Design, synthesis, and evaluation of nonaqueous silylamines for efficient CO2 capture

Article abstract of DOI:10.1002/cssc.201300438

A series of silylated amines have been synthesized for use as reversible ionic liquids in the application of post-combustion carbon capture. We describe a molecular design process aimed at influencing industrially relevant carbon capture properties, such as viscosity, temperature of reversal, and enthalpy of regeneration, while maximizing the overall CO₂-capture capacity. A strong structure-property relationship among the silylamines is demonstrated in which minor structural modifications lead to significant changes in the bulk properties of the reversible ionic liquid formed from reaction with CO ₂.

Full text of DOI:10.1002/cssc.201300438

CHEMSUSCHEM
FULL PAPERS
DOI: 10.1002/cssc.201300438
Design, Synthesis, and Evaluation of Nonaqueous
Silylamines for Efficient CO2 Capture
Jackson R. Switzer,[a] Amy L. Ethier,[a] Emily C. Hart,[b] Kyle M. Flack,[b] Amber C. Rumple,[b]
Jordan C. Donaldson,[a] Ashley T. Bembry,[a] Owen M. Scott,[a] Elizabeth J. Biddinger,[a]
Manish Talreja,[a] Myoung-Geun Song,[b] Pamela Pollet,[b] Charles A. Eckert,[a, b] and
Charles L. Liotta*[a, b]
A series of silylated amines have been synthesized for use as
reversible ionic liquids in the application of post-combustion
carbon capture. We describe a molecular design process aimed
at influencing industrially relevant carbon capture properties,
such as viscosity, temperature of reversal, and enthalpy of re-
generation, while maximizing the overall CO2-capture capacity.
A strong structure–property relationship among the silylamines
is demonstrated in which minor structural modifications lead
to significant changes in the bulk properties of the reversible
ionic liquid formed from reaction with CO2.
Introduction
Despite efforts to develop alternative, carbon-neutral energy
sources, it is projected that coal will remain the predominant
fuel for electricity generation until 2035.[1,2] Coal-fired power
plants produce nearly a third of all CO2 emissions in the United
States, which represent large point sources of CO2 emissions.[3]
With the growing recognition that global CO2 emissions must
be reduced to mitigate their harmful environmental impact,
there is significant interest in the development of post-com-
bustion CO2-capture technologies that can be retrofitted to ex-
isting power plants.[4]
system must operate in a cost-effective manner and minimize
the overall energy required for capture and release.[2]
Reversible ionic liquids have been developed as a new class
of switchable solvents.[8] Our recent work has focused on rever-
sible ionic liquids derived from silylamines.
An example of the basic silylamine structure and its reaction
with CO2 is shown in Scheme 1. Upon reaction with CO2, silyla-
Several gas separation technologies applicable to CO2 cap-
ture are in use today; their feasibility and cost-effectiveness are
highly dependent on capture conditions (e.g., temperature,
pressure, concentration) and purity specifications at release.[4,5]
One of the most mature technologies is the use of an absorp-
tion-based process. In general, viable absorption-based, post-
combustion CO2-capture processes must meet three criteria.
First, the liquid sorbent must be selective for CO2, which com-
prises only a portion of post-combustion coal-fired power
plant flue gas (approximately 16% by volume).[6,7] Second, the
system must also be recyclable; the amount of CO2 released
by just one coal-fired power plant in a year of operation
makes capture by a single-use sorbent unfeasible.[6] Finally, the
Scheme 1. Example trialkylsilamine and its reaction equilibrium with CO2.
mines reversibly form an ammonium carbamate salt, which is
liquid at or near room temperature; the salt is referred to as
the reversible ionic liquid. On heating, the reversible ionic
liquid releases the captured CO2 to reform the silylamine.
We previously demonstrated the use of this type of solvent
system in coupling reactions and separations[9] as well as in
crude oil separation.[10] Most recently, we have shown how re-
versible ionic liquids can be used as a unique class of CO2-ab-
sorption solvents.[11] In addition, we have shown that the rever-
sible ionic liquid is further capable of physically absorbing ad-
ditional CO2.[11] This dual method of capture provides enhanced
CO2 uptake under industrially relevant operating conditions.[12]
Here we describe how we have used iterative molecular
design to optimize the structures of silylamines for CO2 capture
and show how the variation of the length of the tether that
connects the amine to the Si atom and changing the position
of or the addition of a single methyl group to the silylamine
backbone can influence capture properties.
[a] J. R. Switzer, A. L. Ethier, J. C. Donaldson, A. T. Bembry, O. M. Scott,
E. J. Biddinger, M. Talreja, Prof. C. A. Eckert, Prof. C. L. Liotta
School of Chemical and Biomolecular Engineering
Georgia Institute of Technology
311 Ferst Street, Atlanta, GA 30332 (USA)
[b] E. C. Hart, K. M. Flack, A. C. Rumple, M.-G. Song, P. Pollet, Prof. C. A. Eckert,
Prof. C. L. Liotta
School of Chemistry and Biochemistry
Georgia Institute of Technology
911 Atlantic Drive, Atlanta, GA 30332 (USA)
E-mail: charles.liotta@chemistry.gatech.edu
We present 13 silylated amines (Table 1) as potential candi-
dates for an absorption-based CO2-capture process and show
Supporting Information for this article is available on the WWW under
http://dx.doi.org/10.1002/cssc.201300438.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2014, 7, 299 – 307 299

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
TEtSMA
Table 1. Silylamine structures, names, abbreviations, and gravimetric CO2 uptake capacities after reaction with
1 bar CO2 at 258C.
TEtSMA was synthesized through
a three-step procedure as high-
lighted in Scheme 2. The first
step involved the Grignard reac-
tion of trichloro(chloromethyl)si-
lane with ethyl magnesium chlo-
ride to form triethylchlorome-
thylsilane. The triethylsilyl phtha-
limidomethane intermediate was
prepared by reaction of triethyl-
chloromethylsilane with potassi-
um phthalimide and subse-
quently reacted with hydrazine
to afford TEtSMA in an overall
27% isolated yield.
Silylamine
Abbreviation
CO2 uptake
À1]
À1
[molCO molamine
[molCO kgamine
]
2
2
variation of alkyl chain length between Si atom and amine
TEtSMA
0.59Æ0.01
0.59Æ0.01
0.63Æ0.01
0.35Æ0.03
4.09Æ0.08
3.67Æ0.03
3.66Æ0.04
1.9Æ0.2
1-(aminomethyl)triethylsilane
TEtSEtA
2-(aminoethyl)triethylsilane
TEtSA
3-(aminopropyl)triethylsilane
TEtSBA
4-(aminobutyl)triethylsilane
branching along the alkyl chain backbone
aMe-TEtSA
0.63Æ0.03
3.36Æ0.15
4-(triethylsilyl)butyl-2-amine
a,aDMe-TEtSA
bMe-TEtSA
0.48Æ0.03
0.59Æ0.01
2.4Æ0.2
2-methyl-4-(triethylsilyl)butyl-2-amine
TEtSEtA
3.15Æ0.07
2-methyl-3-(triethylsilyl)propylamine
TEtSEtA was prepared as shown
in Scheme 3. Triethylsilane was
reacted with N-vinylphthalimide
in the presence of platinum(0)-
1,3-divinyl-1,1,3,3-tetramethyldi-
siloxane (Pt-DVDS) to yield
phthalimide-protected 2-(amino-
ethyl)triethylsilane. The inter-
mediate was then reacted with
hydrazine to yield TEtSEtA in
25% isolated yield.
unsaturation in the alkyl chain backbone
trans-TEtSA
0.61Æ0.03
0.47Æ0.03
3.56Æ0.20
2.35Æ0.16
trans-3-(triethylsilyl)prop-2-en-1-amine
trans-a,aDMe-TEtSA
trans-2-methyl-4-(triethylsilyl)butyl-2-amine
trans-a,aDMe-TPSA
0.40Æ0.000
1.65Æ0.000
trans-2-methyl-4-(tripropylsilyl)butyl-2-amine
secondary silylamines
STEtSA
0.65Æ0.01
0.62Æ0.02
3.49Æ0.05
3.87Æ0.11
N-methyl-3-(triethylsilyl)propan-1-amine
SDMESA
TEtSA, DMESA, and TEtSBA
N-methyl-3-(aminopropyl)dimethylethylsilane
TEtSA and DMESA were pre-
pared by the hydrosilylation of
allyl amine and the correspond-
ing silane (dimethylethylsilane or
triethylsilylane, respectively) with
additional
DMESA
0.42Æ0.04
2.85Æ0.24
3-(aminopropyl)dimethylethylsilane
how, despite having similar molecular weights, the properties
of these 13 silylated amines differ substantially. The properties
presented include CO2-uptake capacity, reversal temperature,
enthalpy of regeneration, and reversible ionic liquid viscosity.
Results and Discussion
Synthesis
The silylamines were prepared as follows. Detailed experimen-
tal procedures and characterization of the silylamines are in-
cluded in the Supporting Information. Yields and synthesis pro-
cedures were not optimized as the goal was to evaluate the
properties for CO2 capture of the silylamines to establish struc-
ture–property relationships.
Scheme 2. Synthesis of TEtSMA.
Pt-DVDS in toluene at 1108C to yield TEtSA (89% isolated
yield) and DMESA (58% isolated yield) as shown in Scheme 4.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2014, 7, 299 – 307 300

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
a,aDMe-TEtSA
Commercially available a,a-dimethylpropargylamine was react-
ed at room temperature with triethylsilane in the presence of
Pt-DVDS and phosphatrane in THF (Scheme 7). The product,
trans-2-methyl-4-(triethylsilyl)butyl-2-amine, was distilled after
the evaporation of THF. The compound was then reduced
under hydrogenation conditions in a Parr autoclave reactor
with Pd on carbon in hexane at 508C to afford a,aDMe-TEtSA
in an 80% isolated yield.
Scheme 3. Synthesis of TEtSEtA.
Scheme 4. Synthesis of TEtSA (R=CH2CH3) and DMESA (R=CH3).
TEtSBA was formed by the hydrosilylation of 4-amino-1-butene
with Pt-DVDS in toluene at 808C (Scheme 5; 66% yield).
Scheme 7. Synthesis of a,aDMe-TEtSA.
bMe-TEtSA
The b-methyl derivative of TEtSA, bMe-TEtSA, was synthesized
through the hydrosilylation of commercially available 2-methyl-
allylamine (Scheme 8, 76% isolated yield).
Scheme 5. Synthesis of TEtSBA.
aMe-TEtSA
The synthesis of aMe-TEtSA followed a Gabriel synthesis path-
way from 3-chloro-1-butene (Scheme 6). 3-Chloro-1-butene
was reacted with potassium phthalimide and K2CO3 in DMF.
The resulting compound, 3-phthalimido-1-butene, was then re-
acted with Pt-DVDS and triethylsilane in toluene at reflux.
Deprotection was achieved by using anhydrous hydrazine in
anhydrous methanol at 608C to yield aMe-TEtSA in a 67% iso-
lated yield.
Scheme 8. Synthesis of bMe-TEtSA.
trans-TEtSA, trans-a,aDMe-TEtSA, and trans-a,aDMe-TPSA
The general reaction scheme for the synthesis of trans-TEtSA,
trans-a,aDMe-TEtSA, and trans-a,aDMe-TPSA is shown in
Scheme 9. For trans-TEtSA, propargylamine was reacted with
triethylsilane at room temperature in the presence of Pt-DVDS
in THF to provide the product in a 46% isolated yield. For
trans-a,aDMe-TEtSA and trans-a,aDMe-TPSA, a,a-dimethylpro-
pargylamine was reacted at room temperature with the corre-
sponding silane (triethylsilane or tripropylsilane, respectively)
in the presence of Pt-DVDS and phosphatrane in THF to yield
the desired silylamine. trans-a,aDMe-TEtSA and trans-a,aDMe-
TPSA were obtained in 80 and 68% isolated yield, respectively.
Scheme 9. Synthesis of trans-TEtSA (R’=H; R’’=CH2CH3), trans-a,aDMe-
TEtSA (R’=CH3; R’’=CH2CH3), and trans-a,aDMe-TPSA (R’=CH3;
R’’=CH2CH2CH3).
Scheme 6. Synthesis of aMe-TEtSA.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2014, 7, 299 – 307 301

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
STEtSA and SDMESA
theoretical CO2 capacity for the chemical reaction was in fact
0.67 moles of CO2 per mole of amine. Several of the silylamines
(TEtSA, aMe-TEtSA, STEtSA, and SDMESA) discussed herein ap-
proach this limit (Table 1).
The synthetic strategy for the formation monomethylated
TEtSA and DMESA is shown in Scheme 10. TEtSA or DMESA
were reacted with ethyl formate under neat conditions in an
Ar atmosphere at 608C. The amide was then reacted with
LiAlH4 in THF (1m) in THF at 608C to yield STEtSA (73% isolat-
ed yield) or SDMESA (49% isolated yield).
Variation of the distance between the amine and the Si
atom does not adversely affect the CO2 capacity until the
tether is increased to a butyl chain (Table 1). TEtSBA forms
a low-melting solid immediately upon reaction with CO2; the
comparatively lower equilibrium CO2 uptake (i.e., incomplete
reaction) observed is likely the result of occlusion of the silyla-
mine within the solid carbamate salt. Otherwise, the variation
of the linker chain from methylene to ethylene and propylene
did not alter the CO2 uptake significantly; a deviation of only
0.02 moles of CO2 per mole of amine is observed between
TEtSMA, TEtSEtA, and TEtSA.
Compared to TEtSA, the inclusion of a single methyl group
in the a position to the amine (aMe-TEtSA) does not affect the
CO2 capacity; both TEtSA and aMe-TEtSA show a CO2 uptake
of 0.63 moles of CO2 per mole of amine. However, the intro-
duction of a second methyl group at the a position (a,aDMe-
TEtSA) decreases the CO2 capacity on a molar basis by 27%.
The decreased CO2 uptake is a result of the increased steric
hindrance at the reactive amine site, which lowers the stability
of the carbamate species formed upon reaction with CO2. In-
terestingly, the inclusion of a methyl group in the b position
relative to the amine (bMe-TEtSA) causes a slight decrease in
the CO2 capacity; bMe-TEtSA has a capacity of 0.59 moles of
CO2 per mole of amine compared to 0.63 moles of CO2 per
mole of amine for TEtSA.
Scheme 10. Synthesis of STEtSA (R=CH2CH3) and SDMESA (R=CH3).
CO2 capacity
The equilibrium CO2 uptake of each silylamine under 1 bar of
CO2 at 258C was measured by gravimetry (Table 1). The CO2
uptake is reported both in moles of CO2 per mole of amine as
well as moles of CO2 per kilogram of solvent. The former is
useful to discern stoichiometric effects, whereas the latter is
more useful to judge industrial viability compared to other sol-
vent systems.
Unsaturation along the alkyl chain backbone between the Si
atom and amine has little effect on the CO2 uptake compared
to its saturated analogue. Unsaturated trans-TEtSA exhibits
a CO2 capacity of (0.61Æ0.03) moles of CO2 per mole of amine,
whereas saturated TEtSA exhibits a similar CO2 capacity of
(0.63Æ0.01) moles of CO2 per mole of amine. Unsaturated
On average, the silylamines summarized in Table 1 exhibit an
equilibrium CO2 capacity of 0.59 moles of CO2 per mole of
amine. This combines CO2 absorption from the chemical reac-
tion of CO2 with the silylamine and physical absorption of CO2
into the newly formed ionic liquid. Previous silylamine studies
have shown that under 1 bar of CO2 at 258C, physical absorp-
tion contributes approximately 0.01–0.02 moles of CO2 per
mole of amine to the overall CO2 capacity [0.59 moles of CO2
per mole of 3-(aminopropyl)tripropylsilane) (TPSA)].[13] The ma-
jority of the CO2 uptake is the result of the chemical reaction
of CO2 with the silylamine. The conventional stoichiometric
CO2 capacity of an amine that reacts with CO2 to form ammo-
nium-carbamate ion pairs is 0.5 moles of CO2 per mole of
amine. However, our silylamines exhibit an enhanced CO2 ca-
pacity. An in-depth investigation of the chemical reaction of
CO2 with the silylamines has shown that a stabilized carbamic
acid species is formed amidst a network of ammonium-carba-
mate ion pairs (Figure 1).[13] The results established that the
trans-a,aDMe-TEtSA exhibits
a
CO2 capacity of (0.47Æ
0.03) moles of CO2 per mole of amine, whereas saturated
a,aDMe-TEtSA shows a CO2 capacity of (0.46Æ0.01) moles of
CO2 per mole of amine. An additional unsaturated silylamine
with a longer alkyl chain around the Si atom, trans-a,aDMe-
TPSA, was also investigated and its CO2 uptake was slightly
lower than that of the other unsaturated silylamines at
0.40 moles of CO2 per mole of amine.
The secondary silylamine STEtSA exhibits a CO2-uptake ca-
pacity of (0.65Æ0.01) moles of CO2 per mole of amine, which
is on a par with that of its primary amine analogue TEtSA
((0.63Æ0.01) moles of CO2 per mole of amine). The secondary
silylamine SDMESA exhibits a significantly higher CO2-uptake
capacity (0.62 moles of CO2 per mole of amine) than its pri-
mary amine counterpart, which reached a CO2 uptake of only
0.42 moles of CO2 per mole of amine. DMESA immediately
forms a solid upon reaction with CO2, whereas SDMESA does
not.
The CO2 uptake at 408C was investigated for a select group
of silylamines to evaluate the effect of temperature on the CO2
absorption (Table 2). An absorption temperature of 408C is rep-
resentative of the flue-gas temperature of a post-combustion
Figure 1. Carbamic acid stabilized by the ammonium carbamate ion of the
TEtSA reversible ionic liquid.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2014, 7, 299 – 307 302

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
The introduction of a single methyl group, a (aMe-TEtSA) or
b (bMe-TEtSA), to the amine results in a reduction of the rever-
sal temperature of 19 and 148C, respectively. The introduction
of two methyl groups a to the amine results in a further reduc-
tion in the reversal temperature as shown for a,aDMe-TEtSA
(438C). As capture conditions dictate a flue-gas temperature of
408C, a reversal temperature at or below 408C would result in
incomplete conversion to the reversible ionic liquid. Although
the reversal temperature of a,aDMe-TEtSA is low, it evidences
our ability to affect the reversible ionic liquid properties signifi-
cantly by altering the silylamine structure.
Table 2. CO2 uptake capacity at 408C for a select group of silylamines.
Silylamine
CO2 uptake at 408C
À1
À1
[molCO molamine
]
[molCO kgamine
]
2
2
TEtSA
a,aDMe-TetSA
bMe-TetSA
0.603Æ0.002
0.46Æ0.02
3.48Æ0.01
2.29Æ0.08
3.19Æ0.01
0.599Æ0.003
stream in a coal-fired power plant. Of the three silylamines in-
vestigated at 408C, only TEtSA showed a (modest) drop of 5%
in CO2 capacity at this elevated temperature.
The reversal temperatures of reversible ionic liquids with an
unsaturated propyl chain (trans-TEtSA, trans-a,aDMe-TEtSA,
trans-a,aDMe-TPSA) between the amine and Si atom were also
investigated. A mixture of cis and trans unsaturated silylamines
were initially proposed, however, only the trans isomers were
synthesized. These unsaturated reversible ionic liquids show re-
versal temperatures that are over 208C less than those of their
saturated counterparts. We postulate that this must be related
to the entropic effect that results from the inflexible, locked
conformation of the unsaturated reversible ionic liquids.
The reversal temperature of the secondary amines decreases
significantly (30, 378C) compared to that of TEtSA (718C).
Unlike the branched amines, both STEtSA and SDEMSA exhibit
capacities equal to the corresponding primary amine at 258C.
However, the capture conditions must be carefully balanced
because these silylamines may not reach full conversion to the
reversible ionic liquid at 408C.
Reversal temperature
The reversal temperature is the point at which the reversible
ionic liquid releases CO2 and reverts back to the silylamine.
We determined the reversal temperature of the reversible ionic
liquids experimentally by using differential scanning calorime-
try (DSC). The temperatures of the reversal events were deter-
mined by the intersection of the baseline of the event and the
tangent to the peak of the event (see Supporting Information
for an example thermogram).
The lower the temperature of reversal, the smaller the
amount of energy required to heat the reversible ionic liquid
from the capture temperature to the point at which the CO2 is
released. Of the structural modifications investigated, the
effect of branching along the alkyl chain backbone, unsatura-
tion of the propyl backbone, and the order of the amine (18 or
28) showed the strongest influence on reversal temperature
(Table 3). The reversal temperature of TEtSA is included as
a baseline. These experimental results are in agreement with
the trends predicted previously for the reversal temperatures
of these ionic liquids by the quantum-chemical approach
COSMO-RS.[12]
Enthalpy of Regeneration
The enthalpy of regeneration is the primary contributor to the
overall energy required to release absorbed CO2 and regener-
ate the silylamine. The structural modifications presented here
show both positive and negative deviations from the enthalpy
of regeneration of the baseline silylamine TEtSA (Table 4).
The enthalpies of regeneration for the reversible ionic liquids
were calculated from the DSC thermograms. Heat supplied to
We examined the effect of the alkyl tether between the Si
atom and the amine and found that decreasing the alkyl chain
length from three C atoms (TEtSA) to two C atoms (TEtSEtA)
results in the formation of a solid that melts at 498C. The re-
versal temperature increases by almost 408C. Unlike TEtSEtA,
the additional reversible ionic liquids presented in Table 3
show a decrease in the reversal temperature from that of the
TEtSA base comparison.
the system [kJmolCO À1] was calculated based on the gravimet-
2
ric uptake of CO2.
An increase of the length of the alkyl tether between the
amine and Si atom (TEtSBA) results in a significant increase in
the heat of regeneration (152 kJmolCO À1) with an en-
2
thalpy of regeneration almost double that of TEtSA
Table 3. Reversal temperature of reversible ionic liquids influenced by structural
modifications.
(83 kJmolCO À1). This may be a result of intramolecular
2
interactions, which result in the stabilization of the
ammonium-carbamate and carbamic acid species.[13]
As presented in the CO2 capacity section, the intro-
duction of two methyl groups in the a position to
the amine (a,aDMe-TEtSA) results in the incomplete
conversion of the silylamine to the reversible ionic
liquid. This is likely because of steric hindrance,
which causes destabilization of the carbamate ion.[14]
As a result, the enthalpy of regeneration should be
lower, although it appears to be higher than expect-
Silylamine
Reversal temp-
erature [8C]
Silylamine
Reversal temp-
erature [8C]
TEtSMA
TEtSA
aMe-TetSA
bMe-TetSA
trans-a,aDMe-TEtSA
STEtSA
78Æ5
71Æ3
52Æ2
57Æ7
11[c]
TEtSEtA
TEtSBA
a,aDMe-TEtSA
trans-TEtSA
trans-a,aDMe-TPSA[d]
SDMESA
109Æ3[a]
84Æ4[b]
41Æ6
48Æ2
15Æ2
30Æ3
37Æ2
DMESA
80Æ8[e]
[a] Formed solid (m.p. (49Æ1)8C). [b] Formed solid [m.p. (66Æ4)8C]. [c] Reversal tem-
perature difficult to measure because of broad events in the DSC thermogram.
[d] Bottom of reversal curve at 488C. [e] Formed solid [m.p. (66Æ3)8C].
ed (114 kJmolCO À1) compared to TEtSA. However, be-
2
cause of the decreased equilibrium CO2 uptake of
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2014, 7, 299 – 307 303

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
does not follow the viscosity trend of the other rever-
sible ionic liquids. The trend exhibited by TEtSMA,
TEtSA, and TEtSBA may be explained by the proximi-
ty to the Lewis acidic Si atom.
Table 4. Enthalpy of regeneration of reversible ionic liquids influenced by structural
modifications.[a]
Silylamine
Heat of regeneration Silylamine
Heat of regeneration
À1
À1
[kJmolCO
]
[kJmolCO
]
2
2
The branching along the alkyl chain backbone of
the silylamine has a significant effect on viscosity.
With an added methyl group in the a position to the
amine (aMe-TEtSA), the reversible ionic liquid viscosi-
ty increases slightly (6915 vs. 6088 cP). However,
when two methyl groups are in the a position
(a,aDMe-TEtSA), the viscosity decreases to 1257 cP at
258C. The decrease in viscosity is likely because of
the decreased CO2 uptake. Unreacted silylamine (vis-
cosity<100 cP) could still be present. A 1000 cP de-
crease is seen in the reversible ionic liquid viscosity
TEtSMA
TetSA
aMe-TEtSA
bMe-TEtSA
trans-a,aDMe-TetSA
STEtSA
76Æ7
83Æ6
90Æ5
89Æ8
7[d]
TEtSEtA
TEtSBA
a,aDMe-TEtSA
trans-TEtSA
91Æ16[b]
152Æ19[c]
114Æ16
85Æ7
trans-a,aDMe-TPSA[e] 21Æ6
77Æ6
SDMESA
76Æ2
DMESA
130Æ14[f]
[a] Heat of regeneration of reversible ionic liquids that formed solids includes both
melting and CO2-release DSC events. [b] Formed solid [m.p. (49Æ1)8C]. [c] Formed
solid [m.p. (66Æ4)8C]. [d] Enthalpy difficult to measure because of broad events in
the DSC thermogram. [e] Bottom of reversal curve at 488C. [f] Formed solid [m.p.
(66Æ3)8C].
a,aDMe-TEtSA, the enthalpy cannot be directly compared to
Table 5. Reversible ionic liquid viscosity of the silylamines after reaction
with 1 bar CO2 measured at 25 and 408C.
that of its unbranched analogue (TEtSA); the enthalpies are
calculated based on their respective equilibrium CO2 uptake.
The unsaturated amines, with branching in the a position to
the amine (trans- a,aDMe-TEtSA, trans- a,aDMe-TPSA) show
a decrease in the enthalpy of regeneration of 18 and 628C, re-
spectively.
Silylamine
Viscosity [cP]
408C
258C
variation of the alkyl chain length between the Si atom and amine
TEtSMA
TEtSEtA
TEtSA
2373Æ206
solid
6088Æ36
solid
625Æ58
solid
1303Æ100
solid
TEtSBA
Reversible ionic liquid viscosity
branching along the alkyl chain backbone
The viscosities of the reversible ionic liquid species formed
after reaction with 1 bar of CO2 at 258C (at equilibrium conver-
sions) are reported in Table 5. We recognize that some of the
viscosities of the reversible ionic liquids reported at 258C are
not industrially viable. Nevertheless, several important points
must be emphasized: 1) viscosities at equilibrium conversion
at 258C are important to characterize the structure–property
relationships of the reversible ionic liquids, 2) industrial pro-
cesses are usually conducted at temperatures higher than
258C (commonly 408C), which further reduces the viscosity of
the reversible ionic liquid, and 3) in the real world it is not nec-
essary for the CO2 uptake to go to the full equilibrium CO2
uptake, which allows an opportunity to control the viscosity of
the system (see below).
aMe-TEtSA
a,aDMe-TEtSA
bMe-TEtSA
6915Æ431
725Æ164
125Æ22
1257Æ338
5075Æ141
1084Æ62
unsaturation in the alkyl chain backbone
trans-TEtSA
trans-a,aDMe-TEtSA
trans-a,aDMe-TPSA
3889Æ252
362Æ19
<100[a]
<100[a]
<100[a]
<100[a]
secondary silylamines
STEtSA
SDMESA
135Æ14
117Æ16
<100[a]
<100[a]
additional
DMESA
solid
solid
[a] Measurements below 100 cP are below the reliable detection limits of
the instrument used.
The results given in Table 5 demonstrate that the manipula-
tion of the silylamine structure can affect the reversible ionic
liquid viscosity. The reversible ionic liquid viscosities reported
here vary from <100 to over 6000 cP at 258C.
The length of the alkyl tether between the amine and Si
atom has a significant effect on the reversible ionic liquid vis-
cosity. With only a methylene linker, the reversible ionic liquid
viscosity is in the order of 2000 cP, whereas the reversible ionic
liquid forms a solid with the longest linker (butylene). We an-
ticipate that at intermediate lengths (ethylene and propylene)
a trend of increasing reversible ionic liquid viscosity would be
observed.
of bMe-TEtSA compared to that of TEtSA.
Unsaturation in the alkyl chain backbone causes a 36% de-
crease in the reversible ionic liquid viscosity at 258C for trans-
TEtSA compared to the saturated reversible ionic liquid coun-
terpart. A significant reversible ionic liquid viscosity decrease is
also seen in trans-a,aDMe-TEtSA (<100 cP) versus its saturated
analogue a,aDMe-TEtSA (1257Æ338 cP).[15] We found a similarly
low reversible ionic liquid viscosity for trans-a,aDMe-TPSA. No-
tably, these two silylamines have a lower equilibrium CO2 ca-
pacity than the other silylamines as they do not reach their full
theoretical CO2 capacity (0.67 moles of CO2 per mole of amine)
at 258C.
However, although TEtSA, which has a propylene linker, falls
within the viscosity range of TEtSMA and TEtSBA, the reversible
ionic liquid that has an ethylene linker, TEtSEtA, forms a solid
after complete reaction with CO2. It is not clear why TEtSEtA
Although the secondary silylamine STEtSA does not exhibit
a decreased CO2 capacity relative to its TEtSA analogue, it has
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2014, 7, 299 – 307 304

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
a markedly lower viscosity. At (135Æ14) cP, the reversible ionic
liquid viscosity of STEtSA is nearly 98% lower than that of
TEtSA. Similarly, although DMESA forms a solid upon reaction
with CO2, its secondary analogue, SDMESA has an ionic viscosi-
ty of only (117Æ16) cP.
At 408C, the viscosity of the reversible ionic liquid drops by
over 80% on average (excluding those for which measure-
ments are <100 cP or form solids). For example, at 258C, the
viscosity of TEtSMA is (2373Æ206) cP, whereas the viscosity is
only (625Æ58) cP at 408C. As long as the reversal temperature
is above 408C, a significant decrease in CO2 capacity is not
expected.
Figure 3. Viscosity versus CO2 uptake for a,aDMe-TEtSA at 258C.
Control of the reversible ionic liquid viscosity
A unique feature of the reversible ionic liquid systems dis-
cussed herein is that the viscosity varies with CO2 uptake fol-
lowing a hockey-stick-like relationship (Scheme 10). For 3-(ami-
nopropyl)tripropylsilane (TPSA), the viscosity of the partially
converted solvent (which contains both silylamine and reversi-
ble ionic liquid) is 620 cP at a CO2 uptake of 0.45 moles of CO2
per mole of amine at 258C.[11] However, with only a slight in-
crease in CO2 uptake to 0.50 moles of CO2 per mole of amine,
the viscosity increases by over 50% to 1528 cP.
deviates slightly with a sharper rise in viscosity. For example, at
0.45 moles of CO2 per mole of amine, the viscosity of a,aDMe-
TEtSA is nearly 1500 cP, whereas TPSA does not reach a similar
viscosity until 0.50 moles of CO2 per mole of amine.
Conclusions
A systematic process to design CO2-capture solvents that ach-
ieve a high CO2 uptake—while minimizing viscosity and the
energy required for reversal—has been presented. The influ-
ence of the length of the alkyl tether as well as the addition or
change in the position of one or more methyl groups in the si-
lylamine structure has a significant effect on the pertinent CO2-
capture properties of the reversible ionic liquids. 3-(Aminopro-
pyl)triethylsilane (TEtSA) was used as a baseline silylamine to
which structural modifications were made to establish struc-
ture–property relationships.
This curve is even more valuable for the results obtained at
408C (Figure 2). Compared to the 258C curve, it is immediately
apparent that at 408C higher CO2 uptakes can be achieved at
lower viscosities. For example, at 258C with 0.45 moles of CO2
per mole of amine, the viscosity of TPSA is 620 cP. If measured
at 408C, a viscosity of 620 cP is not reached until an uptake
that nears 0.52 moles of CO2 per mole of amine. Notably, at
a CO2 uptake of 0.39 moles of CO2 per mole of amine, the vis-
cosity is only 144 cP—a viscosity that nears industrial viability.
A plot of CO2 uptake versus viscosity for the branched
amine a,aDMe-TEtSA at 258C is shown in Figure 3. Although
a,aDMe-TEtSA does not reach the same CO2 uptake as TPSA at
258C, the curves are similar in shape. The a branching affects
only the end point; although the equilibrium CO2 uptake is
lower, the viscosity exhibits behavior similar to that of TPSA
prior to the end point. At higher conversions, a,aDMe-TEtSA
The CO2 capacity was determined experimentally both in
terms of moles of CO2 per mole of silylamine as well as the
more industrially relevant term of moles of CO2 uptake per
kilogram of silylamine. The influence of the length of the
tether between the Si atom and the amine on CO2 capacity is
evident as silylamines with methylene and propylene tethers
remain liquids in their ionic state and nearly achieve the maxi-
mum theoretical CO2 uptake (0.67 moles of CO2 per mole of
amine), whereas the silylamine with a butylene linker immedi-
ately forms a solid upon reaction with CO2. The energy re-
quired to regenerate the silylamine and release CO2 is compa-
rable for 1-(aminomethyl)triethylsilane (TEtSMA) and TEtSA, but
the viscosity of TEtSMA is less than half of that of TEtSA. The
length of the alkyl silylamine backbone has a definite influence
on capacity, energy required for reversal, and viscosity.
The branching along the alkyl chain backbone was also in-
vestigated. The data show that the introduction of a methyl
group in the a position to the amine (4-(triethylsilyl)butyl-2-
amine; aMe-TEtSA) results in a reduction of 208C in the rever-
sal temperature. Furthermore, the addition of a second methyl
group in the a position to the amine (2-methyl-4-(triethylsilyl)-
butyl-2-amine; a,aDMe-TEtSA) decreases the viscosity of the
reversible ionic liquid to approximately a third of that of TEtSA,
although the CO2 uptake is affected. The effect of branching is
Figure 2. Viscosity versus CO2 uptake for TPSA at 25 and 408C.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2014, 7, 299 – 307 305

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
further emphasized if the alkyl
chain backbone is unsaturated.
trans-2-Methyl-4-(triethylsilyl)bu-
Table 6. Summary of the properties of TEtSMA and trans-TEtSA
Property
Unit
tyl-2-amine
(trans-a,aDMe-
TEtSMA
trans-TEtSA
TEtSA) and trans-2-methyl-4-(tri-
propylsilyl)butyl-2-amine (trans-
a,aDMe-TPSA) have viscosities
of less than 100 cP in part be-
cause of reversal temperatures
below 308C. The removal of the
branching in the unsaturated
À1
CO2 uptake at 258C
CO2 uptake at 258C
reversal temperature
enthalpy of regeneration
reversible ionic liquid viscosity at 258C
reversible ionic liquid viscosity at 408C
[molCO molamine
]
0.59Æ0.01
4.09Æ0.08
78Æ5
0.61Æ0.03
3.56Æ0.20
48Æ2
2
À1
[molCO kgamine
]
2
[8C]
[kJmolCO
[cP]
À1
]
76Æ7
85Æ7
2
2373Æ206
625Æ58
3889Æ252
362Æ19
[cP]
backbone yields
a reversible
tube, porosity C) was weighed and used to sparge dry carbon di-
oxide through the silylamine for 75 min at flow rate of
ionic liquid with a viscosity slightly higher than that of TEtSA
but with a reversal temperature over 208C lower than that of
the saturated species.
a
200 mLminÀ1 at 258C and 1 bar. The resulting reversible ionic
liquid, vial, and fritted tube were then weighed to determine the
CO2 uptake capacity at 258C. To determine the CO2 uptake capaci-
ty at 408C, the same procedure was followed except that the vial
and sample were placed in a 408C sand bath in thermal equilibri-
um with a heated gas line.
There is a strong effect of the order of the amine (18 and 28)
on the reversible ionic liquid properties as evidenced by the
addition of a methyl group on the amine functionality (N-
methyl-3-(aminopropyl)dimethylethylsilane, SDMESA). A 50-fold
decrease in viscosity is observed for this reversible ionic liquid
as well as a reversal temperature that is reduced below 408C
compared to that of TEtSA.
Not only are we able to control the properties of the reversi-
ble ionic liquids by change of position or addition of methyl
groups along the silylamine tether, but we can also control the
viscosity by limiting the conversion to the reversible ionic
liquid. The viscosity can be maintained below 300 cP while still
able to achieve approximately 70% conversion to the reversi-
ble ionic liquid at 408C. This behavior is not only unique to
this system, but also serves as an additional parameter that
may be used to meet criteria for industrial implementation.
Although the selection of a particular CO2-capture solvent is
dependent on the specifications of its end use application, we
have identified two silylamine solvents for further investiga-
tion, TEtSMA and trans-3-(triethylsilyl)prop-2-en-1-amine (trans-
TEtSA), the properties of which are summarized in Table 6.
They both exhibit high CO2-uptake capacities relative to the
other silylamines studied, viable reversible temperatures, and
reasonable enthalpies of regeneration. Their reversible ionic
liquids have lower viscosities than most other silylamine rever-
sible ionic liquids, and furthermore, by controlling the CO2
uptake, the viscosity could be reduced to meet industry-oper-
ating specifications.
Instrumentation
Attenuated total reflectance (ATR) FTIR data were collected by
using a Shimadzu IRPrestige21 spectrometer with a DLaTGS detec-
tor with 32 scans and a resolution of 1 cmÀ1 used in combination
with a Specac, Ltd., heated Golden Gate ATR accessory with dia-
mond crystal and zinc selenide lenses.[11,16]
1H and 13C NMR spectra were recorded by using a Varian Mercury
Vx 400 instrument using CDCl3 as the lock solvent. As a result of
the high viscosities of the ionic forms, the reversible ionic liquids
were synthesized in the NMR tube under solvent-free conditions,
and a capillary tube that contained CDCl3 was placed inside the
sample.
The viscosities of all of the reversible ionic liquids were measured
by using a Rheosys Merlin II viscometer with a temperature-con-
trolled 28 cone and plate system. The densities of all of the silyla-
mines and reversible ionic liquids (used in the gravimetric calcula-
tion of CO2 uptake) were measured in triplicate by using an Anton
Paar DMA 38 Laboratory Density Meter. Elemental analysis (C, N, H)
of molecular liquids was performed at Atlantic Microlab in Nor-
cross, GA. Samples were stored under inert gas and either deliv-
ered in person or sent by FedEx.
DSC measurements were used to determine the reversal tempera-
tures of the reversible ionic liquids and the evaporation tempera-
tures of the silylamines. The instrument used was a Q20 TA Instru-
ments DSC; hermetically sealed aluminum pans were used, into
which the compound under consideration (ꢀ2 mg) was measured.
The temperature was ramped from À40 to 4008C at a rate of
58CminÀ1, and the temperatures of the reversal or boiling events
were determined by the intersect of the baseline of the event and
the tangent to the peak of the event.
Experimental Section
Synthesis
The materials, experimental procedures, and the characterization
for the synthesis of the silylamines are included in the Supporting
Information.
Reversible Ionic Liquid Formation
For all of the reversible ionic liquids discussed, synthesis was per-
formed as follows unless otherwise noted. A 2 dram vial was
weighed; the silylamine (ꢀ1 g) was added and the vial was weigh-
ed again. A porous fritted tube (Ace Glass, 10 mm gas dispersion
Acknowledgements
We thank Phillips 66 for their financial support of this research.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2014, 7, 299 – 307 306

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
[9] R. Hart, P. Pollet, D. J. Hahne, E. John, V. Llopis-Mestre, V. Blasucci, H.
Huttenhower, W. Leitner, C. A. Eckert, C. L. Liotta, Tetrahedron 2010, 66,
1082–1090.
Keywords: amines · carbon · environmental chemistry · ionic
liquids · structure–property relationships
[10] V. Blasucci, C. Dilek, H. Huttenhower, E. John, V. Llopis-Mestre, P. Pollet,
C. A. Eckert, C. L. Liotta, Chem. Commun. 2009, 116–118; V. Blasucci, R.
Hart, V. L. Mestre, D. J. Hahne, M. Burlager, H. Huttenhower, B. J. R. Thio,
P. Pollet, C. L. Liotta, C. A. Eckert, Fuel 2010, 89, 1315–1319.
[11] A. L. Rohan, J. R. Switzer, K. M. Flack, R. J. Hart, S. Sivaswamy, E. J. Bid-
dinger, M. Talreja, M. Verma, S. Faltermeier, P. T. Nielsen, P. Pollet, G. F.
Schuette, C. A. Eckert, C. L. Liotta, ChemSusChem 2012, 5, 2181–2187.
[12] M. Gonzalez-Miquel, M. Talreja, A. L. Ethier, K. Flack, J. R. Switzer, E. J.
Biddinger, P. Pollet, J. Palomar, F. Rodriguez, C. A. Eckert, C. L. Liotta, Ind.
Eng. Chem. Res. 2012, 51, 16066–16073.
[1] U.S. Energy Information Administration, Washington DC, 2011.
[2] U.S. Energy Information Administration, Washington DC, 2000.
[3] A. A. Olajire, Energy 2010, 35, 2610–2628.
[4] E. S. Rubin, H. Mantripragada, A. Marks, P. Versteeg, J. Kitchin, Prog.
Energy Combust. Sci. 2012, 38, 630–671; A. B. Rao, E. S. Rubin, Environ.
Sci. Technol. 2002, 36, 4467–4475.
[5] G. Sartori, D. W. Savage, Ind. Eng. Chem. Fundam. 1983, 22, 239–249.
[6] D. M. D’Alessandro, B. Smit, J. R. Long, Angew. Chem. 2010, 122, 6194–
6219; Angew. Chem. Int. Ed. 2010, 49, 6058–6082.
[13] J. R. Switzer, A. L. Ethier, K. M. Flack, E. J. Biddinger, L. Gelbaum, P. Pollet,
C. A. Eckert, C. L. Liotta, Ind. Eng. Chem. Res. 2013, 52, 13159–13163.
[14] T. Mimura, H. Simayoshi, T. Suda, M. Iijima, S. Mituoka, Energy Convers.
Manage. 1997, 38, S57–S62.
[15] We could not accurately measure viscosities lower than 100 cP with our
apparatus, but for the purpose of this paper it is sufficient to know
they were below 100 cP (industrially viable limit).
[7] K. S. Fisher, K. Searcy, G. T. Rochelle, S. Ziaii, C. Shubert, Pittsburgh, PA,
Advanced Amine Solvent Formulations and Process Integration for Near-
Term CO2 Capture Success, DOE Final Report (Grant DE-FG02-
06ER84625), 2007.
[8] P. G. Jessop, D. J. Heldebrant, X. W. Li, C. A. Eckert, C. L. Liotta, Nature
2005, 436, 1102–1102; L. Phan, D. Chiu, D. J. Heldebrant, H. Huttenhow-
er, E. John, X. W. Li, P. Pollet, R. Y. Wang, C. A. Eckert, C. L. Liotta, P. G.
Jessop, Ind. Eng. Chem. Res. 2008, 47, 539–545; L. Phan, J. R. Andreatta,
L. K. Horvey, C. F. Edie, A.-L. Luco, A. Mirchandani, D. J. Darensbourg,
P. G. Jessop, J. Org. Chem. 2008, 73, 127–132; B. E. Gurkan, J. C. de La
Fuente, E. M. Mindrup, L. E. Ficke, B. F. Goodrich, E. A. Price, W. F.
Schneider, J. F. Brennecke, J. Am. Chem. Soc. 2010, 132, 2116–2117.
[16] S. G. Kazarian, B. J. Briscoe, T. Welton, Chem. Commun. 2000, 2047–
2048.
Received: May 7, 2013
Published online on November 7, 2013
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2014, 7, 299 – 307 307