Hydrophobic Polymeric Solvents for the Selective Absorption of CO2 from Warm Gas Streams that also Contain H2 and H2O

Hydrophobic Polymeric Solvents for the Selective Absorption of CO2from Warm Gas Streams that also Contain H2 and H2OR. M. Enick,†,‡,* P. Koronaios,‡ C. Stevenson,‡ S. Warman,‡ B. Morsi,†,‡ H. Nulwala,†,§ and D. Luebke∥

†NETL RUA Faculty Fellow, National Energy Technology Laboratory 626 Cochrans Mill Road P.O. Box 10940, Pittsburgh,Pennsylvania 15236-0940, United States‡University of Pittsburgh, Department of Chemical and Petroleum Engineering, 1249 Benedum Hall, 3700 O’Hara Street, Pittsburgh,Pennsylvania 15261, United States§Carnegie Mellon University, Department of Chemistry, 4400 Fifth Avenue, Pittsburgh Pennsylvania 15213, United States∥US DOE NETL, 626 Cochrans Mill Road P.O. Box 10940, Pittsburgh, Pennsylvania 15236-0940, United States

ABSTRACT: The hydrophobic polymers polydimethyl siloxane (PDMS) and polypropyleneglycol dimethylether (PPGDME)may provide an alternative to physical solvents based on the hydrophilic polymer polyethyleneglycol dimethylether (PEGDME)for the precombustion capture of CO2 from the warm, high pressure stream that also contains H2O and H2. PPGDME can bemade with a linear repeat unit (PPGDMEl, poly(1,3-propanediol) dimethylether) or a branched repeat unit (PPGDMEb,poly(1,2-propanediol) dimethylether). The solubility of CO2 and H2 in each of the four solvents of specified average molecularweight (PEGDME 250, PDMS 550, PPGDMEl 678 and PPGDMEb 430) is determined between 25 and 120 °C at pressures to10 MPa. CO2 is much more soluble in each solvent than H2; however, the solubility of CO2 decreases as the solubility of H2increases with increasing temperature. PPGDMEl 678 and PPGDMEb 430 are comparable CO2 solvents. PPGDMEl 678 absorbsless H2 than all the other solvents, while PPGDMEb 430 absorbs significantly more H2. PDMS 550 is a very good CO2 solvent,absorbing more CO2 than all of the other solvents at all temperatures except for PEGDME 250 at 25 °C. PDMS 550 absorbsmore H2 than all of the other solvents.

■ INTRODUCTION

Chemical and physical absorption methods have beensuggested for the precombustion and postcombustion captureof CO2, respectively.

1 At low CO2 partial pressures of ∼0.01−0.015 MPa associated with postcombustion capture of CO2from flue gas, amine solutions are favored because they willreact with dilute concentrations of CO2. For example, a 30 wt %solution of monoethanamine (MEA) in water will bind CO2 asa water-soluble ammonium carbamate at a 2:1 molar ratio ofMEA:CO2, enabling this solution to absorb about 11 wt %CO2. CO2 release and solvent regeneration is accomplished viaheating.1,2 Precombustion capture of CO2 is typicallyaccomplished with physical solvents, given that the high partialpressure of CO2 in the fuel gas stream is sufficient to dissolvesignificant amounts of CO2 into the solvent without the needfor chemical reaction. For example, at 25 °C, if the partialpressure of CO2 is about 1 MPa, about 5 wt % CO2 willdissolve in the polymeric Selexol solvent that is commonlyemployed for low temperature CO2 absorption. Regenerationof the solvent and release of the CO2 can be accomplished withtemperature increase and/or pressure reduction.Numerous small volatile compounds such as methanol and

acetone3 and polymers have been considered as physical CO2solvents.4 The most common polymeric CO2 solvent is basedon an extremely hydrophilic polymer, polyethyleneglycoldimethylether (PEGDME). For example, the solvent used inthe Selexol process5,6 is a proprietary formulation that is rich inPEGDME. Given that the CO2 solvent strength of PEGDMEincreases with decreasing temperature and increasing pressure,it is not surprising that in most IGCC plant designs, the CO2

absorption is typically conducted at high pressure and lowtemperature (∼40 °C); under these conditions, most of thewater vapor in the post-WGSR stream has been condensed andseparated from the CO2- and H2-rich gas stream that is fed tothe absorption column. Therefore, the complete miscibility ofPEGDME and water is not problematic because there is verylittle water vapor (on a mass basis) in the gas entering theabsorption column even though the gas is saturated with water.The energetic penalties and capital costs associated with

cooling the fuel gas stream to 40 °C are substantial. Consider atypical IGCC fuel gas stream (31 mol % CO2, 43% H2, 23%H2O, and 3% of other gases such CO, COS, H2S) leaving theWGSR at 250 °C and 5.5 MPa. The stream could be cooledisobarically to its dew point of about 180 °C before thecomposition would change due to water condensation. Processmodeling at the US DOE NETL has indicated that if CO2 canbe selectively removed from within the WGSR or from thepost-WGSR stream using physical solvents with little or nocooling, and if the remaining H2−H2O gas mixture iscombusted to generate the hot, high pressure gas stream thatis expanded in the gas turbine, the IGCC plant thermalefficiency could increase by 2−3 percentage points.7

The objective of this work is to identify promising physicalsolvents for selective CO2 capture from a hot or warm highpressure gas stream rich in CO2, H2O, and H2. The maindisadvantage of PEGDME for this higher temperature

Received: August 29, 2013Revised: October 15, 2013Published: October 17, 2013

Article

pubs.acs.org/EF

© 2013 American Chemical Society 6913 dx.doi.org/10.1021/ef401740w | Energy Fuels 2013, 27, 6913−6920

absorption is its complete miscibility with water. PEGDME-richsolvents, such as Selexol, would remove both the water and theCO2 from the hot, humid fuel gas. This study focuses onhydrophobic solvents that would absorb as much CO2 and aslittle H2O and H2 as possible at temperatures above 40 °C.Although phase behavior studies involving these polymeric

solvents have been conducted, most reports focus on CO2solubility at low temperature. For example, our prior study ofphase behavior at 25 °C4 considers PEGDME 250 and threehydrophobic polymers, polypropyleneglycol dimethyletherbased on the branched (1,2-propanediol) monomer(PPGDMEb 230), perfluoropolyether (PFPE 960), andpolydimethylsiloxane (PDMS 237). (In each case, the averagemolecular weight of the polymer is provided after its acronym.)Concerning hydrophobicity, PDMS and PFPE are completelyimmiscible with water, PEGDME is completely miscible withwater in all proportions, and PPGDMEb absorbs several weightpercent water at ambient temperature. When CO2 solventstrength is assessed on a weight percentage basis, PEGDMEand PPGDMEb are comparable CO2 solvents. PDMS dissolvesslightly less CO2 than the polyethers at the same temperatureand pressure, and PFPE absorbs significantly less CO2.

4 Thesolubility of CO2 in polyethylene glycols of varying molecularweight, ranging from ethylene glycol monomethyl ether up toPEGDME 250 has also been reported at temperatures up to 60°C.8 An earlier paper9 reported similar results for the solubilityof CO2 in Selexol. The solubility of CO2 in a fluorinatedsilicone oil, trifluoropropylmethylsiloxane, with a kinematicviscosity of 300 centistokes has also been determined.10 TheCO2-philicity of all of these solvents is attributable to multiple,favorable Lewis acid:Lewis base interactions between CO2 andthe monomeric unit of each polymer.High pressure hydrogen solubility in these polymers has not,

to our knowledge, been previously published.The current study is a comparison of the solubility of CO2

and H2 in PEGDME (the main constituent of the hydrophilicsolvent Selexol) and three types of hydrophobic solvents overthe 25−120 °C range. The four solvents are shown in Figure 1:PEGDME 250, PPGDMEb 430 (poly(1,2-propylene glycol)dimethyl ether), PPGDMEl 678 (poly (1,3-propylene glycol)dimethylether), and PDMS 550 (polydimethyl siloxane).Because PDMS is available in a very broad range of repeat

units, the effect of average molecular weight on CO2 and H2solvent strength was determined at 25 °C using PDMS 237,550, 1250, 2000, and 3870.

■ EXPERIMENTAL SECTIONSolvents. PDMS 237, 550, 1250, 2000, and 3780 (Gelest Inc. 99+

%), PEGDME 250 (Aldrich 99+%), and PPGDMEb 430 (PolymerSource, 99+%) were used as received. The Selexol solvent waspurchased from Univar USA Inc. in January 2009 and used as received.

Poly(1,3-propanediol) 650 was kindly provided as a gift by DuPontvia a Materials Transfer Agreement and was used as received for thesynthesis of PPGDMEl 678. All other chemicals used in the synthesisof PPGDMEl, including triethylamine (TEA), methanesulfonyl-chloride, anhydrous tetrahydrofuran (THF), sodium methoxide,sodium hydroxide (NaOH), hydrochloric acid (HCl), magnesiumsulfate (MgSO4), sodium chloride (NaCl) and dichloromethane, wereobtained at high purity (Aldrich 99%) and used as received.

PPGDMEl is synthesized in a two-step process illustrated in Figure2. In the first step, a 2 molar solution of poly(1,3-propanediol) (1equiv) in dry tetrahydrofuran and 4 equiv of triethylamine (TEA) arecombined under nitrogen. Methanesulfonyl-chloride is addeddropwise by an addition funnel at 0 °C over a period of 3 h. Themixture is allowed to warm to room temperature and stirred overnight.The mixture is then filtered to remove the TEA-Cl salts, and thesolvent is then removed under reduced pressure. The obtainedproduct is then dissolved in dichloromethane and washed with water,0.5 molar solution of HCl, water, and saturated-NaCl solution. Thesolution is then dried over magnesium sulfate and filtered. Dichloro-methane is then removed under vacuum. The product is dried underhigh vacuum at 50 °C yielding 4,8,12,16,20,24,28,32,36,40-decaoxa-tritetracontane-1,43-diyldimethanesulfonate as a viscous liquid (90%yield).

In the second step, 2.4 equiv of sodium methoxide is added in dryTHF. In this suspension, dried 4,8,12,16,20,24,28,32,36,40-decaoxa-tritetracontane-1,43-diyldimethanesulfonate is added dropwise usingan addition funnel at 0 °C accompanied with mechanical stirring.Mechanical stirring is needed for this reaction due to formation of a gelduring the reaction. The reaction mixture is stirred vigorously for 4 hand small aliquots of the reaction mixtures are taken and analyzed withATR-FTIR. Upon completion, wet methanol is added in the reactionvessel to neutralized excess sodium methoxide. The mixture is thenfiltered and solvent is removed under vacuum. The product isdissolved in dichloromethane and washed with 0.5 M solution ofNaOH and 1 M solution of HCl followed by a water wash, andsaturated-NaCl solution wash. The solution is then dried over MgSO4and filtered. Dichloromethane is removed under vacuum and theproduct is dried under high vacuum at 50 °C yielding 88% as a paleyellow oil. FTIR showed no −OH peak present in the dried productwhich was further confirmed via NMR. 1H NMR (400 MHz, CDCl3):δ 3.22−3.58 (Br-s, 44 H, O−CH2−), 3.10- 3.22 (Br-s, 6H, −O−CH3),1.46−1.99 (Br-s, 22H, CH2−CH2−CH2). The synthesis is shown inFigure 2:

Gases. Carbon dioxide (Matheson Gas, 99.999%) and hydrogen(Matheson Gas, 99.999%) were both used without further purification.

Solubility of Water in Solvents, And Gel Formation. Smallamounts of water are added gravimetrically from a syringe to aspecified mass of solvent in a vial. The mixture is then stirred for 10min with a small magnetic stir bar. Water addition continues inincrements of about 0.10 wt % until the transparent mixture reaches acloud point, indicative of an excess water phase. The mixture with thehighest concentration of water that remains a stable, transparent, singlephase at 25 °C is reported as the solubility of water in the solvent.Mixtures containing water-saturated solvent and a small amount ofexcess water are then observed under quiescent conditions todetermine if gels form.

Solvent Density. The density of the solvents at ambient pressureis measured at 25 and 125 °C with an aluminum pychnometer(Fisher). The pychnometer volume is determined via calibration withwater at 25 °C, and the volume at 120 °C is calculated using theFigure 1. Chemical structures of the four polymeric solvents.

Energy & Fuels Article

dx.doi.org/10.1021/ef401740w | Energy Fuels 2013, 27, 6913−69206914

thermal expansion of aluminum combined with the appropriateexpressions for the change in pychnometer dimensions withtemperature.11,12

CO2 and H2 Solubility. All phase behavior measurements of eachbinary system are carried out in an invertible, high pressure, variable-volume, windowed, agitated, view cell (Schlumberger). The stainlesssteel cell has 8.00 in. high, 1.50 in. wide, 0.50 in. thick pressure-retaining borosilicate windows on opposing sides of the vessel. Thecell, which is rated to 69 MPa at 180 °C, houses a thick-walled Pyrextube (1.25 in. ID, 1.75 in. OD, 8.00 in. length, Schlumberger Ltd.) withpolished ends that contains a floating piston and has a working(sample) volume of up to 100 mL. A transparent overburden fluid(e.g., water, silicone oil) resides beneath the floating piston and aroundthe piston; therefore, the wall of the tube is not subject to a differentialpressure. The piston retains a Buna-N O-ring that maintains a sealbetween the sample volume above the piston and the overburden fluidbelow the piston. A ∼0.05 MPa pressure drop is required to move thepiston. Larger O-rings are seated above and below the polished ends ofthe Pyrex tube.The solubility of CO2 and H2 in the solvents is determined using

standard nonsampling techniques for bubble point detection describedin detail for CO2 bubble points elsewhere.

3,4 In all cases, a single phasemixture of known overall composition is very slowly expanded until asingle, persistent bubble is observed in equilibrium with the liquidphase. The bubble point data is determined at 298 K, 313 K, 353 K,and 393 K at CO2 mass fractions ranging between 0.04 and 0.25 or H2mass fractions ranging between 0.00005 and 0.00025 (50 to 250 ppm).The high pressure apparatus is housed in an air bath environmentalchamber (Cincinnati Sub Zero Products Inc.) capable of controllingthe temperature between 253 and 453 K, as measured with a type Kthermocouple to an accuracy of ±0.2 K.In a typical experiment for a CO2 bubble point, 30 g of solvent is

loaded gravimetrically from a syringe into the Pyrex tube on top of thepiston. The tube, piston, and solvent are then placed in the topopening of the view cell. The steel end-cap of the view cell, which alsocontains the mixer and the port for the venting or addition of gas intothe sample volume, is then bolted to the top of the vessel. A computer-controlled positive displacement (PD) pump (Schlumberger) is usedto displace the overburden fluid into the bottom of the view cell belowthe piston, thereby displacing the piston and the solvent upward anddecreasing the gas-filled volume above the solvent. This process ishalted when nearly all of the gas has been displaced from the samplevolume. The cell is then isolated by closing a valve on top of the cell,and the solvent is then compressed to ∼10 MPa using the overburdenfluid PD pump. Liquid CO2 at ambient temperature is thencompressed to the same pressure in a second PD pump. Tubingcarries the CO2 from the pump to the other side of the valve thatretains the solvent within the sample volume of the cell. The valve isthen quickly opened, and the CO2 pump is advanced as the

overburden fluid pump is retracted at the same volumetric rate. Thisallows for the well-controlled addition of CO2 into the sample volume.When the desired volume of CO2 has been introduced to the samplevolume, both pumps are turned off, and the valve on top of the cell isclosed. By recording the initial and final volume of CO2 in the finelycalibrated PD pump (0.01 mL), and the initial and final temperatureand pressure of the CO2, an equation of state13 can be used todetermine the amount of CO2 displaced into the sample volume. Thecell can then be heated to the desired temperature, and the samplevolume, which contains known amounts of solvent and CO2, is thencompressed via the slow addition of overburden fluid to the bottom ofthe cell. As the pressure is increased to ∼14 MPa, the CO2 and solventare mixed with the magnetically driven slotted fin impeller until asingle, clear, homogeneous liquid phase (L) or fluid phase (F) isachieved. The sample volume is then very slowly expanded at constanttemperature until a bubble point is observed. Pressure is determinedfrom a certified Heise gauge (14 MPa ± 0.03 MPa) that measures thepressure of the overburden fluid. Bubble point measurements arerepeated five times, and the average value is reported as the bubblepoint.

Because the (mass) solubility of hydrogen in these solvents issignificantly less than that of CO2 in the pressure range of interest, it isnot possible to accurately add very small amounts of dense H2 to thecell using the same technique. Therefore, the method of charging H2

involves loading the Pyrex tube and solvent into the cell and thenflushing the space above the liquid solvent with low pressure H2 inorder to displace the air from the sample volume. The gas and liquidare not mixed during this process in order to reduce the dissolution ofH2 in the solvent during this process. The flow of hydrogen is thenstopped and the sample volume, which contains the solvent and H2 atatmospheric pressure, is isolated. The volume of H2 is determined asthe product of the height of the sample volume above the liquid (asmeasured with a cathetometer) and the circular cross-sectional area ofthe tube. The mass of H2 in the sample volume is calculated as theproduct of gas volume and H2 density;

13 the amount of H2 that mayhave dissolved in the solvent at 0.1 MPa during the introduction of H2

is assumed to be negligible. The H2-solvent mixture is thencompressed and stirred until a single phase is attained.

■ RESULTS

Specific Gravity and Thermal Expansion. The specificgravity of each solvent at 25 and 120 °C is provided in Table 1.Each solvent becomes less dense as temperature increases. Thedensity values of the polyether solvents are comparable to thatof water at 25 °C (0.9970).13

Absorption of Water. PEGDME is miscible with water inall proportions; a single phase was always observed when water

Figure 2. General synthesis scheme of PPGDMEl.

Energy & Fuels Article

dx.doi.org/10.1021/ef401740w | Energy Fuels 2013, 27, 6913−69206915

and PEGDME were mixed at ambient temperature. No gelformation occurred.PDMS and water are completely immiscible; regardless of

the ratio of PDMS and water, two phases are always observed.For example, for each of the five PDMS solvents (MW 237,550, 1250, 2000, and 3780), a single drop of water could not bedissolved in the PDMS. This finding is consistent with Miller’sprevious observation that PDMS 550 and water are completelyimmiscible, even at 120 °C and 69 MPa.4 Gels or emulsionswere not observed for any PDMS-water mixtures, even afterprolonged contact.When water is added quickly to PPGDMEl and PPGDMEb at

23 °C, they absorb 3.0 and 2.0 wt % water, respectively. This isconsistent with previous work4 reporting the solubility of waterin PPGDMEb as 2.1 wt %. However, upon exposure to smallamounts of excess liquid water under quiescent conditions,PPGDMEl forms a rigid opaque gel in about 1 h, Figure 3. Ifthe gel is heated to 50 °C, then the gel breaks and the solventand water become free-flowing liquids. But the gel reformsslowly at ambient temperature.

PPGDMEb also slowly forms a gel in the presence of excesswater. Under quiescent conditions, a thin gel forms at thePPGDMEb−water interface. If the phases are thoroughlymixed, then all of the PPGDMEb will gel. The gel breaks atabout 80 °C and slowly reforms upon cooling.The formation of gels in the presence of very small amounts

of water could prove problematic in the handling and storage ofthe PPGDME solvents.Solubility of CO2 in Solvents. Bubble point loci at 25 °C

for mixtures of CO2 with each of the solvents are shown inFigure 4. The strongest CO2 solvents (a desirable trait for theproposed application) will exhibit the lowest pressure for aspecified mass fraction of CO2 dissolution. Alternately, thestrongest solvents absorb the greatest mass fraction of CO2 at aspecified pressure. At 25 °C PEGDME 250 is the strongestCO2 solvent. PPGDMEb 430 is the next best CO2 solvent,exhibiting a bubble point pressure about 10% greater than that

of PEGDME 250 at the same mass fraction of CO2, or about20% less dissolved CO2 at the same pressure. At concentrationsup to 0.10 mass fraction CO2 in the solvent, PPGDMEl 678 is aslightly weaker solvent than PPGDMEb 430, especially athigher concentrations of CO2. PDMS 237 is a slightly weakerCO2 solvent than PPGDMEb over the entire concentrationrange studied.The effect of solvent molecular weight was determined only

with PDMS. As the molecular weight of PDMS increases from237 to 3870, the bubble point pressure over the entire range ofmass fraction of dissolved CO2 increases steadily. This reflects adiminishing solvent strength with increasing PDMS molecularweight. The increase in the bubble point pressure for thedissolution of CO2 in PDMS may be due to steric effects causedby the longer chains. The longer silicone chains may reduce theaccess of CO2 to the oxygen atoms, onto which the carbon ofthe CO2 is attracted, and as a result reduce the solubility. Theresults are shown in Figure 4:The effect of temperature on the solubility of CO2 in these

solvents is illustrated in Figures 5− 8 for PEGDME 250,PPGDMEb 430, PPGDMEl 678, and PDMS 550, respectively.In all cases, the solubility of CO2 decreases with increasingtemperature. This reflects the diminishing strength of thefavorable Lewis acid:Lewis base interactions between CO2 andthe polymer with increasing temperature.The increase in bubble point pressure for PDMS 550 with

increasing temperature is less than observed for the polyethers,Figures 5−7. This is also illustrated in Figure 9, which providesthe bubble point pressure for all of the solvents as a function oftemperature for the 0.10 mass fraction CO2 dissolved in thesolvent.The increase in bubble point pressure with temperature is

very similar for the PEGDME 250, PPGDMEb 430, andPPGDMEl over the entire temperature range. Figure 8 alsoillustrates that while PEGDME 250 is the strongest CO2solvent at 25 °C and 40 °C, at 80 °C and 120 °C PDMS550 is the strongest CO2 solvent.

Table 1. Specific Gravity (i.e., Density in g/mL) of the FourSolvents at 25 and 120 °C

PDMS 550 PEGDME PPGDMEb PPGDMEl

25 °C 0.870 1.008 0.990 0.988120 °C 0.774 0.924 0.924 0.927density change −11.0% −8.3% −6.6% −6.2%

Figure 3. PPGDMEl linear slowly forms a gel under quiescentconditions in the presence of excess water. The light brown opaquephase is the gel, the transparent liquid is excess water. PPGDMEb gelsare similar in appearance, although they form more slowly.

Figure 4. The bubble point pressure for various CO2 solvents versusthe mass fraction of CO2, all at 25 °C. The strongest CO2 solvent isPEGDME 250, the weakest CO2 solvent is PDMS 3870. The size ofthe data markers reflects the range of the five bubble point pressurescollected during each experiment.

Energy & Fuels Article

dx.doi.org/10.1021/ef401740w | Energy Fuels 2013, 27, 6913−69206916

The changes in solubility of CO2 in these four solvents is alsomodeled by determining the Henry’s law constant for eachbinary at each temperature, and then using those results to

determine the enthalpy of dissolution. The Henry’s lawconstant is determined using eq 1,

=K P w/H (1)

where KH is the Henry’s law constant (MPa), P is the CO2pressure (MPa), and w is the dimensionless mass fraction. CO2concentrations up to and including the 0.15 mass fraction dataare used because the data was linear through the origin andtherefore reflected the slope of the bubble point curve atinfinite dilution.14

Table 2 lists the Henry’s law constants for each CO2-solventsystem at 25 °C, 40 °C, 80 °C, and 120 °C, based on a bestlinear fit of the bubble point data up to 15 wt % CO2. Henryconstants derived from literature are also provided. In theliterature data8,9 the Henry’s law constant was provided interms of mole fraction CO2 dissolved; these values wereconverted to mass fractions for comparison with our results.The enthalpy of dissolution, ΔH, can be determined from

the slope of the natural logarithm of the Henry’s law constant(MPa) against the inverse absolute temperature (K), accordingto the van’t Hoff formula, eq 2:

δδ

Δ =HR T

ln K1/

H

(2)

Figure 5. The effect of temperature on bubble point pressure for theCO2−PEGDME 250 mixture at 25 °C, 40 °C, 80 °C, and 120 °C.

Figure 6. The effect of temperature on bubble point pressure for theCO2−PPGDMEb 430 mixture at 25 °C, 40 °C, 80 °C, and 120 °C.

Figure 7. The effect of temperature on bubble point pressure for theCO2−PPGDMEl 678 mixture at 25 °C, 40 °C, 80 °C, and 120 °C.

Figure 8. The effect of temperature on bubble point pressure for theCO2−PDMS 550 mixture at 25 °C, 40 °C, 80 °C, and 120 °C.

Figure 9. The effect of temperature on bubble point pressure for 0.10mass fraction CO2 in PEGDME 250, PPGDMEb 430, PPGDMEl 678,and PDMS 550 at 25 °C, 40 °C, 80 °C, and 120 °C.

Energy & Fuels Article

dx.doi.org/10.1021/ef401740w | Energy Fuels 2013, 27, 6913−69206917

The results, illustrated in Figure 10, are used to determinethe enthalpy of dissolution values provided in Table 3.

The enthalpy of CO2 dissolution of PDMS 550 is lower thanany of the other solvents, which reflects that the solubility ofCO2 in PDMS 550 decreases more slowly with temperaturethan the other three solvents.Solubility of H2 in Solvents. The solubility of hydrogen in

each of the solvents is provided in Figures 11−14. As expected,the solubility of H2 in each solvent is significantly less than thesolubility of CO2 over the temperature range studied. Animportant distinction, however, is that the solubility of H2 inthe solvents increases with increasing temperature. The increasein hydrogen solubility with temperature over wide temperatureranges has been observed previously for solutions of H2 inmany solvents, such as creosote,15 water,16 methanol,17 andtoluene.18 Bruner’s study19 substantiated this behavior for 10organic solvents over the 298−373 K temperature range. Arecent study on this phenomenon in H2-ionic liquid systems20

concluded that the primary factors responsible for this behavioris the extreme lightness and small intermolecular forces of

hydrogen molecules, which approach those of an ideal gas.Because of its small size and the absence of intermolecularforces, the increasing H2 solubility in the solvent was attributed

Table 2. Henry’s Law Constants, KH in MPa, for CO2−Solvent Systems at 25 °C, 40 °C, 80 °C, and 120 °C Basedon Solubility Data to 15 wt % CO2

a

25 °C 40 °C 80 °C 120 °C

PDMS 550 23.0 26.5 39.1 51.4PEGDME 250 16.4 20.9 38.2 58.6PEGDME 2508 15.6Selexol 2509 16.8PPGDMEb 430 25.1 32.8 51.4 72.8PPGDMEl 678 27.0 31.3 53.8 75.5

aLiterature data also provided for comparison.

Figure 10. The van’t Hoff plot for the CO2−solvent binary systems.

Table 3. Enthalpy of Dissolution for the CO2−SolventBinary Mixtures for the 25−120 °C Temperature Range

ΔHdiss

PDMS 550 −8.4 kJ/molPEGDME 250 −13.3 kJ/molPPGDMEb 430 −10.8 kJ/molPPGDMEl 678 −10.9 kJ/mol

Figure 11. Lines showing the bubble point lines for PEGDME 250 at25 °C, 40 °C, 80 °C, and 120 °C versus the H2 concentration in ppm.

Figure 12. Lines showing the bubble point lines for PPGDMEb at 25°C, 40 °C, 80 °C, and 120 °C versus the H2 concentration in ppm.

Figure 13. Lines showing the bubble point lines for PPGDMEl at 25°C, 40 °C, 80 °C, and 120 °C versus the H2 concentration in ppm.

Energy & Fuels Article

dx.doi.org/10.1021/ef401740w | Energy Fuels 2013, 27, 6913−69206918

to the expansion of the solvent upon heating (all of the solventsused in this study exhibited decreasing density with increasingtemperature, Table 1). Unlike CO2, which exhibits multiple,specific, favorable thermodynamic interactions with the solventsthat diminish in strength with increasing temperature, there areno analogous favorable thermodynamic interactions for the H2-solvent mixtures, and therefore no associated decrease in themagnitude of such interactions that diminishes H2 solubility.This density decrease is associated with increased free voidswithin the large polymer molecules into which the H2 canreside. This effect is thought to be very significant because eventhe slightest decrease in density resulting from a temperatureincrease results in very small free spaces that the extremelysmall hydrogen molecule can still manage to fit into, thusincreasing solubility. This concept is confirmed by calculationsperformed from experimental data of gas solubility in organicsolvents.21

Below is a table (Table 4) of the Henry’s law constant forsolubility of H2 in the four solvents at 25 °C, 40 °C, 80 and 120

°C. The H2−Selexol Henry constant based on monitoringpressure reduction of H2 in the vapor space above Selexol in aclosed stirred vessel22 is also provided for comparison with theH2-PEGDME 250 result.The poorest H2 solvent, which is the desired attribute for the

proposed application, is PPGDMEl 678, followed by PEGDME250, both of which are linear polymers. PPGDMEb and PDMS550, with one or two pendent methyl groups extending fromeach repeat unit of the polymer backbone, respectively, werethe best solvents for H2.

The van’t Hoff results for this data set are illustrated inFigure 15, are used to determine the enthalpy of dissolutionvalues, shown in Table 5.

One way to quantify how much more soluble CO2 is than H2in the solvents (on a mass basis) is to determine the ratio of theH2 to CO2 Henry’s law constants found in Tables 4 and 2,respectively. These values are provided in Table 6.

These results illustrate approximately how much moresoluble CO2 is than H2 in the solvent on a mass basis at aspecified temperature. For example, CO2 is 1646 times moresoluble than H2 in PEGDME 250 at 25 °C, based on acomparison of pure gas solubility. In all cases, the ratiodecreases with increasing temperature, an unfavorable trend forthe proposed application of absorbing CO2, but not H2, at hightemperature. PEGDME 250 exhibits the most favorable values.PPGDMEb and PPGDMEl have the next most favorable ratios.PDMS 550 exhibits the lowest ratios, especially at the lowesttemperature.

Figure 14. Lines showing the bubble point lines for PDMS 550 at 25°C, 40 °C, 80 °C, and 120 °C versus the H2 concentration in ppm.

Table 4. Henry’s Law Constant in GPa (Using theConcentration As the Weight Fraction of H2) for the FourSolvents at Four Temperatures up to 120 °C

25 °C 40 °C 80 °C 120 °C

PDMS550 9.0 8.4 7.4 5.8PEGDME 250 27.0 26.0 20.2 15.2Selexol 250a 25.7PPGDMEb 430 16.1 15.3 13.0 11.1PPGDMEl 678 30.0 28.5 24.7 20.0

aThe Selexol result was obtained using the pressure reductiontechnique.22

Figure 15. The van’t Hoff plot for the H2−solvent binary systems.

Table 5. Enthalpy of Dissolution for the H2−Solvent BinaryMixtures for the 25−120 °C Temperature Rangea

ΔHdiss

PDMS 550 +4.3 kJ/molPEGDME 250 +6.0 kJ/molPPGDMEb 430 +4.1 kJ/molPPGDMEl 678 +3.8 kJ/mol

aNote that the + sign reflects that the solubility of H2 increases withtemperature.

Table 6. Ratio of the Henry’s Law Constants for H2−Solventto the Henry’S Law Constant for CO2−Solvent System at 25,40, 80, and 120 °Ca

25 °C 40 °C 80 °C 120 °C

PDMS 550 391 317 189 113PEGDME 250 1646 1244 528 259PPGDMEb 430 1111 910 459 265PPGDMEl 678 641 466 253 152

aThe ratio reflects how much more soluble CO2 is in a solvent than H2(mass basis) at the same temperature.

Energy & Fuels Article

dx.doi.org/10.1021/ef401740w | Energy Fuels 2013, 27, 6913−69206919

■ CONCLUSIONSThe selective removal of CO2 from water and H2 in the hotpost-WGSR stream of an IGCC power plant could increase theefficiency of an IGCC by several percentage points. PEGDMEis an excellent solvent for absorbing CO2 but not H2, but it isextremely hydrophilic. PPGDME absorbs only several masspercent water at 25 °C, but it slowly forms a reversible gel at 25°C when left in contact with excess liquid water. PPGDMEb430 and PPGDMEl 678 are comparable CO2 solvents thatabsorb less CO2 than PEGDME 250 over the entiretemperature range. PPGDMEl 678 is the poorest H2 solvent,followed by PEGDME 250. PEGDMEb 430 absorbs muchmore H2 than either PEGDME 250 or PPGDMEl 678,probably due to the increased free volume associated withthe pendent methyl group of the PPGDMEb repeat unit. Thesolubility of water in PDMS is undetectable, and PDMS doesnot gel in the presence of liquid water. PDMS 550 absorbsslightly less CO2 than PEGDME 250 at 25 °C, and about thesame amount of CO2 as PEGDME 250 at 40 °C. At 80 and 120°C PDMS 550 is the best CO2 solvent. PDMS 550 absorbsmore H2 than any of the other solvents, however, probably dueto the presence of two pendent methyl groups. Increasingtemperature diminishes the strength of the favorablethermodynamic interactions between CO2 and all the solvents,thereby lowering CO2 solubility, while increasing the freevolume that enhances the solubility of H2. These trends areunfavorable for the proposed application.Given its low cost, availability in bulk quantities, thermal

stability, environmentally benign nature, low viscosity, minimalhealth and safety concerns, and availability over a widemolecular weight range, PDMS appears to be the mostpromising candidate for the proposed. However, modeling ofan IGCC power plant should be performed to ascertainwhether the solubility of H2 in the PDMS is unacceptably high.Further, the performance of other silicone oils, such aspolymethylphenyl siloxane and polydimethyldiphenyl siloxane,should be assessed.

■ AUTHOR INFORMATIONCorresponding Author*E-mail: [email protected] authors declare no competing financial interest.

■ ACKNOWLEDGMENTSAs part of the National Energy Technology Laboratory’sRegional University Alliance (NETL-RUA), a collaborationinitiative of the NETL, this technical effort was performedunder the RES contract DE-FE0004000.

■ REFERENCES(1) MacDowell, N.; Florin, N.; buchard, A.; Hallett, J.; Galindo, A.;Jackson, G.; Adjiman, C. S.; Williams, C.; Shah, N.; Fenell, P. AnOverview of CO2 Capture Technologies. Energy Environ. Sci. 2010, 3,1645−1669.(2) Porcheron, F.; Gibert, A.; Mougin, P.; Wender, A. HighThroughput Screening of CO2 Solubility in Aveous MonoamineSolutions. Environ. Sci. Technol. 2011, 45, 2486−2492.(3) Miller, M. B.; Luebke, D. R.; Enick, R. M. CO2-philic Oligomersas Novel Solvents for CO2 Absorption. Energy Fuels 2010, 24, 6214−6219.(4) Miller, M. B.; Chen, D-L; Xie, H-B; Luebke, D. R.; Johnson, J. K.;Enick, R. M. Solubility of CO2 in CO2-philic Oligomers;

COSMOtherm Predictions and Experimental Results. Fluid PhaseEquil. 2009, 287, 26−32.(5) McKetta, J. J.; Cunningham, W. A., Encyclopedia of ChemicalProcess and Design; Marcel Deccar: New York, 1995.(6) Reighard, T. S.; Lee, S. T.; Olesik, S. V. Determination ofMethanol/CO2 and Acetonitrile/CO2 Vapor/Liquid Phase EquilibriaUsing a Variable-Volume View Cell. Fluid Phase Equilib. 1996, 123,215−230.(7) DOE/NETL. Current and Future IGCC Technologies; A PathwayStudy Focused on Non-Carbon Capture Advanced Power Systems R&DUsing Bituminous CoalVolume 1, report DOE/NETL-2008/1337,Oct. 16, 2008.(8) Henni, A.; Tontiwachwuthikul, P.; Chekma, A. Solubilities ofCarbon Dioxide in Polyethylene Glycol Ethers. Can. J. Chem. Eng.2005, 83, 358−361.(9) Xu, Y.; Scutte, R. P.; Helper, L. G. Solubilities of Carbon Dioxide,Hydrogen Sulfide and Sulphur Dioxide in Physical Solvents. Can. J.Chem. Eng. 1992, 70, 569−573.(10) Wedlake, G. D.; Robinson, D. B. Solubility of Carbon dioxide insilicone Oil. J. Chem. Eng. Data 1979, 24, 305−6.(11) Klemens, P. G. Thermal Expansion of Aluminum at LowTemperatures. Phys. Rev. 1960, 120, 843−844.(12) Atkins, P.; J. de Paula Physical Chemistry, 8th ed.; W.H. Freeman:New York, 2006.(13) National Institute of Standards and Technology, StandardReference Database Number 69, 2011.(14) Thormeier, K. Solubility of the Noble Gases in Liquid Sodium.Nucl. Eng. Design 1970, 14, 69−82.(15) Prather, J. W.; Ahangar, A. M.; Pitts, W. S.; Hemnley, J. P.;Tarrer, A.R. R.; Guin, J. A. Solubility of Hydrogen in Creosote Oil atHigh Temperatures and Pressures. Ind. Eng. Chem., Process. Des. Dev.1977, 16 (3), 267−270.(16) Baranenko, V. I.; Kirov, V. S. Solubility of Hydrogen in Water ina Broad Temperature and Pressure Range. Atomnaya Energiya 1989,66 (1), 24−28.(17) Descampes, C.; Coquelet, C.; Bouallou, C.; Richon, D.Solubility of hydrogen in methanol at temperatures from 248.41 to308.20 K. Thermochim. Acta 2005, 430, 1−7.(18) Simnick, J. J.; Sebastian, H. M.; Lin, H.-M.; Chao, K.-C.Solubility Of Hydrogen In Toluene At Elevated Temperatures AndPressures. J. Chem. Eng. Data 1978, 23 (4), 339−340.(19) Brunner, E. Solubility Of Hydrogen In 10 Organic Solvents At298.15, 323.15, And 373.15 K. J. Chem. Eng. Date 1985, 30, 269−273.(20) Raeissi, S.; Peters, C. J. Solubility in Ionic Liquids, IncludingExperimental Data in [bmim][Tf2N]. AIChE J. 2012, 58 (11), 3553−3559.(21) Korosy, F. Two Rules Concerning Solubility of Gases andCrude Data on Solubility of Krypton. Trans Faraday Soc. 1937, 33,416−425.(22) Heintz, Y. J.; Sehabiague, L.; Morsi, B. I.; Jones, K. L.; Luebke,D. B.; Pennline, H. W. Hydrogen Sulfide and Carbon DioxideRemoval from Dry Fuel Gas Streams Using as Ionic Liquid as aPhysical Solvent. . Energy Fuels 2009, 23, 4822−4830.

Energy & Fuels Article

dx.doi.org/10.1021/ef401740w | Energy Fuels 2013, 27, 6913−69206920