Survey of Carbon Dioxide Capture in Phosphonium-Based Ionic

Feb 2, 2011 - We rationalized that polyamines, such as DETA (DETA = diethylenetriamine), would be excellent absorbers of carbon dioxide in nonvolatile...
0 downloads 0 Views 2MB Size
Download PDF
ARTICLE pubs.acs.org/IECR

Survey of Carbon Dioxide Capture in Phosphonium-Based Ionic Liquids and End-Capped Polyethylene Glycol Using DETA (DETA = Diethylenetriamine) as a Model Absorbent§ Naomi D. Harper,† Katie D. Nizio,† Arthur D. Hendsbee,† Jason D. Masuda,† Katherine N. Robertson,† Luke J. Murphy,† Michel B. Johnson,‡ Cory C. Pye,† and Jason A. C. Clyburne*,† †

The Maritimes Centre for Green Chemistry, Department of Chemistry, Saint Mary’s University, Halifax, Nova Scotia B3H 3C3, Canada ‡ Department of Chemistry and Institute for Research in Materials, Dalhousie University, Halifax, Nova Scotia B3H 4J3, Canada

bS Supporting Information ABSTRACT: We have performed a survey of carbon dioxide capture in phosphonium-based ionic liquids and end-capped polyethylene glycol using DETA (DETA = diethylenetriamine) as a model absorbent. The carbon dioxide adduct of DETA is zwitterionic and forms readily in both ionic and polymeric media. Crystals of the adduct have been isolated, and its X-ray crystal structure is reported. The adduct is hygroscopic and decomposes with heating, with the onset of decomposition occurring at ∼85 °C. Computational studies were performed examining possible structures for the adducts formed between DETA and CO2 in the gas phase. Also described is a study of the thermal stability of phosphonium ionic liquids and selected polymeric solvents under neutral and oxidizing conditions. Heat capacity measurements were also performed on a number of ionic liquids.

1. INTRODUCTION Nonvolatile solvents are extremely attractive reagents for the development of new separation technologies. Among the nonvolatile solvents are three broad classes of chemicals: ionic liquids,1,2 polymeric solvents,3,4 and deep eutectic materials.5 All of these compounds are used in industrial processes, and new applications for these materials are identified each year.6 In addition to their oft touted green chemical significance, which is primarily earned by their relatively nonvolatile nature, these novel compounds are attractive due to their unique material properties. These properties, which include low volatility, chemical resistance, and tunable solvent characteristics, make them extremely attractive for separations of small molecules, particularly gases.7 Currently there is a strong need to reduce carbon dioxide (CO2) emissions. Although much of this reduction can be achieved through increases in efficiency, there is still a requirement for improved carbon capture methods as well as development of the subsequent sequestration technologies.8 Three methods appear promising for carbon capture. First, CO2 can be produced in a relatively pure form through the oxycombustion method.9 In this method, capture of CO2 is easy because it is the only gaseous product of the combustion process. Second, the precombustion method10-12 involves separating fossil fuels into hydrogen and carbon dioxide before they are burnt. The fuel is converted into a synthetic gas “syngas” comprising carbon monoxide/dioxide and hydrogen/water. The “syngas” can be reacted with steam at high pressure (the water gas shift reaction) to produce a mixture with higher concentrations of CO2 (35-45%) and hydrogen, from which the CO2 can then be captured. Finally, CO2 can be captured from flue gas.8 This method is more challenging because flue gas is a mixture of gases. Perhaps the simplest flue gas in existence is that obtained by the combustion of natural gas. r 2011 American Chemical Society

Under dry conditions it contains ca. 11% CO2, 1% O2, and 88% N2.13 Other flue gases, particularly those from coal combustion, are typically contaminated with sulfur-containing compounds, mercury, and other byproduct, thus making purification more challenging.13 Selectively capturing components of flue gases is technically demanding, but it is made easier if the gas of interest is a Lewis acid. Acidic gases, such as carbon dioxide and hydrogen sulfide, are easier to capture. Scrubbing technology using amines is well developed and extensively used in natural gas sweetening technologies.14 Rochelle et al. have written extensively on the use of amines for CO2 capture by gas scrubbing.15,16 One of the most common techniques utilizes monoethanolamine (MEA) in aqueous solution for carbon dioxide capture. MEA reacts to form a temperature sensitive carbamate composed of an ammonium cation and a carbamate anion (Scheme 1a).17 Ionic liquids have previously been recognized as useful materials for the separation of carbon dioxide from gas mixtures.18 Perhaps this is best illustrated by the following four task specific ionic liquids (TSILs) where either the cation19,20 or the anion21-23 has been functionalized with a primary amine. The first such liquid, shown in Scheme 2a, contains an imidazolium cation and is reported to be very useful for carbon capture.19 This TSIL demonstrates the concept of carbon capture very well, but as the authors describe in the original report, it suffers from high viscosity19 and tar formation 24 during the capture reaction. Nonsubstituted Received: August 17, 2010 Accepted: December 29, 2010 Revised: December 1, 2010 Published: February 2, 2011 2822

dx.doi.org/10.1021/ie101734h | Ind. Eng. Chem. Res. 2011, 50, 2822–2830

Industrial & Engineering Chemistry Research Scheme 1. Reaction Scheme Showing the Reaction Stoichiometry of MEA (1a) and DETA (1b) with CO2

phosphonium TSILs possessing amino-acid based anions21 (Scheme 2b) and amino-phosphonium substituted TSILs with amino-acid based anions20 (Scheme 2c) have also been used to trap CO2. Initial studies showed the latter capturing twice the amount of gas as the former.20 However, a more recent publication suggests that phosphonium based ionic liquids with amino acid derived anions can also capture carbon dioxide with a 1:1 stoichiometry.21 On the whole, all of these capture systems appear to be robust but are limited by the fact that the TSIL is both the capture material and dispersant. Finally, absorption of CO2 by a TSIL synthesized from renewable materials, (2-hydroxyethyl)-trimethylammonium (S)-2-pyrrolidine-carboxylic acid salt, [Choline][Pro], (Scheme 2d) in an ionic liquid/polyethylene glycol mixture ([Choline][Pro]/PEG200) has also been reported. Inclusion of PEG was found to enhance the kinetics of the absorption of CO2.23 Rather than direct incorporation of amino-functionalized cations or anions, recent reports describe the use of imidazolium-based ionic liquids with amines added as the CO2 capture reagents.24,25 These systems appear to work well, but it is important to note that they are based on imidazolium ionic liquids. We26 and others27 have reported extensively on the sensitivity of these materials. The C2-H unit is extremely reactive, deprotonation requiring only mild bases such as 1,4-diazabicyclo[2.2.2]octane (DABCO) and 3-hydroxyquinuclidine, or amines, to initiate decomposition.28 Furthermore, it is well-known that they react with oxygen,29 and they are more temperature sensitive than other ionic liquid alternatives.30 We have been interested in the reactions of basic materials in ionic liquids for some time.31 We have previously reported the formation of persistent solutions of BH3 in phosphonium ionic liquids32 as well as the preparation of both electron rich and highly basic materials in these ionic liquids.33-35 A natural extension of this work is the capture of another acidic gas, CO2, in phosphonium-based ionic liquids. We have also extended the study to include liquid polymers. Choice of Capture Molecule and Nonvolatile Solvents Used in the Study. The capture of carbon dioxide can be achieved using a number of different molecules, including primary, secondary, or tertiary amines as well as other bases. In industrial CO2 capture processes, monoalkanolamines are typically used as trapping agents. In such cases, two molecules of absorbent are required to trap one molecule of carbon dioxide. We rationalized that polyamines, such as DETA (DETA = diethylenetriamine), would be excellent absorbers of carbon dioxide in nonvolatile solvents since only one molecule of absorbent is required to bind one molecule of CO2 (Scheme 1b). Furthermore, a somewhat larger amine will have a higher boiling point thus entraining the capture molecule more effectively in the dispersal medium (i.e., the

ARTICLE

nonvolatile solvent). The boiling point of DETA is 207 °C, whereas the boiling point of the common capture molecule, ethanolamine, is 170 °C.36 From a financial perspective, DETA is an industrial commodity, and its cost is comparable to, or less than, that of ethanolamine. The use of DETA to capture carbon dioxide has recently been reported in both an elegant NMR study, which described the formation of CO2 adducts at a variety of pressures,37 and a patent which alluded to the use of DETA as an aqueous amine for CO2 capture.38 Particularly interesting in the latter report was the description of a method to steam strip the CO2 from the adduct formed. For carbon capture it is beneficial that the medium in which the capture reaction is performed be nonvolatile. In this light, nonvolatile solvents such as ionic liquids, polymeric solvents, and deep eutectic solvents are all attractive options. In this work, we chose to examine phosphonium-based ionic liquids. These materials are extremely robust and are much more resistant to reactions with bases such as amines than are imidazolium-based ionic liquids.30 For example, in the Baylis-Hillman reaction, amines react with the imidazolium ion to cause a significant decrease in yield.39 This reaction, in our opinion, would lead to significant consumption of the ionic liquid during CO2 capture, which would be detrimental to the performance of the system. We also speculated that the carbamate salts formed in the trapping procedure would be sufficiently basic as to also react with the solvent, and therefore imidazolium ionic liquids were to be avoided. Phosphonium ionic liquids are commercially available40 and have the longest track record of any ionic liquid being used in an industrial setting,41 having been utilized commercially in the preparation of 2,5-dihydrofuran.41,42 Phosphonium ionic liquids are quite viscous, but we have found that when a solute such as an amine is added to the ionic liquid, the viscosity is significantly decreased. Furthermore, the viscosity is also extremely sensitive to temperature, and even modest warming of a solution can result in a significant decrease in its viscosity.40 TSILs are generally expensive due to the multistep synthetic method required for their preparation. It was apparent to us that while TSILs, such as the tethered amine functionalized imidazolium salts, effectively capture CO2, it is unlikely that such materials will ever be used on a large scale for carbon sequestration. Tar formation is a clear indication that the imidazolium ion behaves as a noninnocent material in the presence of a strong base vis-a-vis the carbamate ion. We will not address the numerous pitfalls of imidazolium ions in this paper, but their reactivity at the acidic C2 site is well documented, as was mentioned above.26-28 Polymeric solvents are attracting the attention of industry as they are relatively easily prepared and affordable.3,4 Other attractive features of this class of solvent include biodegradability, unique solvation properties, and nontoxicity.3,4 For these reasons, we have included polymeric solvents in this investigation of novel carbon capture systems. In this paper we 1. Examine the long-term thermal stability of ionic liquids and polymeric solvents in air as a model for their extended use in an oxygen-containing environment. 2. Examine the reaction of DETA with CO2 in both ionic liquids and polymeric solvents. 3. Examine the thermal stability of the DETA-CO2 complex and discuss methods for releasing carbon dioxide from the complex in ionic liquids and polymeric solvents. 2823

dx.doi.org/10.1021/ie101734h |Ind. Eng. Chem. Res. 2011, 50, 2822–2830

Industrial & Engineering Chemistry Research

ARTICLE

Scheme 2. Task Specific Ionic Liquids Used in Carbon Capture

Table 1. Onsets of Decomposition Temperatures and Maximum Operating Temperatures Obtained for a Series of Ionic Liquids and Polymeric Solvents in a Neutral Environment (N2 Atmosphere) onset of decomposition temperaturea (°C)

maximum operating temperatureb (°C)

tetradecyl(trihexyl)phosphonium chloride

284

220

tetradecyl(trihexyl)phosphonium decanoate

275

170

tetradecyl(trihexyl)phosphonium dicyanimide

335

230

tetradecyl(trihexyl)phosphonium bis(triflamide) 1-butyl-3-methylimidazolium bis(triflamide) IL001

372 391

270 280

poly(ethylene glycol) dimethyl ether (MN = ∼250)

124

35

poly(ethylene glycol) dimethyl ether (MN = ∼500)

225

125

polydimethylsiloxane (trimethylsiloxy terminated, MW = 770)

114

25

polydimethylsiloxane (trimethylsiloxy terminated, 50 cSt)

307

200

(8-12% phenylmethylsiloxane)-(88-92% dimethylsiloxane)

342

230

(18-22% diphenylsiloxane)-(78-82% dimethylsiloxane)

353

230

sample

a

Onset of decomposition temperatures were defined as the temperature at which the solvent lost 5% of its mass while heating from 25 to 500 at 5 °C/min under a nitrogen atmosphere. b Maximum operating temperatures correspond to the maximum temperature at which an isothermal heating period of 8 h under a nitrogen atmosphere results in less than a 5% mass loss of the sample.

4. Report the crystal structure of the DETA-CO2 complex isolated from an ionic liquid.

2. RESULTS AND DISCUSSION As mentioned in the Introduction, flue gas contains trace amounts of oxygen in a balance of carbon dioxide and nitrogen. Our experience with nonvolatile solvents and small molecule chemistry indicated that the presence of oxygen would prove to be a challenge for any carbon capture system utilizing a nonvolatile solvent. Thus, we have assessed the thermal stability of a series of phosphonium ionic liquids and polymeric solvents under both neutral and oxidative conditions. Using thermogravimetric analysis (TGA), an onset of decomposition temperature and a maximum operating temperature, were measured under neutral conditions (i.e., nitrogen atmosphere) for each ionic liquid and polymeric solvent investigated (Table 1). This allowed the useful working temperature range to be determined for each solvent. For full experimental details see the Supporting Information. Overall, the ionic liquids and polymeric solvents were found to be less thermally stable than initially expected; most were only stable up to a temperature which was approximately 100 °C below their corresponding decomposition onset temperature. Furthermore, the maximum operating temperatures are expected to decrease with longer isothermal periods at high temperatures. Therefore, despite ionic liquids and polymeric solvents having no measurable vapor pressure at room temperature, thermogravimetric analysis revealed that the nonvolatile solvents investigated herein do decompose and/or evaporate into volatile components upon prolonged heating at elevated temperatures.

Additionally, the results showed that the lower molecular weight polymeric solvents are not as thermally stable as the higher molecular weight liquid polymers. Therefore, as polymer chain length increases, the thermal stability of the molecule must also increase. This increase in stability is likely due to the higher temperatures needed to untangle and to break the additional chain interactions expected in the longer chain polymers prior to their decomposition and/or evaporation (see below). In addition to thermogravimetric analysis, long-term heating investigations of the solvents were performed as seven-day stability experiments at 100 °C under neutral (argon atmosphere) and oxidative (open to the atmosphere) conditions. These experiments were carried out to determine whether the solvents were decomposing or evaporating into volatile components at elevated temperatures. Overall, the ionic liquids were found to change color, with the most extreme changes occurring under oxidative conditions (see the Supporting Information for the series of color change photographs assembled for each solvent). This intense change in color suggests, qualitatively, that the ionic liquids may be decomposing at high temperatures. The polymeric solvents, on the other hand, showed little or no color change under either neutral or oxidative conditions. Thus, the polymeric solvents appear to be evaporating, rather than decomposing, on heating. To further confirm this hypothesis, poly(ethylene glycol) dimethyl ether (MN = ∼250) was distilled at 150 °C. Infrared and NMR spectroscopy performed on the distillate and the original poly(ethylene glycol) dimethyl ether (MN = ∼250) sample revealed that the two liquids were identical. This supports the earlier proposal that polymeric 2824

dx.doi.org/10.1021/ie101734h |Ind. Eng. Chem. Res. 2011, 50, 2822–2830

Industrial & Engineering Chemistry Research

ARTICLE

Figure 1. Temperature dependence of the molar heat capacity of ionic liquids (data in Table S2 of the Supporting Information). The dotted lines joining the measured experimental points represent the assumed relatively linear rate of increase of heat capacity with temperature over the range studied. Key: IL 1 1-butyl-3-methylimidazolium chloride, IL 2 1-butyl-3-methylimidazolium bis(triflamide), IL 3 tetradecyl(trihexyl)phosphonium chloride, IL 4 tetradecyl(trihexyl)phosphonium decanoate, IL 5 tetradecyl(trihexyl)phosphonium dicyanimide, and IL 6 tetradecyl(trihexyl)phosphonium bis(triflamide).

solvents evaporate, rather than decompose, at elevated temperatures. In order to investigate the drastic color changes observed for the ionic liquids under oxidative conditions and to determine if the solvents might be reacting with either oxygen or water from the atmosphere to cause decomposition, deoxygenated water was added to the ionic liquid tetradecyl(trihexyl)phosphonium chloride, and the mixture was heated under neutral conditions. The resulting color change was found to closely resemble the color change observed under neutral conditions (i.e., with no water present), suggesting that atmospheric oxygen, and not water, may be causing the extreme color changes observed. The color change could result from O2 f O2 3 - formation, a common pathway of oxygen reactivity in ionic liquids, and/or subsequent reactions with the solvent.29 Furthermore, it was noted that the most drastic color changes were observed in those ionic liquids containing more strongly coordinating anions such as chloride. It may be that oxygen reactivity is enhanced by strongly coordinating anions and reduced by weakly coordinating anions. We speculate that O2, dissolved in the ionic liquid, may be more polarized and hence more reactive in more strongly coordinating ionic liquids. NMR spectroscopy was employed in an effort to determine the nature and structure of the colored impurities generated in the ionic liquids upon heating. Proton (1H) and phosphorus (31P) NMR spectra were obtained for each phosphonium ionic liquid sample before and after heating in both neutral and oxidative environments. Both the 1H and 31P NMR spectra obtained showed no changes between the samples before and after heating in either

environment. 31P NMR spectra revealed one resonance peak at 32-33 ppm as expected.43 Thus, the concentration of the colored impurities generated must have been too low to be detected in the NMR experiments run. Infrared spectroscopy performed on all ionic liquid and polymeric solvent samples before and after heating revealed several new peaks generated after heating under oxidative conditions. None of these peaks were assigned to a specific chemical structure, but some correspond to OH-containing fragments. Under neutral conditions, virtually no new peaks were observed. This suggests decomposition and/or evaporation into volatile components occurs more rapidly in an oxidative environment. On the whole, the ionic liquids and polymeric solvents investigated herein appear to have better overall thermal stability in a neutral environment than in an oxidizing environment. The heat capacity of a compound is defined as the ability of that compound to store heat. Generally, the energy can be stored in translational, vibrational, and rotational modes of the component molecules. A molecule containing more atoms should have more energy storage modes and thus a higher heat capacity. This general trend is observed in the heat capacities of the measured ionic liquids as shown in Figure 1. From the figure, it is clear that the imidazolium based ILs have lower molar heat capacities than the phosphonium ILs. Direct comparison of the pairs {IL1 and IL3}, chloride, and {IL2 and IL6}, bis(triflamide), show that the imidazolium salts have heat capacities roughly 700 J/mol*K lower than their phosphonium counterparts. This is as expected on the basis of their much lower 2825

dx.doi.org/10.1021/ie101734h |Ind. Eng. Chem. Res. 2011, 50, 2822–2830

Industrial & Engineering Chemistry Research

ARTICLE

Table 2. Specific Heats of Selected Ionic Liquids - Measured and Literature Values reference values molar heat capacities (J/mol*K)/temperature (K) compound 1-butyl-3-methylimidazolium chloride

this work

Holbrey45

Fredlake46

Yamamoro47

289.4/300.9

299/343.1

322.7/298

∼310/300

436/293.1

536.3/298

Troncoso48

Ge44

Bhlokin49

a 567.33/298.15

599/298

565.1/298.15

306.7/370.9 1-butyl-3-methylimidazolium

577.4/301.8

bis(triflamide)

608.8/372.7

tetradecyl(trihexyl)phosphonium

1208.7/301.6

bis(triflamide)

1335.7/373.0

molecular weights. Ge et al.44 also found the heat capacity of [P66614][NTf2] to be much greater than that of [C4mim][NTf2], with a difference of 767 J/mol*K at 298 K, which they attributed to the cumulative effect of increasing the length of the hydrocarbon chains on the cation. For both the imidazolium and phosphonium ILs the molar heat capacity increases with molecular weight of the anion (the cation being unchanged). The only exception to this observation is IL4 (the phosphonium decanoate salt) which has the highest measured molar heat capacity at 300 K but has only the second highest molecular weight. At ∼370 K, the order is as expected on the basis of molecular weight, chloride < dicyanimide < decanoate < bis(triflamide). Ge et al.44 also noted significant changes in the heat capacity as a function of the anion, demonstrating that the heat capacity increases with the size of the anion. The authors note, that as found for other properties, such as density and viscosity, the anion type has a greater impact than the cation on the heat capacity. Ge et al.44 also observed that in all cases, the relationship between heat capacity and temperature was approximately linear, with only about a 10% increase observed over the temperature range used (293-358 K). If the molar heat capacities increase in a linear fashion over the measured temperature range, then one can compare the slopes of the dotted lines joining the two measured points shown for each IL in Figure 1. The two imidazolium ILs show very similar rates of change over the temperatures studied, with the slopes of the two lines being very similar. The rate of increase in the heat capacity of the imidazolium salts with temperature is slower than that of the phosphonium ILs (slope is lower) as expected based on the same molecular weight arguments. The phosphonium ILs also show a similar rate of change (slope of the connecting lines) except for the decanoate salt (IL4) which has a much lower slope than the other three phosphonium ILs. The reason for this discrepancy is not clear. There are a number of literature values available to compare with the current experimental results (see Table 2). The majority of these values are for the 1-butyl-3-methylimidazolium chloride and bis(triflamide) salts. For the chloride salt, the measured molar heat capacities of 289.4 J/mol*K (300.9 K) and 306.7 J/mol*K (370.9 K) are somewhat lower than the reported literature values. For 1-butyl-3-methylimidazolium bis(triflamide) there is even more literature data available with which to compare the current experimental results, 577.4 J/mol*K (301.8 K) and 608.8 J/mol*K (372.7 K). Our results are only slightly higher than those of Troncoso et al.48 and Blohkin et al.49 and fall squarely in the range of all the reported results. Both Troncoso et al.48 and Ge et al.44 include discussion of the effect of impurities, such as chloride ion and water, on the measured heat capacities. Both groups observe that chloride has little effect on the measured heat

b 566.47/298.15 1366/298

capacity when present in relatively small amounts. Ge et al.44 found that the presence of water had a more important effect on lowering the heat capacity than the presence of halide in the ionic liquids they studied. In contrast to the imidazolium ILs, literature data have only been reported for one of the phosphonium ionic liquids studied in this work. The measured molar heat capacity for tetradecyl(trihexyl)phosphonium bis(triflamide) 1208.7 J/mol*K (301.6 K) and 1335.7 J/mol*K (373.0 K) can be compared to the values reported by Ge et al.44 It may be that our relatively lower values are a result of impurities (water) in the sample. On a molar basis the heat capacities of the ionic liquids studied are all much greater than that of water (75.38 J/mol*K at 298 K). However, their measured specific heat capacities on a gram basis (see Table S2 in the Supporting Information) are all much lower than that of water (4.184 J/g*K at 298 K) because of their higher molecular weights. This is even more generally true; compared with traditional organic solvents, the molar heat capacities of ionic liquids are normally higher while the specific heat capacities are usually lower, a property that can be exploited in green chemistry processes. For practical applications, volumetric heat capacities are often the key numerical index. These values have been calculated for the ionic liquids studied, using the specific heats recorded experimentally and density data available in the literature50-52 (see Table S2 in the Supporting Information). The calculated volumetric heat capacities ranged from 1.461 J/ml*K at 301 K for tetradecyl(trihexyl)phosphonium chloride to 2.011 J/ml*K at 372 K for tetradecyl(trihexyl)phosphonium dicyanimide. All of these values are much lower than the volumetric heat capacity of water, which is 4.180 J/ml*K at 298 K. Using ionic liquids as alternative solvents to water in carbon capture systems should result in considerable energy savings. Comparison values were calculated for the polymeric solvents, poly(ethylene glycol) 200 and 400 [PEG 200 and PEG 400] using data available in the literature. Although not exactly the solvents studied in this work, where PEG dimethyl ethers of nominal molecular weights 250 and 500 were utilized, the values should be close enough for valid comparison. Using the data reported by Francesconi et al.53 PEG 200 was determined to have a specific heat of 417.2 J/mol*K, an average molecular weight of 192 g/mol, a specific heat of 2.17 J/g*K, a density of 1.116 g/mL, and a volumetric heat capacity of 2.43 J/mL*K at 303.15 K and atmospheric pressure. Under the same conditions, PEG 400 was found to have a specific heat of 774.5 J/mol*K, an average molecular weight of 365 g/mol, a specific heat of 2.12 J/g*K, a density of 1.118 g/mL, and a volumetric heat capacity of 2.37 J/mL*K. The polymeric solvents generally appear to have specific heats and volumetric heat capacities that fall between those of the 2826

dx.doi.org/10.1021/ie101734h |Ind. Eng. Chem. Res. 2011, 50, 2822–2830

Industrial & Engineering Chemistry Research

ARTICLE

Figure 2. Vapor pressure of varying concentrations of diethylenetriamine (DETA) by volume in tetradecyl(trihexyl)phosphonium chloride between 50 and 130 °C.

Table 3. CO2 Capture Results Averaged over 3 Runs solvent: IL101 or PEG sample

DETA

H2O

amount CO2 bound*

g

mmol

g

mmol

g

mmol

g

mmol

mole ratio CO2:DETA

4.54 5.39

8.74 -----

0.25 0.25

2.42 2.42

---------

---------

0.07 0.16

1.59 3.64

0.66:1 1.50:1

IL 101 þ DETA

4.62

8.90

0.47

4.56

-----

-----

0.07

1.59

0.35:1

PEG þ DETA

5.33

-----

0.50

4.85

-----

-----

0.25

5.68

1.17:1

IL 101 þ DETA

4.19

8.07

0.96

9.31

-----

-----

0.08

1.82

0.20:1

PEG þ DETA

5.30

-----

0.97

9.40

-----

-----

0.17

3.86

0.41:1

10% DETA þ 10% Water IL 101 þ DETA þ H2O

4.60

8.86

0.46

4.46

0.49

27.20

0.03

0.68

0.15:1

PEG þ DETA þ H2O

5.30

-----

0.50

4.85

0.47

26.09

0.16

3.64

0.75:1

5% DETA IL 101 þ DETA PEG þ DETA 10% DETA

20% DETA

*

Exposure to CO2 lasted for 3 min at ambient temperature and pressure.

ionic liquids studied and water but closer to the ionic liquid range than to water. The vapor pressures of mixtures containing 10 or 20% DETA by volume as well as pure diethylenetriamine and pure tetradecyl(trihexyl)phosphonium chloride were recorded over a temperature range of 30 to 130 °C at a rate of one reading per second (Figure 2). Tetradecyl(trihexyl)phosphonium chloride and diethylenetriamine were dried and sparged with argon prior to use (see the Supporting Information). The data obtained are consistent with the hypothesis that the vapor pressure of the amine capture reagent should remain low under normal operating conditions. Entraining DETA in the ionic liquid results in a significant lowering of its vapor pressure over the solution. Carbon Dioxide Capture Reactions. Diethylenetriamine (DETA) formed clear, smooth-running solutions with both tetradecyl(trihexyl)phosphonium chloride and poly(ethylene glycol) dimethyl ether, MN = ∼250 (PEG). The solutions are

hygroscopic and can absorb considerable water when exposed to air. Furthermore, they readily react with carbon dioxide, even from the atmosphere, at ambient pressure. Solutions containing upward of 20% carbon dioxide were prepared and left sealed. They maintained their integrity for weeks. Upon exposure to carbon dioxide, solutions containing DETA react rapidly to form the carbamate zwitterionic material (Scheme 1b). This compound is a light, fluffy white material which, when exposed to water, even in trace amounts, turns to a sticky gum. The quantity of carbon dioxide absorbed is a function of added amine and water as summarized in Table 3. It is important to note that one molecule of absorbent amine effectively binds one molecule of carbon dioxide. Large scale reactions were also performed and broadly demonstrate that the CO2 capture systems reported herein are capable of scale-up (see Supporting Information, Table S4). The reaction of DETA with CO2 occurs rapidly. In one case, crystals of the adduct were isolated from the phosphonium ionic 2827

dx.doi.org/10.1021/ie101734h |Ind. Eng. Chem. Res. 2011, 50, 2822–2830

Industrial & Engineering Chemistry Research

ARTICLE

Figure 3. Molecular structure (left) and packing diagram (right) of the DETA-CO2 adduct.

Table 4. Energy (kJ/mol) of Diethylenetriamine þ CO2 Products Relative to the Most Stable Conformer (end-C1 #2) DETA þ CO2

a

HF/6-31Ga

HF/6-31þGa

MP2/6-31Ga

MP2/6-31þGa

6.51 -17.98

6.55 -18.64

5.29 -10.98

5.20 -18.38 11.09

NH2(CH2)2NH(CH2)2NHCOOH NH2(CH2)2NH(CH2)2NHCOOH

C1 #1 C1 #2a

NH2(CH2)2NH(CH2)2NHCOOH

C1 #3

10.26

10.39

10.80

NH2(CH2)2NH(CH2)2NHCOOH

C1 #4

5.18

5.59

5.92

6.68

H2N(CH2)2NCOOH(CH2)2NH2

Cs #1

96.63

99.68

75.88

75.73

H2N(CH2)2NCOOH(CH2)2NH2

Cs #2

88.74

92.05

67.67

68.72

H2N(CH2)2NCOOH(CH2)2NH2

C1 #1

20.08

20.60

5.20

4.43

Relative to DETA þ CO2.

liquid, tetradecyl(trihexyl)phosphonium chloride, as shiny white plates. Numerous adducts are theoretically possible,37 however, under the conditions tested, only the zwitterionic adduct shown in Scheme 1b was recovered. The crystal structure of this adduct is shown in Figure 3. Solution state spectroscopic data were difficult to obtain because the adduct was extremely insoluble, but elemental analysis and IR data are in accord with the crystallographic information. Notable IR spectral features include a broad amine N-H stretch peak at 3360 cm-1, a sharp amine N-H stretch peak at 3248 cm-1, and a sharp amine N-H bend peak at 1572 cm-1. In addition, a CdO stretch absorption at 1649 cm-1 is present, confirming the presence of CO2. The adduct, shown in Figure 3, is discrete but exhibits extensive intermolecular interactions within the layers forming twodimensional sheets. These sheets layer to form the crystal lattice and, on the microscopic level, are consistent with the macroscopic plate structure of the crystalline material. Thermogravimetric studies indicate that the adduct has an onset of decomposition temperature of 82 °C (corresponding to a 5% mass loss, at a heating rate of 5 °C/min under a nitrogen atmosphere). Furthermore, mass spectrometry studies demonstrate that the adduct releases CO2 between the temperatures of 75 and 145 °C. See the Supporting Information for TGA and MS scans. As described above, at these temperatures the phosphonium ionic liquids are extremely stable (as are 1-butyl-3-methylimidazolium bis(triflamide) and the polymeric solvents studied). Lastly, we emphasize that the adduct is a zwitterion, and not a discrete salt, i.e. it is not an ion pair. This portends well for favorable release energetics since lower cohesive forces hold the crystal together. Theoretical Calculations. Reaction of diethylenetriamine with carbon dioxide in the gas phase can give one of two possible carbamic acids. For terminal N-substitution, we examined four

Figure 4. The most stable conformation calculated in the gas phase for the neutral DETA-CO2, adduct C1 #2.

possible structures varying in the placement and orientation of the carboxyl group. Calculations were performed with Gaussian 03,54 using a stepping stone approach, details of which are given in the Supporting Information. The C1 #2 structure (Figure 4) was the most stable of the four. For central N-substitution, neither of the two Cs structures was a minimum, and upon relaxing the symmetry a major conformational change took place to give the C1 structure. All were of higher energy than the terminally substituted conformers. Three zwitterionic forms of the carbamic acid complex (NH3þ(CH2)2NH(CH2)2NHCOO- (ZW1), NH2(CH2)2NH2þ(CH2 )2 NHCOO- (ZW2), and NH 2 (CH 2 )2 N(COO -)(CH2 )2 NH3 þ (ZW3)) were also studied. ZW1 and ZW3 were not stable at the three lowest levels of theory, undergoing either a proton transfer from N to O, or a proton transfer from N to N(COO) with elimination of CO2. ZW2 also underwent elimination at HF/STO-3G and proton transfer at the MP2 levels but was stabilized enough at the other HF levels by hydrogen bonding to remain zwitterionic. It is quite clear that the zwitterionic forms are not stable in the gas phase. The naturally occurring amino acids are also neutral in the gas-phase, becoming zwitterionic only in solution or the solid state. The results of these calculations indicate that terminal addition to give the nondeprotonated adduct results 2828

dx.doi.org/10.1021/ie101734h |Ind. Eng. Chem. Res. 2011, 50, 2822–2830

Industrial & Engineering Chemistry Research in the most stable product being formed. This is entirely consistent with our experimental observations. Our current work is focused on bringing the rather disparate results reported herein together into one optimized system for carbon dioxide capture. To this end we are combining ionic liquids and polymeric solvents with amines suitable for effective carbon capture to identify a system that remains liquid under standard operating conditions, even after addition of CO2. Mixtures with a high capacity for CO2 but which remain as nonviscous liquids under all working conditions are in the process of being identified. Altering the combination of DETA with other amines is being used to control the physical state of the product after carbon capture. The energetics of complex formation with the chosen amine(s) must balance complex stability with facile removal of CO2 when desired. The best system will be chosen based on these criteria but it should also be stable, green, and inexpensive.

3. CONCLUSION We have performed a series of carbon dioxide capture experiments in phosphonium-based ionic liquids and end-capped polyethylene glycol using DETA (DETA = diethylenetriamine) as a model absorbent. The carbon dioxide adduct of DETA forms readily in both ionic and polymeric media. Crystals of the adduct have been isolated, and its X-ray crystal structure is reported. The material consists of discrete zwitterionic species that form a layered structure exhibiting extensive hydrogen bonds within the layers and minimal interactions between the layers. We have also studied the thermal stability of phosphonium ionic liquids and selected polymeric solvents under neutral and oxidizing conditions and shown that the presence of oxygen in the gas sample significantly affects the stability of the ionic liquid. ’ ASSOCIATED CONTENT

bS Supporting Information. General experimental procedures, thermogravimetric analysis experiments, long-term thermal stability testing, heat capacity measurements, vapor pressure measurements, carbon dioxide capture experiments, and theoretical calculations. This material is available free of charge via the Internet at http://pubs.acs.org. ’ AUTHOR INFORMATION Corresponding Author

*E-mail: [email protected].

’ ACKNOWLEDGMENT Funding was provided by the Natural Sciences and Engineering Council of Canada (NSERC) through the Discovery Grants Program to J.D.M. and J.A.C.C. J.A.C.C. acknowledges generous support from the Canada Research Chairs Program, the Canadian Foundation for Innovation, and the Nova Scotia Research and Innovation Trust Fund. The work was supported in part by GreenCentre Canada and Springboard. We are also grateful for the Nuclear Magnetic Resonance Research Resource (NMR-3) at Dalhousie University for NMR data acquisition. ’ ADDITIONAL NOTE § Taken in part from the thesis titled “A Critical Analysis of the Thermal Stabilities of a Series of Ionic Liquids and Polymeric

ARTICLE

Solvents under Neutral, Reductive, and Oxidative Conditions” by Katie D. Nizio, Saint Mary’s University, 2009.

’ REFERENCES (1) Holbrey, J. D.; Seddon, J. R. Ionic Liquids. Clean Products Processes 1999, 1, 223–236. (2) Smiglak, M.; Metlen, A.; Rogers, R. D. The Second Evolution of Ionic Liquids: From Solvents and Separations to Advanced Materials Energetic Examples from the Ionic Liquid Cookbook. Acc. Chem. Res. 2007, 40, 1182–1192. (3) Heldebrant, D. J.; Witt, H. N.; Walsh, S. M.; Ellis, T.; Rauscher, J.; Jessop, P. G. Liquid polymers as solvents for catalytic reductions. Green Chem. 2006, 8, 807–815. (4) Chen, J.; Spear, S. K.; Huddleston, J. G.; Rogers, R. D. Polyethylene glycol and solutions of polyethylene glycol as green reaction media. Green Chem. 2005, 7, 64–82. (5) Li, X.; Hou, M.; Han, B.; Wang, X.; Zou, L. Solubility of CO2 in a Choline Chloride þ Urea Eutectic Mixture. J. Chem. Eng. Data 2008, 53, 548–550. (6) Poster Program. Bio/Green Applications of Ionic Liquids. 2009. http://www.coil-3.org/media/UserFiles/File/Coil3/Poster_Programme. pdf (accessed June 26, 2009). (7) Brennecke, J. F.; Maginn, E. J. Ionic Liquids: Innovative Fluids for Chemical Processing. AIChE J. 2001, 47, 2384–2389. (8) Hermann, W.; Bosshard, P.; Hung, E.; Hunt, R.; Simon, A. J. An Assessment of Carbon Capture Technology and Research Opportunities. Stanford University Global Climate & Energy Project Technical Assessment Report 2005, 1–20. (9) Favre, E.; Bounaceur, R.; Roizard, D. A hybrid process combining oxygen enriched air combustion and membrane separation for postcombustion carbon dioxide capture. Sep. Purif. Technol. 2009, 68, 30–36. (10) Meis, N. N. A. H.; Bitter, J. H.; de Jong, K. P. Support and Size Effects of Activated Hydrotalcites for Precombustion CO2 Capture. Ind. Eng. Chem. Res. 2010, 49, 1229–1235. (11) Agarwal, A.; Biegler, L. T.; Zitney, S. E. Superstructure-based Optimal Synthesis of Pressure Swing Adsorption Cycles for Precombustion CO2 Capture. Ind. Eng. Chem. Res. 2010, 49, 5066–5079. (12) Li, X.-S.; Xia, Z.-M.; Chen, Z.-Y.; Yan, K.-F.; Li, G.; Wu, H.-J. Gas Hydrate Formation Process for Capture of Carbon Dioxide from Fuel Gas Mixture. Ind. Eng. Chem. Res. 2010, 49, 11614–11619. (13) Hesketh, H. E. Air Pollution Control; Ann Arbor Science Publishers, Inc.: MI, 1979; pp 41-89. (14) Huttenhuis, P. J. G.; Agrawal, N. J.; Hogendoorn, J. A.; Versteeg, G. F. Gas solubility of H2S and CO2 in aqueous solutions of N-methyldiethanolamine. J. Petrol. Sci. Eng. 2007, 55, 122–134. (15) Rochelle, G. T. Amine Scrubbing for CO2 Capture. Science 2009, 325, 1652–1654. (16) Zhou, S.; Chen, X.; Nguyen, T.; Voice, A. K.; Rochelle, G. T. Aqueous Ethylenediamine for CO2 Capture. ChemSusChem 2010, 3, 913–918. (17) Tobiesen, F. A.; Svendsen, H. F.; Mejdell, T. Modeling of Blast Furnace CO2 Capture Using Amine Absorbents. Ind. Eng. Chem. Res. 2007, 46, 7811–7819. (18) Huang, J.; R€uther, T. Why are Ionic Liquids Attractive for CO2 Absorption? An Overview. Aust. J. Chem. 2009, 62, 298–308. (19) Bates, E.; Mayton, R.; Ntai, I.; Davis, J. H. CO2 Capture by a Task-Specific Ionic Liquid. J. Am. Chem. Soc. 2002, 124, 926–927. (20) Zhang, Y.; Zhang, S.; Lu, X.; Zhou, Q.; Fan, W.; Zhang, X. Dual Amino-Functionalised Phosphonium Ionic Liquids for CO2 Capture. Chem.;Eur. J. 2009, 15, 3003–3011. (21) Zhang, J.; Zhang, S.; Dong, K.; Zhang, Y.; Shen, Y.; Lv, X. Supported Absorption of CO2 by Tetrabutylphosphonium Amino Acid Ionic Liquids. Chem.;Eur. J. 2006, 12, 4021–4026. (22) Gurkan, B.; de la Fuente, J.; Mindrup, E.; Ficke, L.; Goodrich, B.; Price, E.; Schneider, W.; Brennecke, J. Equimolar CO2 Absorption by Anion-Functionalized Ionic Liquids. J. Am. Chem. Soc. 2010, 132, 2116– 2117. 2829

dx.doi.org/10.1021/ie101734h |Ind. Eng. Chem. Res. 2011, 50, 2822–2830

Industrial & Engineering Chemistry Research (23) Li, X.; Hou, M.; Zhang, Z.; Han, B.; Yang, G.; Wang, X.; Zou, L. Absorption of CO2 by ionic liquid/polyethylene glycol mixture and the thermodynamic parameters. Green Chem. 2008, 10, 879–884. (24) Camper, D.; Bara, J. E.; Gin, D. L.; Noble, R. D. RoomTemperature Ionic Liquid - Amine Solutions: Tunable Solvents for Efficient and Reversible Capture of CO2. Ind. Eng. Chem. Res. 2008, 47, 8496–8498. (25) Bara, J. E.; Carlisle, T. K.; Gabriel, C. J.; Camper, D.; Finotello, A.; Gin, D. L.; Noble, R. D. Guide to CO2 Separation in Imidazolium-Based Room-Temperature Ionic Liquids. Ind. Eng. Chem. Res. 2009, 28, 2739– 2751. (26) Ramnial, T.; Taylor, S. A.; Bender, M. L.; Gorodetsky, B.; Lee, P. T. K.; Dickie, D. A.; McCollum, B. M.; Pye, C. C.; Walsby, C. J.; Clyburne, J. A. C. Carbon-Centered Strong Bases in Phosphonium Ionic Liquids. J. Org. Chem. 2008, 73, 801–812. (27) Dupont, J.; Spencer, J. On the Noninnocent Nature of 1,3Dialkylimidazolium Ionic Liquids. Angew. Chem., Int. Ed. 2004, 43, 5296–5297. (28) Nair, V.; Bindu, S.; Sreekumar, V. N-Heterocyclic Carbenes: Reagents, Not Just Ligands! Angew. Chem., Int. Ed. 2004, 43, 5130– 5135. (29) Islam, M. M.; Imase, T.; Okajima, T.; Takahashi, M.; Niikura, Y.; Kawashima, N.; Nakamura, Y.; Ohsaka, T. Stability of Superoxide Ion in Imidazolium Cation-Based Room-Temperature Ionic Liquids. J. Phys. Chem. A 2009, 113, 912–916. (30) Fraser, K. J.; MacFarlane, D. R. Phosphonium-Based Ionic Liquids: An Overview. Aust. J. Chem. 2009, 62, 309–321. (31) Ramnial, T.; Ino, D. D.; Clyburne, J. A. C. Phosphonium Ionic Liquids As Reaction Media for Strong Bases. Chem. Commun. 2005, 325–327. (32) Ramnial, T.; Hauser, M. K.; Clyburne, J. A. C. PhosphoniumBased Ionic Liquids as Efficient Borane Carriers. Aust. J. Chem. 2006, 59, 298–301. (33) Canal, J. P.; Ramnial, T.; Dickie, D. A.; Clyburne, J. A. C. From the reactivity of N-heterocyclic carbenes to new chemistry in ionic liquids. Chem. Commun. 2006, 17, 1809–1818. (34) Ramnial, T.; Taylor, S. A.; Clyburne, J. A. C.; Walsby, C. J. Grignard Reagents in Ionic Solvents: Electron Transfer Reactions and Evidence for Facile Br-Mg Exchange. Chem. Commun. 2007, 2066– 2068. (35) Ramnial, T.; Taylor, S. A.; Bender, M. L.; Gorodetsky, B.; Lee, P. T. K.; Dickie, D. A.; McCollum, B. M.; Pye, C. C.; Walsby, C. J.; Clyburne, J. A. C. Carbon-Centered Strong Bases in Phosphonium Ionic Liquids. J. Org. Chem. 2008, 73, 801–812. (36) CRC Handbook of Chemistry and Physics, 63rd ed.; CRC: Boca Raton, FL, 2002. (37) Hartono, A.; da Silva, E. F.; Grasdalen, H.; Svendsen, H. F. Qualitative Determination of Species in DETA-H2O-CO2 System Using 13 C NMR Spectra. Ind. Eng. Chem. Res. 2007, 46, 249–254. (38) Ouimet, M. A. Process for the Recovery of Carbon Dioxide from a Gas Stream. U.S. Patent 20080159937A1, July 3, 2008. (39) Aggarwal, V. K.; Emme, I.; Mereu, A. Unexpected side reactions of imidazolium-based ionic liquids in the base-catalysed Baylis-Hillman reaction. Chem. Commun. 2002, 1612–1613. (40) Bradaric, C. J.; Downard, A.; Kennedy, C.; Robertson, A. J.; Zhou, Y. Industrial preparation of phosphonium ionic liquids. Green Chem. 2003, 5, 143–152. (41) Plechkova, N. V.; Seddon, K. R. Applications of ionic liquids in the chemical industry. Chem. Soc. Rev. 2008, 37, 123–150. (42) Freemantle, M. Ionic Liquids Make Splash in Industry. Chem. Eng. News 2005, 83, 33–38. (43) Ramnial, T. M. The Chemistry of Imidazolium Salts and Phosphonium-Based Ionic Liquids. Ph.D. Thesis, Simon Fraser University, Burnaby BC, 2006. (44) Ge, R.; Hardacre, C.; Jacquemin, J.; Nancarrow, P.; Rooney, D. W. Heat Capacities of Ionic Liquids as a Function of Temperature at 0.1 MPa. Measurement and Prediction. J. Chem. Eng. Data 2008, 53, 2148–2153.

ARTICLE

(45) reference to the Holbrey et al. data as quoted in Soriano, A. N.; Agapito, A. M.; Lagumbay, L. J.; Caparanga, A. R.; Li, M. A simple approach to predict molar heat capacity of ionic liquids using groupadditivity method. J. Taiwan Inst. Chem. Eng. 2010, 41, 307–314. (46) Fredlake, C. P.; Crosthwaite, J. M.; Hert, D. G.; Aki, S. N.; Brennecke, J. F. Thermophysical Properties of Imidazolium-Based Ionic Liquids. J. Chem. Eng. Data 2004, 49, 954–964. (47) Yamamuro, O.; Minamimoto, Y.; Inamura, Y.; Hayashi, S.; Hamaguchi, H. Heat capacity and glass transition of an ionic liquid 1-butyl-3-methylimidazolium chloride. Chem. Phys. Lett. 2006, 423, 371–375. (48) Troncoso, J.; Cerdeiri~ na, C. A.; Sanmamed, Y. A.; Romaní, L.; Rebelo, L. P. Thermodynamic Properties of Imidazolium-Based Ionic Liquids: Densities, Heat Capacities, and Enthalpies of Fusion of [bmim][PF6] and [bmim][NTf2]. J. Chem. Eng. Data 2006, 51, 1856– 1859. (49) Blokhin, A. V.; Paulechka, Y. U.; Strechan, A. A.; Kabo, G. J. Physicochemical properties, structure, and conformations of 1-Butyl-3methylimidazolium bis(trifluoromethanesulfonyl)imide [C4mim]NTf2 ionic liquid. J. Phys. Chem. 2008, B112, 4357–4364. (50) Huddleston, J. G.; Visser, A. E.; Reichert, W. M.; Willauer, H. D.; Broker, G. A.; Rogers, R. D. Characterization and Comparison of Hydrophilic and Hydrophobic Room Temperature Ionic Liquids Incorporating the Imidazolium Cation. Green Chem. 2001, 3, 156–164. (51) Kilaru, P; Baker, G. A.; Scovazzo, P. Density and surface tension measurements of imidazolium-, quaternary phosphonium-, and ammonium-based room-temperature ionic liquids: data and correlations. J. Chem. Eng. Data 2007, 52, 2306–2314. (52) CYPHOS IL 103 Phosphonium Ionic Liquid. Cytec Industries Inc.: 2005. http://www.cytec.com/specialty-chemicals/PDFs/PhosphoniumSalts/CYPHOSIL103.pdf (accessed July 8, 2010). (53) Francesconi, R.; Bigi, A.; Rubini, K.; Comelli, F. Molar Heat Capacities, Densities, Viscosities, and Refractive Indices of Poly(ethylene glycols) þ 2-Methyltetrahydrofuran at (293.15, 303.15, and 313.15) K. J. Chem. Eng. Data 2007, 52, 2020–2025. (54) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Montgomery, J. A., Jr.; Vreven, T.; Kudin, K. N.; Burant, J. C.; Millam, J. M.; Iyengar, S. S.; Tomasi, J.; Barone, V.; Mennucci, B.; Cossi, M.; Scalmani, G.; Rega, N.; Petersson, G. A.; Nakatsuji, H.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Klene, M.; Li, X.; Knox, J. E.; Hratchian, H. P.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Ayala, P. Y.; Morokuma, K.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Zakrzewski, V. G.; Dapprich, S.; Daniels, A. D.; Strain, M. C.; Farkas, O.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Ortiz, J. V.; Cui, Q.; Baboul, A. G.; Clifford, S.; Cioslowski, J.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.; Challacombe, M.; Gill, P. M. W.; Johnson, B.; Chen, W.; Wong, M. W.; Gonzalez, C.; Pople, J. A. Gaussian 03, Revision D.02; Gaussian, Inc.: Wallingford, CT, 2004.

2830

dx.doi.org/10.1021/ie101734h |Ind. Eng. Chem. Res. 2011, 50, 2822–2830