reduce the treatment capacity of the 8788



Article

Figure 4. Pressure drops in the tube side (a) and pressure drop percentage (b) during extended time absorption step in two PEEK-S membrane modules in series.

3.3.2. Optimal Absorption Step Duration for PEEK-S Module System. Similar to the tests described above, two PEEK-S membrane modules were connected in series. The feed pressure was kept at 100 psig, and the absorption behavior of 900 s for the PEEK-S membrane module system was studied at different temperatures. As shown in Figure 4a, the extended time absorption behavior for the PEEK membrane module system was quite different from the ceramic module system. Actually with the hollow fiber-based PEEK-S modules, the absorption of gas was much faster. Most of the absorption was completed within

100 s. If we set the optimal absorption step duration time to 30 s, around 3.33% of the total extended absorption time, we see from Figure 4b that 55−65% of total absorption has taken place. Based on these results, for two PEEK modules in the series mode system, its optimal absorption step duration time could be fixed at 30 s shorter than 120 s for the ceramic system. Due to the very high surface area per unit volume of the PEEK system, much more rapid absorption takes place into the surrounding liquids compared to that in the ceramic tubule system. In PSMAB process, rapid initial absorption is important. After this initial period, there is slow diffusion and 8789



Article

Figure 5. Pressure change in tube side and product composition for different helium product release amounts for two PEEK-S hollow fiber modules in series.

Figure 6. Pressure in the tube side and product changes with different helium product release quantity for three ceramic modules in series.

absorption into the large volume of the surrounding liquid, as we see in Figures 3 and 4. 3.4. Influence of the Quantity of Helium-Rich Product Withdrawal. For PSMAB process, the key issue is to increase the contacting opportunity between the gas and the absorbent. From this point of view, a longer and thinner hollow fiber will be favorable. Therefore we will assume at the end of the absorption step, gas composition in the tube side is more likely to have a concentration distribution along the fiber length.

Since the gas at the module end for helium-rich product withdrawal passed through the whole fiber and had more contacting opportunity with the absorbent in shell side, its CO2 concentration should be much lower than the other end. Thus it is necessary to investigate the influence of helium withdrawal quantity on its product quality. The 5-valve system was applied without recycling the middle part gas. Helium-rich product withdrawal quantity was controlled by varying this step duration. 8790



Article

3.4.1. PEEK-S System. For tests described below, two PEEKS membrane modules were put together in series, and pure [bmim][DCA] used as absorbent. All tests were carried out at room temperature with a feed pressure of 140 psig. The optimal absorption step time employed was 30 s. Figure 5 presents pressure changes in tube side within one complete cycle. It started from a negative pressure and then reached 140 psig after the feed gas introduction. At the end of 30 s absorption time, pressure in the tube side dropped to 133 psig. Next by adjusting the He-rich product withdrawal time, different pressure drops were observed. Next middle part gas release left the tube side at one atmospheric pressure, and last step CO2-rich product withdrawal brought the tube side to a negative pressure again. Different helium-rich product withdrawal durations will lead to different pressure drops in tube side. Longer duration will bring more pressure drop that means more gas in the tube side will be taken out as a helium-rich product, and thus product quality will be influenced. As shown in Figure 5, starting from 133 psig, when pressure at the end of helium withdrawal step decreased to 100 psig (1 s), we had a helium product with CO2 concentration as low as 5.5%. While pressure was reduced to 78 psig (2 s), the CO2 concentration in helium-rich product increased to 7.7%; 65 psig pressure generated 10.0% concentration (3 s); 47 psig brought a 13.5% concentration (5 s); and 30 psig lead to the poorest product with CO2 concentration up to 17.5% (10 s). In terms of CO2-rich product, since it was withdrawn from the same atmospheric pressure in the tube side, its quality was stable maintaining a CO2 concentration at 77.9−78.4%. Based on these results, there should be a trade-off between the helium-rich product quantity and its quality. We need to find a balance between these two factors by determining a proper helium-rich product withdrawal duration time in a practical application. 3.4.2. Ceramic System. Compared with PEEK-S hollow fibers, the ceramic module has a tubule with a much larger diameter. In order to find out if this will have any impact, similar tests were carried out with three ceramic membrane modules in series and the feed pressure was 100 psig. As determined earlier the optimal absorption time was set as 120 s in each cycle. As shown in Figure 6, unlike the PEEK-S module test results, He-rich product withdrawal quantity showed little influence on its product quality in the case of ceramic modules. For the first 3 different pressure drops during helium-rich product withdrawal [from 95 to 64 psig (1 s), 44 psig (2 s), and 21 psig (4 s)], we achieved almost the same helium-rich product with CO2 concentration around 33.1−33.8%. Only when the pressure in the tube side dropped to atmospheric (6 s) CO 2 concentration in helium-rich product increased obviously to 37.4%, since we almost took all gas in the tube side out as helium-rich product. This meant that compared with the PEEK-S system, because of its much larger tubule diameter, even when we had a Teflon rod to reduce the feed gas volume, there was still too much feed gas in the tube side waiting to be treated. Only a very small part of the feed gas could be satisfactorily absorbed due to a limited contacting area. Therefore for the ceramic system, it is more likely that the gas concentration distribution along the tube length was not fully developed as in the PEEK modules. In terms of CO2-rich product, its quality was stable maintaining a CO2 concentration at 47.5−48.4%. It can be clearly seen that PEEK-S generated much better absorption behavior and gas products than the ceramic system,

comparing results from Figures 5 and 6. This was in accordance with our expectation in Section 3.3.2. 3.5. Influence of Test Temperature and Dendrimer Addition to IL on PSMAB Performance with Ceramic System. Precombustion syngas is generated in a state of high temperature; thus it is important to investigate PSMAB process performance at higher temperatures. Also as reported earlier,24−26 amine groups of polyamidoamine (PAMAM) dendrimer can react with CO2; therefore the addition of dendrimer to solvent [bmim][DCA] is expected to enhance its CO2 solubility performance.21 To investigate the influence of dendrimer on the PSMAB process, especially at a high test temperature, a set of tests with three ceramic modules in series was carried out with pure [bmim][DCA] and different dendrimer−IL mixtures as absorbents. The feed pressure was kept at 150 psig, and different test temperatures were used. For all tests, the 5-valve system was used with middle part gas recycled. Figure 7a illustrates the pressure change in one complete cycle running for 188 s. It started from around 8 psig, which was from middle part gas recycle (usually for 3-valve system the starting pressure in tube side should be negative). Within 2 s of feed gas introduction duration, the pressure in the tube side reached 150 psig and then experienced 120 s of absorption and dropped differently with different absorbents in the shell side and running at different temperatures. Continued helium-rich product withdrawal decreased the pressure in the tube side to around 40 psig within 2 s. Next middle part gas recycle reduced the pressure in the tube side to 13 psig further within another 2 s. The CO2-rich product withdrawal driven by a vacuum pump left the tube side with a negative pressure after 60 s duration. Finally the middle part gas recycle increased the pressure in the tube side from negative to 8 psig within 2 s. Figure 7b−e presents detailed pressure drops during the absorption step when different liquid mixtures were used as absorbents and when running at different temperatures. When pure IL was used as absorbent, we did not see obvious pressure drop differences at different temperatures. When dendrimer was mixed with IL and acted as absorbents, the test temperature had considerable influence as seen in Figure 7c−e. At room temperature, pressure drops formed in absorption step were very small for all used liquid mixtures. Increasing the temperature shows a positive effect and increases the pressure drop, meaning that the gas solubility has been facilitated. This can be explained as follows: Adding dendrimer to IL will increase liquid viscosity greatly, and viscous liquid does not favor rapid gas solubility. When the temperature is increased, the liquid mixture viscosity decreases; this is why a larger pressure drop during absorption step could be seen with dendrimer−IL mixtures. What should be mentioned also is that when 30 and 45 wt % dendrimer−IL mixtures were used as absorbents, the gas solubility was totally controlled by viscosity since increasing temperature will increase the pressure drop during the absorption step at the same time. In the case of 15 wt % dendrimer−IL mixture, from room temperature to 75 °C, the gas solubility was viscosity controlled, while at 100 °C a lower pressure drop could be seen, meaning that now the gas solubility was temperature controlled. Helium- and CO2-rich products generated from different absorbents at different temperatures can be seen in Figure 8. When pure IL was applied as the absorbent, temperature shows obvious negative influence. Even we did not see an obvious pressure drop difference during the absorption step. The composition of gas absorbed was altered since the temperature 8791



Article

increase has different effects on CO2 and helium. The CO2 concentration in the helium-rich product increased from 33.6% at room temperature to 36.1% at 100 °C, while at the same time, the CO2-rich product suffered a concentration drop from 51.7% at room temperature to 47.6% at 100 °C. When

dendrimer was added to the IL and used as an absorbent, unlike pure IL, the temperature increase shows a positive effect. Increasing temperature decreased the CO2 concentration in the helium-rich product and increased the CO2 concentration in the CO2-rich product under most conditions. Among all tested

Figure 7. continued 8792



Article

As mentioned earlier, the ceramic module has a limited effective area and a large amount of feed gas in the tube side; therefore we did not see great improvements when the dendrimer−IL mixture was applied as the absorbent at high temperatures. The suitability of ceramic modules for use at high

absorbents, 15 wt % dendrimer−IL mixture shows the best performance at 100 °C. It yielded a helium-rich product with CO2 concentration of 34.1% and a CO2-rich product with CO2 concentration of 50.6%. Both were much better than pure IL generated products at 100 °C.

Figure 7. continued 8793



Article

Figure 7. Pressure change in the tube side during one complete cycle (a) and pressure drops within absorption step for different absorbents: (b) pure IL; (c) 15 wt % dendrimer in IL; (d) 30 wt % dendrimer in IL; and (e) 45 wt % dendrimer in IL at different test temperatures for three ceramic modules in series.

liquid surrounding the fibers unlike that in PEEK-S in which fibers were spaced apart with considerable liquid in between. After optimal absorption step duration was determined, a set of PSMAB process tests with the PEEK-L module were carried out with different feed pressures at room temperature. The 5valve system was applied with the middle part gas recycled. Pressure changes in the tube side for different feed pressures are presented in Figure 10. As we can see for all tested feed pressures from 100 to 250 psig, after the absorption step, most gas in tube side was taken out as a helium-rich product. We already know if too much gas in the tube side was withdrawn as helium product, its quality will be degraded. This arrangement is favorable for CO2-rich products since it will be withdrawn from a lower pressure in the tube side. Figure 10 also clearly shows how the feed pressure will influence pressure drop within the absorption step. Higher pressure will bring increased gas solubility leading to larger pressure drops in the tube side. When the feed gas pressure was 100 psig, the pressure drop within absorption step was 5.80 psig. It increased to 15.68 psig when the feed gas pressure was maintained at 250 psig. Influence of feed gas pressure on qualities of the two product streams is presented in Figure 11. An increase in feed gas pressure means more gas will be introduced into the membrane tube side and contacted with IL to be absorbed. Higher pressure means higher CO2 partial pressure that will lead to a higher absorption, resulting in a larger pressure drop. More gas absorbed in the IL means we will have a better CO2-rich product. On the other hand, as we mentioned before, the PEEK-L module may have the problem of insufficient amounts of surrounding absorbent, thus higher feed pressure will definitely bring a larger burden on the limited absorbent.

temperature enables us to carry out the tests listed above, and the results are encouraging enough for us to learn how a dendrimer can facilitate the whole PSMAB process and improve the products. This will be very helpful for further investigation. 3.6. Influence of Feed Pressure on PSMAB Performance with PEEK-L System. PEEK-L module has shown excellent pressure withstanding capability when breakthrough pressures were measured at room temperature compared to PEEK-S due to tube sheet leakage. In this section, based on PEEK-L membrane module, the influence of feed pressure on PSMAB process will be investigated at room temperature with pure IL as absorbent. As before, for a brand new membrane module, we determined its optimal absorption duration first. Thus a 900 s duration absorption process was repeated with the PEEK-L module. Test results are shown in Figure 9. From Figure 9b we know that more than 55% of the total absorption was complete within the first 60 s; thus the optimal absorption step duration will be 60 s for the PEEK-L system. If we compare pressure drop curves of Figure 9a with Figures 3a and 4a, one can conclude that the pressure drop rate for the PEEK-L module was larger than that of the ceramic system but less than that of the PEEK-S system. We already know that the PEEK hollow fiber is better than the ceramic tubule, but since PEEK-L and PEEK-S are using exactly the same fibers, why do they have different pressure drop rates? As we mentioned in Section 2.1, the main difference between PEEK-L and PEEK-S is that all the fibers are helically wound and tightly bunched together in PEEK-L. So the reason for a lower pressure drop rate of PEEK-L is that there was reduced 8794



Article

Figure 8. Product changes from different absorbents at different test temperatures for three ceramic modules in series: (a) helium- and (b) CO2-rich products.

CO2 concentration in CO2-rich products increased from 84.4% to 88.2%. The DOE target for CO2 concentration in the CO2-rich stream is >95%. During our studies on various aspects of the PSMAB cycle, the maximum CO2 concentration in the CO2rich product obtained has been increased to what we mentioned earlier, namely, 88.2%. There is significant room

This is not favorable for the helium-rich product. Results in Figure 11 are in good accordance with our analysis: Higher feed pressure will generate a better CO2-rich product while degrading the helium-rich product at the same time. When feed gas pressure increased from 100 psig to 250 psig, CO2 concentration in the helium-rich product increased from 16.5% to 21.5%, meaning quality degradation. At the same time the 8795



Article

highly CO2-rich gas with much less CO2-rich gas in various dead volumes in the system. Specifically the tube-side header section of the membrane module and the potted section of the module at the feed end and related connections introduce much less CO2-rich gas into the CO2-rich product. If we increase the gas volume located in the hollow fiber tube side which is being treated vis-à-vis the dead volume regions at the feed end, we expect a significant enhancement in the final product purity at the feed end. The modules we have used are short and have relatively a small volume of gas in the tube side of the active region of the hollow fibers. Our further efforts to improve CO2-rich product quality include reducing dead volume from both membrane module itself and all related connections, using dendrimer-containing IL absorbents to increase solubility selectivity of CO2 over He and PSMAB process design optimization being carried out now. The results of these efforts will be reported later. 3.7. Simulated Two-Stage Membrane Absorption Process. According to test results of Section 3.6, to generate a good quality CO2-rich product we took most of the gas in tube side out as helium-rich product; thus we had difficulty in achieving a high-quality helium product. Restricted by high pressure and CO2 concentration in the feed gas, we may have difficulty in achieving both high-quality helium- and CO2-rich products by just an one stage absorption process. One possibility to improve helium-rich product quality is to have a second PSMAB process treatment using the helium-rich product from first stage as the feed gas that we call a simulated two-stage membrane absorption process. Knowing helium-rich product quality could be controlled by its withdrawal quantity and assuming that we now have a helium-rich product with 14.0% CO2 concentration and 100 psig pressure from the first stage, using PEEK-L module and 3valve system, pure IL was applied as an absorbent, and a set of tests was carried out at different temperatures. The pressure changes in tube side are shown in Figure 12. Within 5 s of the feed gas introduction duration, pressure in the tube side

Figure 9. Pressure drops in tube side (a) and pressure drop percentage (b) during extended time absorption step in one PEEK-L membrane module.

for improvement in this value. One of the primary sources of CO2 concentration reduction is mixing of the purified desorbed

Figure 10. Pressure change in tube side during one complete cycle at different feed pressures for PEEK-L system at room temperature. 8796



Article

and at the same time the CO2 concentration in CO2-rich products decreased from 20.6% to 17.1% (this gas could also be recycled). Now if we combine the test results from Sections 3.6 and 3.7, we have a possible arrangement to achieve both high-quality helium and CO2 products. A 2-stage membrane absorption system will be designed as follows: The first stage will generate a large amount of helium-rich product with a relatively high CO2 concentration at higher pressure. A high-quality CO2-rich product with CO2 concentration not lower than 90% will also be generated at the same time. The second stage will use the helium-rich product from first stage as feed gas, and a high-quality helium-rich product with CO2 concentration not higher than 5.0% will be generated. The CO2-rich product generated from this stage usually has a lower CO2 concentration and can be recycled. Results of further efforts, such as running the PSMAB process with PEEK-L system at high pressure and temperature combining different liquid mixtures as absorbent, process redesign and optimization to improve product qualities are being carried out and will be reported later. 3.8. Comparison of PEEK and Ceramic Membrane Module Systems. We have seen that there are considerable differences in performance between the PEEK membrane modules and the ceramic membrane module systems. Obviously the performance of PEEK system is much better. In order to find out the reason, a detailed comparison between them is provided in Table 4. In the PSMAB process, absorbent capability and membrane module contacting area per unit feed gas volume will be the two main factors that will determine the final separation achieved for a given cycle configuration. According to dimensional parameters listed in Table 4, in terms of the ratio between the effective contacting area and the corresponding feed gas

Figure 11. Product changes from different feed pressures for the PEEK-L system at room temperature.

increased from negative to 100 psig. Then with next 60 s absorption duration, a certain pressure drop will happen for different temperatures (higher temperature usually decrease gas solubility, thus less pressure drop will be seen at higher temperature, as shown in Figure 12). A continued 20 s heliumrich product withdrawal duration allows us to take most of the gas in tube side out; now pressure in tube side was around 10 psig. The final CO2-rich product withdrawal duration of 60 s will give the tube side a negative pressure again. Detailed information about product qualities is shown in Figure 13. As revealed by pressure drop differences from different temperatures, temperature increase will definitely degrade the product quality in both sides. When temperature increased from room temperature to 100 °C, CO2 concentration in helium-rich products increased from 4.2% to 5.8%;

Figure 12. Pressure change in the tube side during one complete cycle at different test temperatures for PEEK-L system at 100 psig with a 14% CO2 containing feed gas. 8797



Article

Figure 13. Product changes from different feed temperatures for PEEK-L system at 100 psig with a 14% CO2 containing feed gas.

Table 4. Estimated Dimensional Calculations for PEEK Hollow Fiber Module and Ceramic Tubule Membrane-Based Modules module

OD, cm

ID, cm

L, cm

VVF

(V, cm3)a

(A, cm2)b

(A/V, cm−1)c

ceramic ceramicd PEEK

0.57 0.57 0.0452

0.37 0.37 0.0290

44.0 44.0 34.3

∼0.4 ∼0.4 ∼0.4

7.33 3.84 0.0356

31.5 31.5 1.95

4.30 8.19 54.7

Feed gas volume in one fiber. bEffective contacting area for one fiber based on the outer diameter. cRatio between A and V. dA Teflon rod was inserted into the tube to reduce the volume.

a

designed to improve gas product qualities. An absorption experiment of extended duration was designed to determine the optimal absorption step duration. An increase in the amount of helium-rich product withdrawal will degrade its quality; thus, the product quality could be controlled by varying the quantity of helium-rich stream withdrawal. A test temperature increase will degrade product qualities when pure IL is used as an absorbent. Adding dendrimer in IL could greatly mitigate temperature influence especially at higher temperatures. Feed pressure increase will degrade helium-rich product quality, while CO2-rich product quality will be improved. A simulated two-stage membrane absorption system and attempted tests showed it will help to achieve a much better helium-rich product. Dimensional calculations revealed that PEEK fibers have a much larger ratio of contacting area per unit feed gas volume leading to a much better PSMAB performance. We have demonstrated the feasibility of the novel 5-valve PSMAB system for precombustion syngas CO2 removal process. Additional efforts to enhance the absorption process and improve the product quality that can meet the environmental and application requirements are being carried out and will be reported later. The present membrane contactor results are probably the most demanding performance in terms of temperature and pressure for any gas−liquid membrane contactor in open literature.

volume, for the ceramic module (with a Teflon rod in the tube side), it is only 8.19 cm−1. It is much lower than 54.71 cm−1 for a PEEK hollow fiber. This could directly explain why the PEEK module showed much higher absorption rates than the ceramic membrane module. This also explains why PEEK modules have much better PSMAB performance than the ceramic modules. In other words, if ceramic modules with similar dimensions as PEEK fibers could be successfully prepared, much better separation results and promising application potential for this kind of module that show stable separation results under high temperature could be expected. An ideal membrane contactor should in addition have high breakthrough pressure, appropriate surface hydrophobicity, low pore transport resistance, and a significant gap between neighboring fibers so that there is sufficient liquid for absorption without sacrificing a high surface area per unit volume (see Section 3.6). 3.9. Additional Considerations. Syngas has a few impurities, such as H2S, CO etc. It is useful to enquire whether there are any downsides to using ILs in the presence of H2S. We have not located any literature indicating adverse chemical reactions of H2S with the IL under consideration. Literature data on solubility of H2S in a variety of ILs are available.27−29 In case there are any adverse reactions with trace H2S amounts, the PSMAB device allows easy discharge and replenishment of the IL through two valves on the shell side as and when necessary.

■

4. CONCLUDING REMARKS A novel PSMAB process has been developed for absorption treatment of precombustion syngas, and a 5-valve system was

AUTHOR INFORMATION

Corresponding Author

*Tel: 973-596-8447. Fax: 973-642-4854. E-mail: sirkar@njit. edu. 8798



Article

Notes

separation of CO2 by suing novel facilitated transport membrane at elevated temperatures and pressures. J. Membr. Sci. 2007, 291, 157. (19) Yang, Hongqun; Xu, Zhenghe; Fan, Maohong; Gupta, Rajender; Slimane, Racid B; Bland, Alan E; Wright, Ian Progress in carbon dioxide separation and capture: A review. J. Environ. Sci. 2008, 10, 14. (20) Harrison, Douglas P. Sorption-enhanced hydrogen production: a review. Ind. Eng. Chem. Res. 2008, 47, 6486. (21) Chau, J.; Obuskovic, G.; Jie, X M; , and Sirkar, K. K.; . Solubilities of CO2 and He in ionic liquid containing Poly (amido amine) dendrimer Gen 0. Ind. Eng. Chem. Res., Accepted for publication. (22) Bhaumik, S.; Majumdar, S.; Sirkar, K. K. Hollow-fiber membrane-based rapid pressure swing absorption. AIChE J. 1996, 42, 409. (23) Obuskovic, G.; Poddar, T. K.; Sirkar, K. K. Flow swing membrane absorption-permeation. Ind. Eng. Chem. Res. 1998, 37, 212. (24) Kovvali, A. S.; Chen, H.; Sirkar, K. K. Dendrimer Membranes: A CO2-Selective Molecular Gate. J. Am. Chem. Soc. 2000, 122, 7594. (25) Kovvali, A. S.; Sirkar, K. K. Dendrimer Liquid Membranes: CO2 Separation from Gas Mixtures. Ind. Eng. Chem. Res. 2001, 40, 2502. (26) Kosaraju, P.; Kovvali, A. S.; Korikov, A.; Sirkar, K. K. Hollow Fiber Membrane Contactor Based CO2 Absorption-Stripping Using Novel Solvents and Membranes. Ind. Eng. Chem. Res. 2005, 44, 1250. (27) Pomelli, C. S.; Chiappe, C.; Vidis, A.; Laurenczy, G.; Dyson, P. J. Influence of the Interaction between Hydrogen Sulfide and Ionic Liquids on Solubility: Experimental and Theoretical Investigation. J. Phys. Chem. B 2007, 111, 13014. (28) Jalili, A. H.; Rostami, M. R.; Ghotbi, C.; Jenab, M. H.; Ahmadi, A. N. Solubility of H2S in Ionic Liquids [bmim][PF6], [bmim][BF4], and [bmim][Tf2N]. J. Chem. Eng. Data 2009, 54, 1844. (29) Shiflett, M. B.; Niehaus, A. M. S.; Yokozeki, A. Separation of CO2 and H2S Using Room-Temperature Ionic Liquid [bmim][MeSO4]. J. Chem. Eng. Data 2010, 55, 4785.

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This material is based upon work supported by the Department of Energy, National Energy Technology Laboratory under award no. DE-FE0001323.

■

REFERENCES

(1) Lashof, Daniel A.; Ahuja, Dilip R. Relative contributions of greenhouse gas emissions to global warming. Nature 1990, 334, 529. (2) Changes in atmospheric constituents and radiative forcing. Intergovernmental Panel on Climate Change’s Fourth Assessment Report; IPCC: Geneva, Switzerland, 2007; Chapter 2, p 135. (3) Azar, Christian; Lindgren, Kristian; Larson, Eric; Mollersten, Kenneth Carbon capture and storage from fossil fuels and biomass − costs and potential role in stabilizing the atmosphere. Clim. Change 2006, 74, 47. (4) Rubin, Edward S.; Chen, Chao; Rao, Anand B. Cost and performance of fossil fuel power plants with CO2 capture and storage. Energy Policy 2007, 35, 4444. (5) Kanniche, Mohamed; Gros-Bonnivard, Rene; Jaud, Philippe; Valle-Marcos, Jose; Amann, Jean-Marc; Bouallou, Chakib Precombustion, post-combustion and oxy-combustion in thermal power plant for CO2 capture. Appl. Therm. Eng. 2010, 30, 53. (6) Tan, Yewen; Douglas, Mark A.; Thambimuthu, Kelly V. CO2 capture using oxygen enhanced combustion strategies for natural gas power plants. Fuel 2002, 81, 1007. (7) Merkel, Tim C.; Lin, Haiqing; Wei, Xiaotong; Baker, Richard Power plant post-combustion carbon dioxide capture: An opportunity for membranes. J. Membr. Sci. 2010, 359, 126. (8) Shiflett, Mark B.; Yokozeki, A. Solubilities and diffusivities of carbon dioxide in ionic liquids: [bmim][PF6] and [bmim][BF4]. Ind. Eng. Chem. Res. 2005, 44, 4453. (9) Yokozeki, A.; Mark, B. Shiflett. Hydrogen purification using room-temperature ionic liquids. Appl. Energy 2007, 84, 351. (10) Yuan, Xiaoliang; Zhang, Suojiang; Chen, Yuhuan; Lu, Xingmei; Dai, Wenbin; Mori, Ryohei Solubilities of gases in 1,1,3,3Tetramethylguanidium Lactate at Elevated Pressures. J. Chem. Eng. Data 2006, 51, 645. (11) Yuan, Xiaoliang; Zhang, Suojiang; Liu, Jun; Lu, Xingmei Solubilities of CO2 in hydroxyl ammonium ionic liquids at elevated pressures. Fluid Phase Equilib. 2007, 257, 195. (12) Raeissi, Sona; Peters, Cor J. A potential ionic liquid for CO2separating gas membranes: selection and gas solubility studies. Green Chem. 2009, 11, 185. (13) Aki, Sudhir N. V. K.; Mellein, Berlyn R.; Saurer, Eric M.; Brennecke, Joan F. High −pressure phase behavior of carbon dioxide with imidazolium-based ionic liquids. J. Phys. Chem. B 2004, 108, 20355. (14) Goodrich, Brett F.; de la Fuente, Juan C.; Gurkan, Burcu E.; Zadigian, David J.; Price, Eric A.; Huang, Yong; Brennecke, Joan F. Experimental measurements of amine-functionalized anion-tethered ionic liquids with carbon dioxide. Ind. Eng. Chem. Res. 2011, 50, 111. (15) Wytze Meindersma, G.; Galan Sanchez, Lara M.; Hansmeier, Antje R.; de. Haan, Andre B. Application of task-specific ionic liquids for intensified separations. Monatsh. Chem. 2007, 138, 1125. (16) Myers, Christina; Pennline, Henry; Luebke, David; IIconich, Jeffery; Dixon, JaNeiIIe K.; Maginn, Edward J.; Brennecke, Joan F. High temperature separation of carbon dioxide/hydrogen mixtures using facilitated supported ionic liquid membranes. J. Membr. Sci. 2008, 322, 28. (17) Zou, Jian; Winston Ho, W. S. CO2-selective polymeric membranes containing amines in crosslinked poly(vinyl alcohol). J. Membr. Sci. 2006, 286, 310. (18) Yegani, R.; Hirozawa, H.; Teramoto, M.; Himei, H.; Okada, O.; Takigawa, T.; Ohmura, N.; Matsumiya, N.; Matsuyama, H. Selective 8799


Preliminary Studies of CO2 Removal from Precombustion Syngas

Recommend Documents