Energy Fuels 2010, 24, 6214–6219 Published on Web 10/25/2010
: DOI:10.1021/ef101123e
CO2-philic Oligomers as Novel Solvents for CO2 Absorption Matthew B. Miller,*,†,‡ David R. Luebke,† and Robert M. Enick†,‡ †
National Energy Technology Laboratory, Pittsburgh, Pennsylvania 15236, United States, and ‡Department of Chemical Engineering, University of Pittsburgh, Pittsburgh, Pennsylvania 15261, United States Received August 20, 2010. Revised Manuscript Received October 7, 2010
Desirable properties for an oligomeric CO2-capture solvent in an integrated gasification combined cycle (IGCC) plant include high selectivity for CO2 over H2 and water, low viscosity, low vapor pressure, low cost, and minimal environmental, health, and safety impacts. The neat solvent viscosity and solubility of CO2, measured via bubble-point loci and presented on a pressure-composition diagram (weight basis), and water miscibility in CO2-philic solvents have been determined and compared to results obtained with Selexol, a commercial oligomeric CO2 solvent. The solvents tested include polyethyleneglycol dimethylether (PEGDME), polypropyleneglycol dimethylether (PPGDME), polypropyleneglycol diacetate (PPGDAc), polybutyleneglycol diacetate (PBGDAc), polytetramethyleneetherglycol diacetate (PTMEGDAc), glyceryl triacetate (GTA), polydimethyl siloxane (PDMS), and perfluorpolyether (PFPE) that has a perfluorinated propyleneglycol monomer unit. Overall, PDMS and PPGDME are the best oligomeric solvents tested and exhibit properties that make them very promising alternatives for the selective absorption of CO2 from a mixed gas stream, especially if the absorption of water is undesirable.
but in the IGCC process, the syngas is used to make power. After the syngas is produced, CO is converted to CO2 and H2 via the water-gas shift reaction, which yields a stream rich in CO2 and H2 and also contains water. The precombustion CO2-capture step would be placed downstream of the watergas shift reactor. The process stream exiting the water-gas shift stage and entering the CO2-removal stage is made up of approximately 38-45% CO2 and 50-55% H2.3 Once selective CO2 capture is complete, the H2- and water-rich balance of the fuel gas could then be used to produce power via expansion through a combustion turbine. Leaving the water with the hydrogen will result in a greater mass flow rate in the fuel gas stream, generating additional power in the gas turbine, and will also reduce energy penalties associated with driving water out of the solvent during regeneration. Therefore, the solvent for this separation should selectively extract only CO2, absorb large amounts of CO2, and have a low viscosity that minimizes pumping costs and mass-transfer resistances. The objective of this study is to assess the viability of several oligomeric, CO2-philic CO2-capture solvents and to compare their CO2 absorption, hydrophobicity, and viscosity to the commercial oligomeric CO2 solvent used in the Selexol process. The identification of a CO2 solvent with superior properties to the Selexol solvent (high CO2 absorption, hydrophobicity, and low viscosity) has the potential to lead to greater IGCC power plant efficiency. This study does not include comparisons of other properties that are very relevant to a large-scale absorption process, including the determination of solvent vapor pressure, corrosion rates with common materials of construction, or rates of solvent thermal degradation. The solvents selected for this study were chosen based on prior evidence of their ability to completely dissolve in
Introduction Concern about the rise in atmospheric CO2 concentration and its potential implications to global climate change has led to diverse research efforts related to CO2 capture and subsequent long-term geological sequestration. There are a variety of opportunities available for CO2 capture from large industrial point sources, including natural gas processing and cleanup (PCO2 ∼ 15-300 psi), ethylene oxide synthesis (PCO2 ∼ 50 psi), steam reforming of methane (PCO2 ∼ 70 psi), ammonia synthesis (PCO2 ∼ 100 psi), and power production based on integrated gasification combined cycles (IGCC) (PCO2 ∼ 30-175 psi).1 Processes, such as these and others with high partial pressures of CO2 in flue gas streams, would be best suited for a physical absorption capture process rather than chemical absorption because the high CO2 partial pressure driving force of physical absorption is inherently present. Technologies that yield effluent streams with low partial pressures of CO2 would have to rely on chemical absorption of CO2, such as that accomplished with an aqueous solution of monoethanol amine. IGCC power plants represent a superior option for CO2 capture because of their high CO2 partial pressure synthesis gas (syngas) streams and increased thermal efficiency2 compared to typical pulverized coal (PC) plants. The fuel source, which is usually coal but can include other fossil fuels or alternative carbon sources, including biomass or waste materials, is gasified to produce syngas, which consists primarily of hydrogen and carbon monoxide. From this stream, a multitude of chemicals can be synthesized, including liquid fuels, *To whom correspondence should be addressed. E-mail: mbm35@ pitt.edu. (1) International Energy Agency (IEA) Greenhouse Gas R& D Programme (IEAGHG). Opportunities for Early Application of CO2 Sequestration Technologies; IEAGHG: Cheltenham, U.K., 2002. (2) Descamps, C.; Bouallou, C.; Kanniche, M. Energy 2008, 33, 874– 881. r 2010 American Chemical Society
(3) Department of Energy (DOE)/National Energy Technology Laboratory (NETL). Cost and Performance Baseline for Fossil Energy Plant; DOE/NETL: Pittsburgh, PA, 2007.
6214
pubs.acs.org/EF
Energy Fuels 2010, 24, 6214–6219
: DOI:10.1021/ef101123e
Miller et al.
CO24-8
of each solvent as a viable CO2 solvent. Ideally, when compared to Selexol, the candidate would absorb more CO2, absorb less water, have a lower viscosity, exert a lower vapor pressure, and have the capability of being less costly when manufactured in large volumes.
compressed liquid or supercritical or by recent evidence from our laboratories that indicated that very small oligomers (e.g., di- and trimers) could solubilize significant amounts of gaseous CO2.9 Glycerol triacetate (GTA) was also considered because its structure is analogous to a trimer of poly(vinyl acetate), the most CO2-philic high-molecularweight hydrocarbon-based polymer identified to date.10 Acetate groups have also been identified as a CO2-philic functional group, where peracetylated mono- and disaccharides were dissolved in CO211,12 and silicone polymers functionalized with acetate groups were dissolved in CO2.13,14 Other functional groups identified as CO2-philic in nature include the carbonyl group (CdO),15-17 the oxygen atoms present in ether groups (R-O-R), and fluorine when prevalent throughout a solvent.11,12,18-22 The addition of these groups adds sites where Lewis acid-Lewis base interactions between the solvent and CO2 can occur,13,16,23 thus improving CO2 solubility. All solvents chosen contain these groups populated throughout the monomer, as well as the end groups, in an effort to enhance their CO2-philicity. All of the polyether oligomers were terminated with methyl or acetate groups as opposed to hydroxyl groups, which are known to be CO2phobic moieties that reduce the solubility of CO2 in a solvent. Ideally, the oligomeric solvent should have a high affinity for CO2 that promotes high solubility of CO2, low viscosity to minimize pumping requirements and mass-transfer limitations, low affinity for water to prevent the co-absorption of CO2 and water, low vapor pressure to reduce evaporative losses, low cost, and minimal environmental, health, and safety impacts. These criteria are used to judge the potential
Experimental Section Materials. Polypropyleneglycol dimethylether (PPGDME; Mn, 400; average repeat unit (RU), 6.00; Mn/Mw, 1.08) and polyethyleneglycol dimethylether (PEGDME; Mn, 310; RU, 6.00; Mn/Mw, 1.12) were purchased from Polymer Source and used as received. Polydimethyl siloxane (PDMS; Mn, 550; RU, 6.24) was purchased from Gelest, Inc. and used as received. Polypropyleneglycol diacetate (PPGDAc; Mn, 509; RU, 6.7) and polybutyleneglycol diacetate with a linear (-C4H8O-) monomer unit, also known as polytetramethyleneetherglycol diacetate (PTMEGDAc; Mn, 250; RU, 3.2) or poly(1-4-butanediol)diacetate, were synthesized by Bayer Material Science using PPG and PTMEG as the starting materials and transforming the terminal groups from hydroxyls to acetates. Polybutylenegylcol (PBG) with a branched monomer [-CH(C2H5) CH2O-], received from Huntsman International LLC, was transformed into PBGDAc (Mn, 250; RU, 3.2) by Bayer Material Science. Glyceryl triacetate (GTA) g99.0% was purchased from Fluka and used as received; this compound was considered to be analogous to a trimer of poly(vinyl acetate), the most CO2-philic high-molecular-weight hydrocarbon-based polymer identified to date.10 The Selexol solvent produced by Dow and purchased from Univar was used as received. The structure of each solvent used is shown in Table 1. CO2 was purchased from Penn Oxygen and Supply Company with a purity of 99.99% and used without further purification. Solubility of CO2 in the Solvent; Bubble-Point Pressures. Phase behavior measurements were carried out at 25 and 40 °C in a windowed, agitated, variable volume view cell (Schlumberger) using standard non-sampling techniques, also known as the synthetic method, described in detail elsewhere.9,12,24,25 In a typical experiment, known amounts of solvent and CO2 were introduced to the high-pressure cell and then compressed until a single, transparent phase was attained. The sample volume was then slowly expanded until a bubble point was observed. The overall CO2 mass concentration in the cell at this point represents the amount of CO2 that is absorbed in the oligomer at that (bubble-point) pressure. Bubble-point pressure values are collected over a range of concentrations and can be used to generate a bubble-point locus. Although this apparatus is well-suited to determine the solubility of CO2 in a solvent, it cannot be used to determine the relative amounts of CO2 and H2 that dissolve in the solvent if a CO2-H2 gas mixture is introduced to the cell along with the solvent, yielding vapor-liquid equilibrium. Therefore, a high-pressure, windowed, agitated, variable-volume, phase equilibrium apparatus with liquid recirculation and vaporand liquid-phase sampling capabilities is currently being developed at the National Energy Technology Laboratory of the Department of Energy. In the near future, this new apparatus will be used to determine the CO2/H2 selectivity of Selexol and the most promising solvents identified during this study. Neat Solvent Viscosity Measurements. Neat solvent viscosity measurements were carried out for all oligomers using a Brookfield LVDV-IIþ Pro viscometer, and cone spindle CPE40 was used for all measurements, capable of measuring viscosities of 0.15-3065 cP. A Brookfield TC-602 temperature controller with ethylene glycol as the heating/cooling fluid was used
(4) Pereira, F. E.; Keskes, E.; Adjiman, C. S.; Galindo, A.; Jackson, G. American Institute of Chemical Engineers (AIChE) Annual Meeting; Philadelphia, PA, Nov, 2008. (5) Murrieta-Guevara, F.; Romero-Martinez, A.; Trejo, A. Fluid Phase Equilib. 1988, 44, 105–115. (6) Mutelet, F.; Vitu, S.; Privat, R.; Jaubert, J. Fluid Phase Equilib. 2005, 238, 157–168. (7) Isaacs, E.; Otto, F. D.; Mather, A. E. J. Chem. Eng. Data 1980, 25, 118–120. (8) Chakravarty, T.; Phukan, U. K.; Weiland, R. H. Chem. Eng. Prog. 1985, April, 32. (9) Miller, M. B.; Chen, D.; Xie, H.; Luebke, D. R.; Johnson, J. K.; Enick, R. M. Fluid Phase Equilib. 2009, 287, 26–32. (10) Wang, Y.; Hong, L.; Tapriyal, D.; Kim, I. C.; Paik, I.-H.; Crosthwaite, J. M.; Hamilton, A. D.; Thies, M. C.; Beckman, E. J.; Enick, R. M.; Johnson, J. K. J. Phys. Chem. B 2009, 113, 14971–14980. (11) Rindfleisch, F; DiNoia, T. P.; McHugh, M. A. J. Phys. Chem. 1996, 100, 15581–15587. (12) Shen, Z.; HcHugh, M. A.; Xu, J.; Belardi, J.; Kilic, S.; Mesiano, A.; Bane, S.; Karnikas, C.; Beckman, E. J.; Enick, R. M. Polymer 2003, 44, 1491–1498. (13) Kilic, S.; Michalik, S.; Wang, Y.; Johnson, J. K.; Enick, R. M.; Beckman, E. J. Ind. Eng. Chem. Res. 2003, 42, 6415–6424. (14) Fink, R.; Hancu, D.; Valentine, R.; Beckman, E. J. J. Phys. Chem. B 1999, 103, 6441–6444. (15) Nelson, M. R.; Borkman, R. F. J. Phys. Chem. A 1998, 102, 7860– 7863. (16) Kazarian, S. G.; Vincent, M. F.; Bright, F. V.; Liotta, C. L.; Eckert, C. A. J. Am. Chem. Soc. 1996, 118, 1729–1736. (17) Raveendran, P.; Wallen, S. L. J. Am. Chem. Soc. 2002, 124, 12590–12599. (18) Ji, M.; Chen, X.; Wai, C. M.; Fulton, J. L. J. Am. Chem. Soc. 1999, 121, 2631–2632. (19) Desimone, J. M.; Guan, Z.; Elsbernd, C. S. Science 1992, 257, 945–947. (20) Laintz, K. E.; Wai, C. M.; Yonker, C. R.; Smith, R. D. J. J. Supercrit. Fluids 1991, 4, 194–198. (21) Raveendran, P.; Wallen, S. L. J. Phys. Chem. B 2003, 107, 1473– 1477. (22) Johnston, K. P.; Harrison, K. L.; Clarke, M. J.; Howdle, S. M.; Heitz, M. P.; Bright, F. V.; Carlier, C.; Randolph, T. W. Science 1996, 271, 624–626. (23) Beckman, E. J. Chem. Commun. 2004, 17, 1885.
(24) Enick, R. M; Beckman, E.; Yazdi, A.; Krukonis, V.; Schonemann, H.; Howell, J. J. Supercrit. Fluids 1998, 13, 121. (25) Potluri, V. K.; Xu, J.; Enick, R. M.; Beckman, E. J.; Hamilton, A. D. Org. Lett. 2002, 4, 2333.
6215
Energy Fuels 2010, 24, 6214–6219
: DOI:10.1021/ef101123e
Miller et al.
Table 1. Structures of CO2-philic Oligomers and Glyceryl Triacetate Used in the Absorption of CO2
bubble-point locus is bounded at the 100% solvent end of the Px diagram by the vapor pressure of the solvent. These pressures are too low to be measured precisely using our equipment. The perfluoropolyether (PFPE; n, 5.0) results shown in Figure 1 are reproduced from our previous study.9 Figure 1 illustrates that the solubilities of CO2 in PEGDME, PPGDME, PDMS, PTMEGDAc, PBGDAc, and GTA are comparable. The greatest difference in the bubble-point loci for any of these systems at any composition is only ∼60 psi for the range of weight percent shown. The Selexol solvent required similar pressures as compared to these solvents to dissolve the same amount of CO2 on a weight basis. The PPGDAc and PFPE solvents yielded the lowest CO2 solubility on a weight basis, requiring roughly 50-100 psi more than PPGDME, PEGDME, PDMS, PTMEGDAc, PBGDAc and GTA to dissolve the same amount of CO2. The poor PFPE performance is attributable in part to the results being presented on a weight fraction of interest to engineering design, which favors oligomers with low-molecular-weight monomers. The bubble-point loci results at 40 °C are illustrated in Figure 2. Because of the poor performance at 25 °C and great expense of the PFPE solvent, it was not assessed at 40 °C. As expected, the increase in the temperature decreases the solubility of CO2 in all of the solvents; therefore, the bubble-point pressures have shifted to higher values for any particular weight percent of the solvent. GTA appears to be the best solvent at this temperature, although it must be kept in mind that this is clearly the most volatile of the solvents being considered (bp 180 °C). The oligomeric solvents, Selexol, PPGDME, PEGDME, PDMS, PBGDAc, and PTMEGDAc, give very
to maintain the temperature of the sample cup during each experiment. Each absorbent was tested at three separate shear rates and two different temperatures: our ambient laboratory temperature of 22 and 40 °C. In each experiment, the viscometer was zeroed and 0.5 mL of the solvent, measured using a Fisherbrand Finnpipette to ensure precision and repeatability, was placed in the sample cup. Then, the sample cup was attached to the viscometer, and the gap between the cup and cone was set using the micrometer adjustment ring. Shear rates were then chosen across the range of the viscometer in an effort to capture any non-Newtonian behavior. Solubility of Water in Neat Solvent; Cloud Point Measurements. The solubility of water in each solvent was also assessed at 22 and 40 °C by slowly adding water to the solvent until a cloud point, indicative of the presence of a small aqueous phase in equilibrium with the solvent, was observed. In a typical experiment, 10 g of solvent was placed in a beaker and distilled water was added to the solvent dropwise until a cloud point was observed. The temperature of the beaker was controlled with a water bath and heating plate, and mixing was ensured with a magnetic stir bar and stir plate.
Results and Discussion Solubility of CO2 in the Solvent. The bubble-point loci for the pseudo-binary systems of PPGDME-CO2, PEGDMECO2, PPGDAc-CO2, PTMEGDAc-CO2, PBGDAcCO2, Selexol-CO2, and PDMS-CO2 and the binary system GTA-CO2 are reported on the solvent-rich end (60-100 wt % oligomer) at 25 °C (Figure 1) and 40 °C (Figure 2). Each bubble point was reproduced 6 times, which resulted in a (10 psi measurement variability. This degree of uncertainty is reflected by the size of the data markers in the figures. Each 6216
Energy Fuels 2010, 24, 6214–6219
: DOI:10.1021/ef101123e
Miller et al.
Figure 1. Phase behavior bubble-point loci for all solvents tested at 25 °C.
Figure 2. Phase behavior bubble-point loci for all solvents tested at 40 °C.
with the exception of the viscosity of GTA decreasing to be slightly less than the viscosity of PBGDAc. The effect of small amounts of dissolved water on solvent viscosity has been measured for several solvents. For example, 1 wt % water was added to Selexol and PBGDAc (PDMS is immiscible with water; therefore, it was not selected). The PBGDAc viscosity value decreased from 17 to 15 cP, while the change in the Selexol viscosity of 6.9 cP was not detectable. Further, upon absorption of CO2 at elevated pressures, the solvent-CO2 solution viscosity will diminish, as shown in the literature in polymer-CO2 mixtures.26-29 For example,
similar CO2 solubility results over the composition range. PPGDAc again exhibited the lowest CO2 solubility as presented. Viscosity of the Neat Solvent. Viscosity is an important property in solvent selection because low viscosity will reduce pumping costs and mass-transfer resistances during CO2 absorption. Mass-transfer coefficients are currently being measured and will be published at a later date. The viscosity of each solvent was measured at two temperatures, 22 and 40 °C, at various shear rates. The data are illustrated in Figures 3 and 4, respectively. At each temperature, the PDMS is substantially better than all other solvents tested, with a viscosity of less than 3 cP at both temperatures. The solvents with the next lowest viscosity of approximately 6 cP include PPGDME and Selexol. PBGDAc, GTA, PTMEGDAc, and PPGDAc had higher viscosity values of about 17, 18, 20, and 30 cP, respectively. When the temperature was increased to 40 °C, the viscosities all decreased as expected but their ranking was maintained,
(26) Gerhardt, L. J.; Manke, C. W.; Gulari, E. J. Polym. Sci., Part B: Polym. Phys. 1997, 35, 523–534. (27) Flichy, N. M. B.; Lawrence, C. J.; Kazarian, S. G. Ind. Eng. Chem. Res. 2003, 42, 6310–6319. (28) Royer, J. R.; Gay, Y. J.; Adam, M.; DeSimone, J. M.; Khan, S. A. Polymer 2002, 43, 2375–2383. (29) Dimitrov, K.; Lubomir, B.; Tufeu, R. Macromol. Chem. Phys. 1999, 200, 1626–1629.
6217
Energy Fuels 2010, 24, 6214–6219
: DOI:10.1021/ef101123e
Miller et al.
Figure 3. Viscosity of each solvent at 25 °C.
Figure 4. Viscosity of each solvent at 40 °C.
was roughly 2 wt %. GTA absorbed about 5 wt % water. The ranking of the oligomers from the greatest to least capable of absorbing water is as follows: Selexol ∼ PEGDME . GTA>PPGDME ∼ PPGDAc ∼ PBGDAc ∼ PTMEGDAc . PDMS.
Table 2. Bench-Top Water Miscibility in Solvent Test Results at 22 and 40 °C miscibility (g of H2O/g of solvent) solvent
22 °C
40 °C
Selexol PEGDME PPGDME PPGDAc PTMEGDAc PBGDAc PDMS GTA
fully miscible fully miscible 0.021 0.022 0.017 0.018 immiscible 0.046
fully miscible fully miscible 0.028 0.025 0.024 0.031 immiscible 0.062
Conclusion Oligomeric solvents, including PPGDME, PEGDME, PPGDAc, PTMEGDAc, PBGDAc, PDMS, and GTA, have been examined as potential CO2 absorbents to be implemented in physical absorption processes. With the exception of PEGDME, these solvents have not been previously assessed as CO2 solvents for the IGCC process. In this work, they have been judged on the basis of three important properties of a physical solvent used for CO2 capture: solubility of CO2 in the solvent over a wide range of pressures, solubility of water in the solvent, and neat solvent viscosity. Phase behavior plots on the solvent-rich end have been constructed, illustrating that the ability of the Selexol solvent to absorb CO2 is comparable to that of PEGDME, PDMS, PPGDME, PBGDAc, PTMEGDAc, and GTA at 40 °C. PPGDAc exhibited much lower solubility than these solvents.
the viscosity of PDMS saturated with CO2 at 80 °C and 1261 psi (8.8 wt % CO2) was 50% lower than that of neat PDMS at low shear rates.26 Solubility of Water in the Solvent. The solubility of water in the solvents at 22 and 40 °C is shown in Table 2. PEGDME and Selexol are fully miscible with water at all concentrations. The solubility of water in the other solvents was dramatically lower. PDMS is completely immiscible with water. The solubility of water in PPGDME, PPGDAc, and PTMEGDAc 6218
Energy Fuels 2010, 24, 6214–6219
: DOI:10.1021/ef101123e
Miller et al.
Selexol or PEGDME, PDMS exhibited comparable CO2 absorption, lower viscosity, and complete immiscibility with water. Relative to Selexol or PEGDME, PPGDME exhibited comparable CO2 absorption, comparable viscosity, and dramatically greater hydrophobicity (only 2-3 wt % water dissolves in PPGDME at 22-40 °C).
Viscosity testing showed that PDMS and PPGDME have lower viscosities than PEGDME and Selexol, which may lead to reduced mass-transfer resistances and pumping requirements if these solvents are employed. The remaining solvents, PPGDAc, PTMEGDAc, PBGDAc, and GTA, had viscosity values greater than that of Selexol. The ability of the solvents to absorb water differed dramatically at 22 and 40 °C. Selexol and PEGDME are fully miscible with water at all proportions. PPGDME, PPGDAc, PBGDAc, and PTMEGDAc absorb only 1-3 wt % water. GTA absorbed 4-6 wt % water. PDMS is completely immiscible with water. These properties suggest that PDMS and PPGDME may be very viable CO2 solvents for the IGCC process. Relative to
Acknowledgment. The authors thank Bayer Material Science and GE Global for assistance with synthesis work assisted by them and Huntsman International LLC for supplying samples of PBG used as a starting material. This work was performed in support of the ongoing research in the area of carbon management by the National Energy Technology Laboratory under the RDS contract DE-AC26-04NT41817.
6219
