High gas storage capacities for ionic liquids through chemical complexation

Article abstract of DOI:10.1021/ja077233b

Gases with sufficient Lewis acidity or basicity form weak and reversible complexes with carefully selected ionic liquids. We have prepared ionic liquid complexes with PH₃ and BF₃ that provide large gas capacities in the liquid phase

Full text of DOI:10.1021/ja077233b

Published on Web 12/19/2007
High Gas Storage Capacities for Ionic Liquids through Chemical
Complexation
Daniel J. Tempel,* Philip B. Henderson, Jeffrey R. Brzozowski, Ronald M. Pearlstein, and
Hansong Cheng
Air Products and Chemicals, Inc., 7201 Hamilton BouleVard, Allentown, PennsylVania 18195
Received September 18, 2007; E-mail: tempeldj@airproducts.com
Storing, transporting, and delivering toxic, flammable, or reactive
gases present significant technological challenges. An unintended
or uncontrollable release of a hazardous gas from a pressurized
container could result in significant harm to life and the environ-
ment. Alternatives to conventional gas processing methods that rely
on high pressures are needed for many important industrial
applications.1 The unique properties of ionic liquids (ILs)2 offer
an opportunity to develop systems capable of reversibly storing a
variety of molecular species in the liquid phase without imparting
solvent impurities. Here we report a combined experimental and
theoretical study of IL complexes for storing two toxic gases, PH3
and BF3. These gases are used as precursors for selectively doping
silicon semiconductors with phosphorus or boron atoms through
ion implantation.1b,c,3 We show that carefully selected ILs can
Figure 1. Reaction isotherms at 15, 25, 35, and 45 °C for [bmim][Cu2Cl3]
and PH3.
reversibly bind the gas molecules via chemical complexation with
unusually high capacities and moderate heats of reaction.
Most ionic liquids have no measurable vapor pressure, are easy
to process, and can be recycled and reused. In addition, anions and
cations can be selected to provide a wide range of physicochemical
properties. Most studies for gas processes using ionic liquids are
based on dissolution4 with notable exceptions for CO2 and SO2
capture utilizing amine-functionalized ionic liquids.5 Our approach
was to design IL complexes by making use of IL chemical reactivity
toward specific molecular species to drastically increase storage
capacity at low pressures. For storing PH3, we considered Lewis
Figure 2. The calculated electrostatic potentials.
-
acidic ionic liquids such as Al2Cl7 and Cu2Cl3-, which form
complexes with the gas molecule via charge transfer from the
electron lone pair of PH3 to the empty s-orbital of the metal atoms.
In contrast, for BF3, we used Lewis basic ILs to bind the Lewis
acidic gas molecule.
anion formally contains two copper atoms, so 1 mol of IL can
accommodate 2 mol of PH3.
The forward reaction isotherms for complexation of PH3 with
[bmim][Cu2Cl3] are co-incident with the reverse isotherms (Figure
1), demonstrating that complexation is completely reversible and
that the IL-PH3 mixture is stable. The reaction isotherms can be
described by a two-site thermodynamic model for formation of PH3
complexes in solution.11 The heats of reaction for the two sites
predicted by the model are -11.9 and -9.69 kcal/mol. The opti-
mized lowest energy structure of [bmim][Cu2Cl3] exhibits strong
Cu-Cu bonding with a bond length of 2.626 Å, which is elongated
slightly by 0.082 Å upon binding with PH3. The electrostatic poten-
tial of [bmim][Cu2Cl3], projected onto the electron density, indicates
that the Cu atoms are positively charged and thus facilitate binding
with the electron lone pair of PH3 (Figure 2). Each [bmim][Cu2-
Cl3] molecule binds with two PH3 molecules. The calculated Cu-P
distances are 2.199 and 2.201 Å, indicating that covalent bonds
between Cu and P are formed. The calculations yield reaction
energies for PH3 binding at the Cu sites of -12.0 and -7.9 kcal/
mol, in good agreement with the fitted experimental values.
Chemical complexation between the IL and the gas species gives
rise to an unusually high PH3 storage capacity for this material,
while the moderate heats of formation allow the reaction to be
readily reversible.
Experiments were carried out isothermally on a 2 g scale by
cascading known pressures and volumes of gas into a continuously
stirred reactor containing a known quantity of ionic liquid for gas
complexation, and the procedure was reversed to collect data for
gas evolution. The equilibrium gas capacities were determined as
mole ratios6 (moles of complexed gas per mole of IL) by using the
ideal gas law. Reaction isotherms were obtained by plotting the
capacity of the IL for the gas as a function of pressure. In parallel,
quantum chemical calculations using gradient-corrected density
functional theory were performed to understand the interactions
between the IL complexes and gas molecules. The computational
study utilized the Perdew-Wang’s exchange-correlation functional
(PW91) and a double numerical atomic basis set augmented with
polarization functions as implemented in DMol3 package.7 All
molecular structures were fully optimized. Reaction energies were
evaluated with zero-point energy corrections.8
We selected 1-butyl-3-methylimidazolium trichlorodicuprate,
[bmim][Cu2Cl3], for detailed experiments using PH3.9 Trichloro-
dicuprate ILs are readily synthesized through addition of 2 equiv
of copper(I) chloride to 1 equiv of a chloride salt precursor.10 The
9
400
J. AM. CHEM. SOC. 2008, 130, 400-401
10.1021/ja077233b CCC: $40.75 © 2008 American Chemical Society

C O M M U N I C A T I O N S
the 11BF3 concentration in the headspace of the reactor was diluted
to ∼90% as is expected for complete fluoride exchange between
-
BF4 and BF3.
Both PH3 and BF3 are readily recoverable when their complexes
with [bmim][Cu2Cl3] and [bmim][BF4] are exposed to reduced
pressure. The reactions are completely reversible, and the ILs show
no measurable loss in capacity after multiple complexation-
evolution cycles. Without external energy input, the rate of gas
evolution under vacuum is limited by mass transport; however, low
flow rates (5-10 sccm per liter of IL) can be sustained to
equilibrium pressures below 20 Torr. Thus, the systems are suitable
for storing and delivering gases for applications that require low
flow rates such as ion implantation.
In summary, we have demonstrated that the inherent properties
of ILs are ideally suited for selectively storing a large quantity of
pure gas in a small volume at low pressures. Exceptionally high
gas storage capacities for the selected gases were achieved by taking
advantage of weak chemical complexation between the gas species
and the carefully selected ILs. The reverse reaction at room
temperature allows for delivery of the gas at modest flow rates
without stirring. Liquid phase sub-atmospheric storage and delivery
offers an important advantage over conventional compressed sources
for highly reactive gases and represents an important new applica-
tion for ionic liquids.
Figure 3. Reaction isotherms at 25, 35, and 45 °C for [bmim][BF4] and
BF3.
The observed capacity of 1.92 mol/mol at 826 Torr and 15 °C
corresponds to PH3 complexation with 96% of the available copper
coordination sites. In contrast, the chemically unreactive ionic liquid
[bmim][PF6] had a capacity of only 0.03 mol/mol at atmospheric
pressure and temperature. The density difference between the neat
IL and the PH3 complex leads to an increase in volume of ∼20%
upon complexation at 25 °C and atmospheric pressure. The room
temperature 1H-decoupled 31P NMR spectrum of the [bmim][Cu2-
Cl3]/PH3 mixture exhibits a singlet at -171.9 ppm versus 85% H3-
PO4, while pure PH3 in the gas phase exhibits a singlet at -252.6
ppm.
Having shown that a Lewis basic gas can be reversibly
complexed with an IL, we extended the concept to a Lewis acidic
gas, BF3. Our computational study predicted that tetrafluoroborate
salts should be well suited for reversibly complexing Lewis acidic
BF3. The calculated minimum energy structure suggested that
[bmim][BF4] can form an adduct, [bmim][BF3-F-BF3], upon
exposure to BF3, with a heat of reaction of -9.7 kcal/mol. The
distance between B in BF3 and the bridged F is 1.635 Å, indicative
of a covalent bond. The bonding is dictated by the charge transfer
from the negatively charged F in [bmim][BF4] to the electron-
deficient B in BF3 as clearly illustrated by the calculated electrostatic
potential shown in Figure 2. Indeed, this IL complex was found to
exhibit a high capacity for BF3. On the basis of the experimental
reaction isotherms (Figure 3), the heat of reaction for binding with
BF3 was determined to be -6.76 kcal/mol. Assuming 1:1 reactivity,
91% of the IL reactive sites are complexed at 25 °C and 819 Torr.
The liquid volume increases by ∼20% under these conditions. The
BF3 capacity of [bmim][PF6] was less than 0.10 mol/mol at 25 °C
and 760 Torr.
Acknowledgment. We thank Ann Kotz for acquiring NMR
spectra.
Supporting Information Available: Coordinates of fully optimized
molecular structures. This material is available free of charge via the
Internet at http://pubs.acs.org.
References
(1) (a) Kohl, A. L.; Nielsen, R. B. Gas Purification, 5th ed.; Gulf Publishing
Company: Houston, TX, 1997. (b) Tempel, D. J.; Henderson, P. B.;
Brzozowski, J. R. U.S. Patent 7 172 646, 2007. (c) McManus, J. V.; Tom
G. M.; Kirk, R. Ion Implant. Technol. 1995, 94, 523-526.
(2) Wasserscheid, P., Welton, T., Eds. Ionic Liquids in Synthesis; Wiley-
VCH: Weinheim, Germany, 2003.
(3) May, G. S.; Sze, S. M. Fundamentals of Semiconductor Fabrication; John
Wiley & Sons, Inc.: New York, 2004; Chapter 7.
(4) (a) Blanchard, L. A.; Hancu, D.; Beckman, E. J.; Brennecke, J. F. Nature
1999, 399, 28-29. (b) Shiflett, M. B.; Yokozeki, A. Ind. Eng. Chem.
Res. 2005, 44, 4453-4464 and references therein.
(5) (a) Bates, E. D.; Mayton, R. D.; Ntai, I.; Davis, J. H., Jr. J. Am. Chem.
Soc. 2002, 124, 926-927. (b) Wu, W.; Han, B.; Gao, H.; Liu, Z.; Jiang,
T.; Huang, J. Angew. Chem., Int. Ed. 2004, 43, 2415.
(6) Fogg, P. G. T.; Gerrard, W. Solubility of Gases in Liquids; Wiley: New
York, 1991; Chapter 1.
(7) (a) Perdew, J. P.; Wang, Y., Phys. ReV. B 1992, 45, 13244. (b) DMol3;
Accelrys Software, Inc.: San Diego, CA, 2005.
(8) The present theoretical model does not account for solvation effects
important for ILs, and thus the calculated binding energies of gas species
are only semi-quantitative. Further improvement in accuracy is expected
when solvation effects are taken into account.
(9) The use of [bmim][Cu2Br3] for PH3 complexation as well as kinetic
measurements for this and other ILs will be included in future reports.
(10) (a) Axtell, D. D.; Good, B. W.; Porterfield, W. W.; Yoke, J. T. J. Am.
Chem. Soc. 1973, 95, 4555-4559. (b) Bolkan, S. A.; Yoke, J. T. J. Chem.
Eng. Data 1986, 31, 194-197.
(11) The [bmim][Cu2Cl3]/PH3 model is analogous to a two-site Langmuir model
for adsorption of gas onto a solid: (a) Langmuir, I. J. Am. Chem. Soc.
1918, 40, 1361-1403. (b) Myers, A. L. AIChE J. 1983, 29, 691-693.
(c) Mathias, P. M.; Kumar, R.; Moyer, J. D., Jr.; Schork, J. M.; Srinivasan,
S. R.; Auvil, S. R.; Talu, O. Ind. Eng. Chem. Res. 1996, 35, 2477-2483.
(12) Hartman, J. S.; Stilbs, P. J. Chem. Soc., Chem. Commun. 1975, 566-
567.
Both our experimental evidence and theoretical results suggest
that BF3 complexes with the BF4- anion through a bridged fluorine
interaction. The room temperature 19F NMR spectrum of the neat
[bmim][BF4]/BF3 adduct exhibits a sharp singlet at -145.2 ppm
(referenced to CFCl3), which is intermediate between the chemical
shifts from BF3 in the gas phase (-134.3 ppm) and the anion of
neat [bmim][BF4] (-150.6 ppm). The presence of a single 19F NMR
resonance at room temperature can be attributed to rapid exchange
of all the fluorines. Hartman and Stilbs observed the heptafluo-
rodiborate anion, B2F7-, as its tetra-n-butylammonium salt by
19F NMR spectroscopy at temperatures below -110 °C.12 Other
(13) (a) Harris, J. J. Inorg. Chem. 1966, 5, 1627-1629. (b) Brownstein, S.;
Paasivirta, J. Can. J. Chem. 1965, 43, 1645-1649. (c) Brownstein, S.
Can. J. Chem. 1967, 45, 2403-2409. (d) Fleming, S; Lupton, M. K.;
Jekot, K. Inorg. Chem. 1972, 11, 2534-2540. (e) Thomas, M. G.; Schultz,
C. W.; Parry, R. W. Inorg. Chem. 1977, 16, 994-1001. (f) Akiba, K.;
Yamashita, M.; Yamamoto, Y.; Nagase, S. J. Am. Chem. Soc. 1999, 121,
10644-10645.
-
reports citing the formation of BF3 adducts with BF4 salts have
-
been published.13 We obtained further evidence for BF3-F-BF3
bridging by adding 99.9% isotopically enriched 11BF3 to [bmim]-
[BF4] containing a natural isotopic abundance of boron (80%
11BF4-, 20% 10BF4-). Upon reaching equilibrium at ∼760 Torr,
JA077233B
9
J. AM. CHEM. SOC. VOL. 130, NO. 2, 2008 401