Reinventing Design Principles for Developing Low-Viscosity Carbon Dioxide-Binding Organic Liquids for Flue Gas Clean Up

Article abstract of DOI:10.1002/cssc.201601622

Anthropogenic CO₂emissions from point sources (e.g., coal fired-power plants) account for the majority of the greenhouse gases in the atmosphere. Water-lean solvent systems such as CO₂-binding organic liquids (CO₂BOLs) are being developed to reduce the energy requirement for CO₂capture. Many water-lean solvents such as CO₂BOLs are currently limited by the high viscosities of concentrated electrolyte solvents, thus many of these solvents have yet to move toward commercialization. Conventional standard trial-and-error approaches for viscosity reduction, while effective, are time consuming and economically expensive. We rethink the metrics and design principles of low-viscosity CO₂-capture solvents using a combined synthesis and computational modeling approach. We critically study the effects of viscosity reducing factors such as orientation of hydrogen bonding, introduction of higher degrees of freedom, and cation or anion charge solvation, and assess whether or how each factor affects viscosity of CO₂BOL CO₂capture solvents. Ultimately, we found that hydrogen bond orientation and strength is the predominant factor influencing the viscosity in CO₂BOL solvents. With this knowledge, a new CO₂BOL variant, 1-MEIPADM-2-BOL, was synthesized and tested, resulting in a solvent that is approximately 60 % less viscous at 25 mol % CO₂loading than our base compound 1-IPADM-2-BOL. The insights gained from the current study redefine the fundamental concepts and understanding of what influences viscosity in concentrated organic CO₂-capture solvents.

Full text of DOI:10.1002/cssc.201601622

Accepted Article
Title: Reinventing design principles for developing low-viscosity carbon
dioxide binding organic liquids (CO2BOLs) for flue gas clean up
Authors: Phillip Koech, Deepika Malhotra, David J Heldebrant, David C
Cantu, Feng Zheng, Vassiliki-Alexandra Glezakou, and Roger
Rousseau
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: ChemSusChem 10.1002/cssc.201601622
Link to VoR: http://dx.doi.org/10.1002/cssc.201601622
A Journal of
www.chemsuschem.org

ChemSusChem
10.1002/cssc.201601622
FULL PAPER
Reinventing design principles for developing low-viscosity
carbon dioxide binding organic liquids (CO BOLs) for flue gas
2
clean up
a
a
a
b
a
Deepika Malhotra, Phillip K. Koech,* David J. Heldebrant, David C. Cantu, Feng Zheng, Vassiliki-
b
b
Alexandra Glezakou, Roger Rousseau
Dedication ((optional))
Abstract: Anthropogenic carbon dioxide (CO
2
) emissions from point
subsequent storage or utilization is critical to mitigate any
adverse environmental effects. Aqueous amine solvent
technology is at the forefront of CO capture from flue gas.
2
sources, such as coal fired-power plants, account for the majority of
the greenhouse gases in the atmosphere. Capture and storage or
utilization are required to mitigate adverse environmental effects.
However, the energy loss associated with boiling and
condensing water to regenerate the solvent system currently
Aqueous amine-based CO
2
capture solvents are currently
1
considered the industry standard, but deployment to market is
limited by their high regeneration energy. In that context, water-lean
limits its deployment to the market. Various water-lean solvent
systems designed to minimize water content, including task-
2
,3
4
solvent systems such as CO
being developed to reduce this energy requirement for CO
Many water-lean solvents such as CO BOLs are currently limited by
2
binding organic liquids (CO
2
BOLs) are
specific ionic liquids (TSILs),
phase change materials,
5
6,7
2
capture.
nanomaterial organic hybrids, aminosilicones, and siloxylated
8,9
2
amines, have been developed as alternatives to the energy-
intensive aqueous-amine solvents. Although the chemical
reactivity of each of these water-lean solvent classes is different,
they are projected to be drop-in replacements for aqueous
the high viscosities of concentrated electrolyte solvents, thus many
of these solvents have yet to move towards commercialization.
Historically, viscosity reduction has been achieved by systematic
methods such as introduction of steric hindrance on the anion to
minimize the intermolecular cation-anion interactions, fine-tuning the
electronics, hydrogen bonding orientation and strength, and charge
solvation. Conventional standard trial-and-error approaches, while
effective, are time consuming and economically expensive. We
10
solvents, yet operate in presence of little or no water. Water-
lean solvent systems have been projected to offer ~40%
reductions in energy requirements compared to aqueous amines
or other water-based solvents due to lower specific heat
11,12
capacities and lower water content.
In the past decade, our
rethink the metrics and design principles of low-viscosity CO
2
group has focused on development of the water-lean solvent
13,14
capture solvents using a combined synthesis and computational
modeling approach. We critically study the effects of viscosity
reducing factors such as orientation of hydrogen bonding,
introduction of higher degrees of freedom, and cation or anion
charge solvation, and assess whether or how each factor affects
system known as CO
CO BOLs are predicated on Jessop’s switchable solvents that
utilize a nonionic base (amidines or guanidines) with nonionic
aliphatic alcohols. CO BOLs chemically fixate CO as
zwitterionic amidium or guanidinium alkyl carbonate liquids.
These CO BOL solvents have shown excellent CO uptake
capacity, CO mass transfer comparable to that of aqueous
solvents, and favorable thermodynamic properties, albeit at
higher solution viscosities. Viscosities of CO BOLs and other
water-lean solvents have been shown to increase exponentially
2 2
binding organic liquids (CO BOLs).
2
2
2
13,14
viscosity of CO
hydrogen bond orientation and strength is the predominant factor
influencing the viscosity in CO BOL solvents. With this knowledge, a
new CO BOL variant, 1-MEIPADM-2-BOL, was synthesized and
tested, resulting in a solvent that is approximately 60% less viscous
at 25 mol% CO loading than our base compound 1-IPADM-2-BOL.
2
BOL CO
2
capture solvents. Ultimately, we found that
2
2
2
2
2
2
4
,7,8,12,14-16
2
with CO
2
loading.
While thermodynamically favorable,
-loaded CO BOLs (~500 cP) currently
The insights gained from the current study redefine the fundamental
concepts and understanding of what influences viscosity in
the high viscosities of CO
prevents them from being economical. Similarly, other water-
lean solvents such as aprotic heterocyclic anions (AHAs) with
2
2
12
2
concentrated organic CO capture solvents.
17-19
viscosities of ~140 cP,
azole-based TSILs with viscosities of
2
0,21
6
50 cP
1000 cP
and hydroxylated amidines (HAAMS) at viscosities
are likely to be uneconomical until solution
22
>
Introduction
viscosities can be substantially reduced.
We recently performed a comprehensive study on 1-((1,3-
dimethylimidazolidin-2-ylidene)amino)propan-2-ol (1-IPADM-2-
BOL) to measure physical, thermodynamic, and kinetic property
data. Additionally, Aspen Plus™ modeling and bench-scale
testing were performed for both thermal and polarity swing
Carbon dioxide (CO
as power plants, petrochemical industries, and refineries and
2
) capture from anthropogenic sources such
[a]
EnergyProcesses and Materials Division,Pacific Northwest
National Laboratory,Richland,WA99352,USA, E-mail:
phillip.koech@pnnl.gov
10
assisted regeneration (PSAR) CO
2
release modes. This study
revealed that 1-IPADM-2-BOL has the potential to be an
[b]
Physical Sciences Division,Pacific NorthwestNational Laboratory,
Richland,WA99352,USA
energetically viable CO
could be reduced.
2
capture solvent, assuming viscosities
Supportinginformation for this article is given via a link at the end of
the document.
This article is protected by copyright. All rights reserved.

ChemSusChem
10.1002/cssc.201601622
FULL PAPER
To address this challenge, our efforts are currently focused on
We have previously shown that a critical indicator of the high
23
not just designing less-viscous solvent systems for CO
but to learn how to make less-viscous solvents at the molecular
level while retaining adequate CO capture capacity and low
2
capture,
viscosity of CO
2
BOLs is the hydrogen bonding network.
Hydrogen bonding and viscosity correlations in ionic liquids have
24
2
also been discussed at length. TSILs that possess N or O
atoms capable of intramolecular hydrogen bonding have
specific heat capacity and regeneration energy. Historically, a
trial-and-error approach would be employed. As in our own
previous work, such approaches using organic chemistry
significantly lower CO
2
-rich viscosity compared to conventional
25
ILs. Upon CO
2
capture, CO BOLs form a zwitterion capable of
2
14
intuition have been shown to be effective, though it was never
clear why or how conventional viscosity-reducing structural
motifs actually affected viscosity. The solvent development
process can be accelerated using high throughput computational
screening of solvent viscosities from a library of compounds to
down select candidates for synthesis and testing. In this study,
we apply an integrated computationally aided design, synthesis,
and characterization efforts to assess the effects of molecular
design principles that are commonly used to reduce viscosity in
liquids. Each molecule consisting of the motif to be evaluated is
modeled, synthesized, and measured, after which a critical
assessment is made of not just whether, but how viscosity is
affected in solution. Ultimately it is shown that each motif affects
viscosity to some degree, but not through the molecular-level
interactions they were designed for. Here, we rethink and
simplify the design principles for viscosity reduction in organic
forming intra- or intermolecular hydrogen bonds, as shown in
Figure 2. In recent computational studies, we have determined
that a high population of molecules engaged in intramolecular H-
bonding accounts for lower viscosities, while a high population of
23,26
intermolecular H-bonds leads to higher viscosities.
Therefore,
understanding the effect of structural modifications on the inter-
/intramolecular hydrogen bonding, and thereby on viscosity, is a
critical aspect of our molecular design strategy.
2
CO capture solvents.
Figure 2. Inter- vs. intramolecular hydrogen bonding forms in CO2BOLs.
Results and Discussion
23,26
In our recent computational studies,
principle that the molecular structure can be modified to increase
intramolecular H-bonding in CO -bound molecules that will in
we showed proof-of-
Our current 1-IPADM-2-BOL molecule was used as a template
to apply different functionalities in this study (Figure 1). A few
2
23
molecular design criteria were kept to ensure adequate CO
uptake and durability. First, the highly basic guanidine moiety
2
turn induce viscosity reductions. Here, two guanidine-based
CO BOL variants (1-PADM-2-BOL and 1-IPADM-2-BOL)
2
was retained, as it has been shown to capture >90% CO
2
from a
bearing linear and branched alkyl alcohol, respectively, in the
side chain, were chosen as case study to validate the theory-
(
15% CO ) gas stream with acceptable solvent recirculation
2
11
rates. Second, the cyclic guanidine base core was retained to
prevent hydrolysis of the reactive guanidinium moiety of the
CO -bound CO BOL. From this template, viscosity modifying
2 2
23
generated hypothesis. Internal H-bonding was indeed found to
be critical in reducing the viscosity: 1-IPADM-2-BOL is at least
15
five times less viscous than 1-PADM-2-BOL at CO
mol% (Figure 3). This can be explained by the fact that 34% of
CO -bound 1-IPADM-2-BOL displayed an internal H-binding
2
loading of 25
motifs could be introduced to study the effects of maximizing or
minimizing steric hindrance to prevent stacking, minimizing
intermolecular cation/anion interactions, or fine-tuning electronic
effects of the anion.
2
23
conformation, compared to less than 1% for 1-PADM-2-BOL.
Based on this finding, our molecular design strategy was to
retain the high degree of internal hydrogen bonding while
systematically evaluating each viscosity-reducing factor. Thus,
the final design criteria were to retain the 2˚ alcohol moiety of the
1
-IPADM-2-BOL template.
Figure 1. Generic structure representing structural motifs to reduce viscosity
of CO BOL molecules.
2
This article is protected by copyright. All rights reserved.

ChemSusChem
10.1002/cssc.201601622
FULL PAPER
2
Figure 3. Viscosity of 1-PADM-2-BOL and 1-IPADM-2-BOL at CO loading of
CO
2
uptake capacity was found to be 8.7 wt% for 1-IPADM-3-
2
5 mol%.
BOL (Table 1), which was slightly lower than that of 1-IPADM-2-
14
BOL at 9.0 wt% which is attributed to the higher molecular
mass of the former. The uptake capacity for CO BOLs is less
than 50 mol% mainly due increased viscosity which limits CO
mass transfer. The CO -rich 1-IPADM-3-BOL was found to be
visually less viscous than 1-IPADM-2-BOL according to our
preliminary falling-ball viscosity measurements (Supporting
information-S1). However, formal viscosity measurements using
our custom ΔPVT (pressure volume temperature) cell (Figure 4)
showed that the viscosity of 1-IPADM-3-BOL was higher at all
loadings than that of 1-IPADM-2-BOL. The viscosity of 1-
2
Once incorporated, the individual structural motifs in Figure 1
can be systematically studied to determine their impacts on
viscosity. They are referred to as “viscosity-reducing factors” in
the following discussion.
2
2
2
Viscosity-Reducing Factor 1: Steric Tuning of CO BOLs
Intermolecular cation-anion interactions have been hypothesized
to be a factor that contributes to high viscosities of concentrated
2
organic electrolyte solutions such as CO BOLs. It is believed
that minimizing the ion pair interactions between different
molecules could reduce the viscosity of the solvent. Our first
strategy to investigate the viability of minimizing ion pair
interactions was to break molecular symmetry by introducing
different conformers to reduce stacking.
To investigate the potential effect of conformational flexibility of
the guanidinium cation on viscosity, we designed 1-IPADM-3-
BOL, which contains a six-membered tetrahydropyrimidine ring
core while retaining the essential component of a high degree of
IPADM-3-BOL was calculated at 237 cP using a reduced-order
26
2
model for CO BOL viscosity , comparable to the experimentally
measured value of 318 cP. 1-IPADM-3-BOL is more viscous
than 1-IPADM-2-BOL, with a calculated viscosity of 164 cP, and
an experimentally measured on of 171 cP (Table 1).
Computation and experiment indicate that this viscosity-reducing
factor, increase in conformational flexibility, did not have the
desired effect and may not be a viable strategy to reduce
viscosity.
internal hydrogen bonding (Scheme 1). The CO
2
-bound form of
1
-IPADM-3-BOL has conformational flexibility as it can flip
27
between two chair conformations, leading to more degrees of
freedom than the five-membered ring in 1-IPADM-2-BOL. The
addition of boat and chair conformers was hypothesized to
reduce the stacking ability of the molecule, thus minimizing ion
pair association, which would translate to a lower viscosity
solvent (Scheme 1).
We incorporated this motif, synthesizing 1-IPADM-3-BOL using
our standard Vilsmeier salt protocol i.e., condensation of 1-
14
aminopropan-2-ol with Vilsmeier salt
structure of 1-IPADM-3-BOL was confirmed with H and
NMR characterization with the diagnostic peak of guanidine
carbon (C=N) at 161.7 ppm. The CO -bound substrate displayed
a downfield chemical shift for –CH O- from 59.7 ppm to 63.3
2 (Scheme 1). The
1
13
C
2
2
14
ppm, which coincided with our earlier observations. However,
an up-field shift was reported for the signal corresponding to -
C=N- from 161.7 ppm to 159.0 ppm, which could be attributed to
the shielding effect induced by the conformational changes in
2
the CO -bound 1-IPADM-3-BOL.
2
Figure 4. Viscosity vs. loading correlation for selected CO BOLs. 1-
TMSIPADM-2-BOL was nottested because ofits high viscosity.
It was initially unclear why 1-IPADM-3-BOL with higher degrees
of flexibility would have a higher viscosity than 1-IPADM-2-BOL
since both molecules are structurally similar. We turned to
molecular dynamics (MD) simulations to determine the root
cause of this anomaly. Our simulations clearly showed that the
percentage of CO
bonding at 25% mol CO
1%, compared to the 34% for 1-IPADM-2-BOL. Thus, it was
2
-bound molecules with internal hydrogen
Scheme 1. Synthetic methodology and conformational flexibility in CO
2
bound
loading (Pint) for 1-IPADM-3-BOL was
1
-IPADM-3-BOL.
2
2
concluded that the increase in viscosity was due to the reduction
in the Pint and that the introduction of conformational flexibility
The CO
gravimetrically by bubbling CO
2
uptake of 1-IPADM-3-BOL was confirmed
through neat liquid at 25 °C. The
2
This article is protected by copyright. All rights reserved.

ChemSusChem
10.1002/cssc.201601622
FULL PAPER
plays a minimal role (if any) compared to the orientation and
speciation of hydrogen bonding in this molecule.
temperature (Scheme 2). The trimethylsilyl chlorohydrin 5 was
reacted with sodium azide to furnish the corresponding azido-
alcohol 6, which underwent reduction catalyzed by palladium on
carbon to generate silylamino alcohol 7. The condensation of
silylamino alcohol 7 with Vilsmeier salt 8 furnished 8a (1-
TMSIPADM-2-BOL).
Scheme 2.Synthetic methodologyfor the synthesis of1-TMSIPADM-2-BOL.
The gravimetric CO
evaluated by bubbling CO
gave a capacity of 5.5 wt%. This low CO
2
uptake ability of 1-TMSIPADM-2-BOL was
into a neat sample at 25 °C, which
uptake capacity can
2
2
be attributed to destabilized or unbalanced charges of the
zwitterion species, which favor the reverse reaction over the
2
formation of the CO -bound product. This result underscores the
delicate balance that needs to be achieved in order to find an
optimal material with both high CO2 capture capacity and low
Table 1. Gravimetric weight capacities, Pint, and viscosities for selected
CO BOLs (*ND indicates notdetermined).
2
2 2
CO -rich solvent viscosity. The CO bound 1-TMSIPADM-2-BOL
was found to be comparable in viscosity to 1-IPADM-2-BOL. The
carbon dioxide capture ability of 1-TMSIPADM-2-BOL was
Viscosity-Reducing Factor 2a: Anion Charge Solvation
Another viscosity reducing factor typically employed in solvent
development is charge solvation. Weakening the cation-anion
interactions by promoting either intra- or intermolecular
interactions with other structural motifs may reduce the viscosity
of concentrated organic solvents. We postulated that this could
be achieved in two ways: anion (carboxylate) or cation
1
13
further evaluated by both H and C NMR spectroscopy. The
CO -free CO BOLs’ signals are consistent with those previously
observed for alkanolguanidines with the guanidine carbon at
58.0 ppm. The characteristic peak for the CO -bound 1-
TMSIPADM-2-BOL had downfield shift for the signal
2
2
1
2
a
corresponding to -C=N- from 158.0 to 161.8 ppm, which is
14
consistent with our previous report.
(
guanidinium) charge solvation. The first could be achieved via
Preliminary viscosity measurements using
a
falling-ball
addition of a pendant highly oxophilic silane group to coordinate
to oxygen in the alkylcarbonate anion. A second approach would
be to introduce ether motifs that could promote solvation of
cations through coordination by unpaired electrons in ether
viscometer showed that 1-TMSIPADM-2-BOL was initially less
viscous than 1-IPADM-2-BOL but the sample solidified at higher
loadings (Figure S1 in supporting information). This observation
was confirmed in that the initial low lean solvent viscosities that
were predicted with the reduced model (143 cP Table 1). Figure
28,29
groups.
In order to validate this hypothesis, we introduced
structural motifs that were believed to promote either cation or
anion charge solvation.
5
shows the radial distribution function of the interaction of the
oxygen in the alkylcarbonate anion with either the guanidinium
cation or the silicon in the TMS moiety. Here it can be seen that
the carboxylate anion preferentially interacts with the
guanidinium cation at ~2 Å and there is negligible interaction
with the silicon at 5.5 Å. This observation is consistent with the
calculated values of higher Pint for 1-TMSIPADM-2-BOL than 1-
IPADM-2-BOL.
Ultimately, anion charge solvation may not be contributing
towards viscosity reduction and internal hydrogen bonding is still
the predominant factor in controlling the viscosity of CO
With that goal, we designed silane-based CO
TMSIPADM-2-BOL, with an oxophilic silane moiety (Scheme 2,
a). The rationale behind introducing the silane group into
CO BOLs is based on the fact that silicon is highly oxophilic and
2
BOL 1-
8
2
may form a dative bond with the carbonate anion. The dative
bond formation could thereby weaken the intermolecular cation-
anion interactions and lead to reduced solvent viscosity. Our
synthesis of 1-TMSIPADM-2-BOL involved the addition of
trimethylsilylmethylmagnesium chloride 3 to chloroacetaldehyde
2
BOLs. It
is unclear why 1-TMSIPADM-2-BOL solidified at higher CO
2
4
to generate chlorotrimethylsilylpropan-2-ol
5
at room
This article is protected by copyright. All rights reserved.

ChemSusChem
10.1002/cssc.201601622
FULL PAPER
loadings, but regardless, this phase change suggests that this
compound would not be viable in a solvent process and as such
was not tested for formal viscosity measurements on the ΔPVT
cell.
Scheme 3.Synthetic scheme for 1-MEIPADM-2-BOL.
The CO
gravimetrically, yielding a capture capacity of 7.3 wt% (Table 1).
Visual observations showed that the CO -bound 1-MEIPADM-2-
2
uptake for 1-MEIPADM-2-BOL was confirmed
2
BOL was significantly less viscous than 1-IPADM-2-BOL.
Falling-ball viscometer and ΔPVT cell measurements
corroborated our visual observations of reduced viscosity of
2
CO -rich 1-MEIPADM-2-BOL solvent. 1-MEIPADM-2-BOL was
found to be approximately 60% less viscous than our base
compound 1-IPADM-2-BOL at higher loadings (Figure 4),
indicating the ethereal moiety did provide a substantial decrease
2
in viscosity. The reduced model also confirmed a lower CO -lean
solvent viscosity (99 cP calculated, 75 cP measured) compared
to 1-IPADM-2-BOL (164 cP calculated, 171 cP measured).
Our initial conclusion was to attribute the decrease in viscosity to
the proposed cation charge solvation, though this was not the
case. The radial distribution coefficient for 1-MEIPADM-2-BOL
obtained from molecular dynamics simulations (Figure 6)
showed negligible intramolecular or intermolecular guanidinium
interactions with the ether moiety, but preferential interaction
between the alkylcarbonate anion with the guanidine moiety at
Figure 5. Proposed alkylcarbonate-silane interaction in 1-TMSIPADM-2-BOL
and plotof radial distribution vs.distance.
~
1.8 Å. These results indicate that charge solvation did not exist
Viscosity-Reducing Factor 2b: Cation Charge Solvation
Conversely, we explored the viability of cation charge solvation.
We envisioned that the introduction of ether functionalities could
solvate positive charges by interaction with unpaired electrons
on the –O- moiety. Ethereal groups such as those found in
crown ethers have been shown to reduce solvent viscosity. To
test this concept, we designed 1-MEIPADM-2-BOL consisting of
a methyl ether group proximal to the alcohol moiety (Scheme 4).
and was not likely to play a significant role in viscosity.
The Pint value for 1-MEIPADM-2-BOL was calculated to be 52%
from MD simulations, which is significantly higher than the 34%
of 1-IPADM-2-BOL (Table 1). These simulations confirm that the
viscosity reduction could not be attributed to any charge
solvation and that viscosity in these solvents is mainly attributed
to the hydrogen bonding in solution.
29
Synthesis of 1-MEIPADM-2-BOL follows
a similar reaction
sequence as for 1-TMSIPADM-2-BOL, which begins with
displacement of the chloride group of the commercially available
1
-chloro-3-methoxypropanan-2-ol 9 using sodium azide to afford
azido-alcohol 10 (Scheme 3). The azido-alcohol 10 is then
reduced using a palladium-on-carbon catalyst and hydrogen gas
to afford the amino alcohol 11, which was condensed with
Vilsmeier salt 8 to yield 1-MEIPADM-2-BOL. The structure of 1-
1
13
MEIPADM-2-BOL was successfully characterized by H and
C
NMR, and the CO -bound form of 1-MEIPADM-2-BOL exhibited
2
downfield shift for -C=N- signal from 157.8 ppm to 160.1 ppm,
which is in accordance with our previous observations.
This article is protected by copyright. All rights reserved.

ChemSusChem
10.1002/cssc.201601622
FULL PAPER
1
13
Experimental Details – General: H and C NMR spectra were
recorded at 500 MHz and 126 MHz (or 400 MHz and 101 MHz),
respectively, using CDCl
3
as a solvent. The chemical shifts are
1 13
reported in δ (ppm) values ( H and C NMR relative to CHCl
3
, δ
.26 ppm for H NMR and δ 77.0 ppm for C NMR; multiplicities
are indicated by s (singlet), d (doublet), t (triplet), sex (sextet), dd
doublet of doublets), m (multiplet) and br (broad). Coupling
1
13
7
(
constants (J) are reported in Hertz (Hz). All reagents and
solvents were employed without further purification. The
products were purified using a flash chromatography system or
a regular gravity glass column. Thin-layer chromatography (TLC)
was developed on silica gel 60G F254 glass plates.
26
Experimental details for ΔPVT Cell : For viscosity
measurements: CO BOL liquid viscosity as a function of CO
2
2
loading was measured using a custom flow cell developed in-
house for simultaneous measurements of vapor/liquid
equilibrium and physical properties of nonaqueous solvents
requiring less than 50 cc sample size.
quantitative injections of CO gas (Oxarc 99.99%) into the flow
cell was performed at sub-ambient pressures. The amount of
CO absorption was determined from temperature and pressure
measurements and the ideal gas equation of state. The liquid
CO loading was determined from the mass balance. A pump
A sequence of
2
2
2
loop and an internal gas/liquid contactor were used to provide
mixing and facilitate the approach to absorption equilibrium. An
in-line viscosity sensor (Cambridge Viscosity, SPC-372 sensor
head and VISCOpro 2000 viscometer) was used to obtain real-
time viscosity data as the liquid sample was circulated inside the
flow cell. The cell was made of stainless steel high vacuum
components with an internal vapor volume of about 150 cc. The
pump loop consisted of a gear pump (Micropump GA series)
and a custom wetted wall stainless steel contactor (1.2712.4
cm) inside the cell. The cell pressure was measured using
temperature-controlled capacitance diaphragm gauges with an
accuracy of 0.2% (Nor-Cal CDG-100 series). The cell
temperature was controlled within 0.1°C by a large circulating
water bath. The operation range of the flow cell was 20°C-100°C
and vacuum to 1 bar.
Figure 6. Intramolecular interaction in 1-MEIPADM-2-BOL and plot of radial
distribution vs.distance.
Conclusions
In this study, we systematically incorporated conventional
structural motifs that are commonly used to reduce viscosity in
concentrated
organic
electrolyte
solvents.
Synthesis,
experimental testing and molecular simulation have shown that
introduction of moieties that promote higher degrees of freedom
or cation or anion charge solvation provides negligible
contributions to viscosity for CO
found that hydrogen bond orientation and strength were the
main parameters that influenced viscosity in CO BOL solvents.
In this study, a new CO BOL variant 1-MEIPADM-2-BOL that is
2 2
BOL CO capture solvents. We
In a typical flow cell experiment, a known amount of CO
2
-free
2
CO BOL liquid sample was loaded into the cell under nitrogen
2
2
protection and was allowed to reach an initial steady state at the
approximately 60% less viscous than our base compound 1-
IPADM-2-BOL was discovered. The viscosity reduction was
found to be simply due to an increase in intramolecular
hydrogen bonding and not any of the viscosity reducing factors
that we had set out to assess. Ultimately, this study highlights
that the design criteria for lowering viscosity of water-lean
testing temperature under the liquid’s own vapor pressure. A
small amount of CO2 was injected quantitatively from
a
calibrated volume. The cell pressure was allowed to come to a
new steady state under constant temperature. The change in
liquid CO
change and the known cell volume. The steady-state viscosity,
equilibrium pressure, and liquid CO loading were recorded.
Additional CO injections were done to repeat the above process
until a target CO loading was reached. The liquid sample was
2
loading was calculated based on the cell pressure
2
2
CO BOL solvents can be simplified and redefined to focus
2
mainly on promoting a high degree of intramolecular hydrogen
bonding. As a consequence, we are now working toward
designing molecules that have an even higher degree of internal
2
regenerated in situ by raising the water bath temperature to
80°C and reducing the cell pressure using a vacuum pump
(Edwards RV8). The operation of the cell was validated using a
2
hydrogen bonding to continue to reduce the viscosity of CO BOL
solvents. We believe that this new understanding of molecular-
level interactions that govern viscosity will be applicable to other
CO BOL solvent0,11with published viscosity and vapor/liquid
2
1
equilibrium data.
water-lean CO
2
capture solvents and will enable viscosity
2
Computation: Classical MD simulations of CO BOL solvents at
reductions in all solvent systems. These studies are the current
focus of our group.
2
5 mol% CO
2
loadings were done, as described in our previous
23,26
work,
to obtain the percentage of CO -bound molecules with
2
an internal hydrogen bond (Pint) values, and radial distribution
functions. Viscosities were calculated using the reduced model
ExperimentalSection
This article is protected by copyright. All rights reserved.

ChemSusChem
10.1002/cssc.201601622
FULL PAPER
26
(
10) Mathias, P. M.; Zheng, F.; Heldebrant, D. J.; Zwoster, A.; Whyatt,
presented in our recent work using Pint values obtained from
classical MD.
G.; Freeman,C.M.; Bearden,M. D.; Koech,P. ChemSusChem 2015, 8,3617.
11) Mathias,P. M.; Afshar,K.; Zheng,F.; Bearden,M. D.; Freeman,
(
C. J.; Andrea, T.; Koech,P. K.; Kutnyakov, I.; Zwoster,A.; Smith,A. R.;
Jessop, P. G.; Nik, O. G.; Heldebrant, D. J. Energy Environ. Sci. 2013, 6, 2233.
(
12) Zheng,F.; Heldebrant,D.J.; Mathias,P. M.; Koech,P.; Bhakta,
M.; Freeman,C.J.; Bearden,M. D.; Zwoster,A. Energy Fuels 2016, 30,1192.
13) Koech,P. K.; Rainbolt,J.E.; Bearden,M. D.; Zheng,F.;
Acknowledgements
(
Heldebrant,D.J. Energy Environ.Sci. 2011, 4,1385.
(14) Koech,P. K.; Zhang,J.; Kutnyakov, I. V.; Cosimbescu,L.;Lee,
S.-J.; Bowden,M. E.; Smurthwaite,T.D.; Heldebrant,D.J. RSC Advances
The authors acknowledge the U.S. Department of Energy’s
Office of Fossil Energy for funding award number FWP-65872.
Computational resources were provided through a National
Energy Research Scientific Computing Center User Proposal,
and Pacific Northwest National Laboratory (PNNL) Institutional
Computing. The authors acknowledge Maura K. Zimmerschied
for technical editing this paper. PNNL is proudly operated by
Battelle for the U.S. Department of Energy
2
013, 3,566.
15) Heldebrant,D.J.; Koech,P. K.; Ang, M. T. C.; Liang,C.;
Rainbolt,J.E.; Yonker, C.R.; Jessop,P.G. Green Chem. 2010, 12,713.
16) Blasucci,V.; Hart, R.; Mestre,V. L.; Hahne,D.J.; Burlager,M.;
(
(
Huttenhower, H.; Thio, B. J. R.; Pollet, P.; Liotta, C. L.; Eckert, C. A. Fuel 2010,
89,1315.
(17) Gurkan,B.; Goodrich,B.F.; Mindrup,E. M.; Ficke, L. E.; Massel,
M.; Seo, S.; Senftle, T. P.; Wu, H.; Glaser,M. F.; Shah,J. K.; Maginn,E. J.;
Brennecke,J.F.; Schneider,W.F. J. Phys.Chem.Lett. 2010, 1,3494.
(
18) Wu, H.; Shah,J. K.; Tenney,C. M.; Rosch,T.W.; Maginn,E. J.
Ind.Eng.Chem.Res. 2011, 50,8983.
19) Seo,S.; Quiroz-Guzman,M.; DeSilva,M. A.; Lee,T. B.; Huang,
Y.; Goodrich, B. F.; Schneider, W. F.; Brennecke, J. F. J. Phys. Chem. B 2014,
Keywords: hydrogen bonding • viscosity reducing factors •
molecular design • water-lean solvents • molecular simulation
(
1
18,5740.
(
20) Li, A.; Tian,Z.; Yan, T.; Jiang,D.-e.;Dai,S. J. Phys.Chem.B
2
014, 118,14880.
References
(21) Wang,C.; Luo,X.; Luo,H.; Jiang,D.-e.;Li,H.; Dai,S. Angew.
Chem.Int.Ed. 2011, 50,4918.
(
1)
Greenh.Gas Control 2007, 1,396.
2) Wang,C., Mahurin S. M., Luo H., Baker,G. A., Li, H., Dai,S.
Green Chem. 2010, 12,870.
3) Bates,E. D.; Mayton, R. D.; Ntai, I.; Davis,J. H. J. Am.Chem.
Soc. 2002, 124,926.
4) Perry, R. J.; Wood,B. R.; Genovese,S.;O’Brien,M. J.;
Peeters,A. N. M.; Faaij,A. P. C.; Turkenburg,W.C. Int. J.
(22) Kim,M.; Park,J.-W. Chem.Commun. 2010,2507.
(23) Cantu,D. C.; Lee,J.; Lee,M.-S.; Heldebrant,D.J.; Koech,P. K.;
Freeman,C.J.; Rousseau,R.;Glezakou,V.-A. J. Phys.Chem.Lett.2016, 7,
1646.
(24) Hunt,P. A.; Ashworth,C.R.; Matthews,R.P. Chem.Soc.Rev.
2015, 44,1257.
(25) Luo,X. Y.; Fan,X.; Shi,G. L.; Li,H. R.; Wang,C. M. J. Phys.
Chem.B 2016, 120,2807.
(26) Cantu,D. C.; Malhotra,D.; Koech,P.K.; Heldebrant,D.J.;
Zheng,F.; Freeman,C.J.; Rousseau,R.;Glezakou,V.-A. Green Chem.2016
18,6004.
(
(
(
Westendorf, T.; Meketa, M. L.; Farnum, R.; McDermott, J.; Sultanova, I.; Perry,
T. M.; Vipperla,R.-K.;Wichmann,L.A.; Enick,R. M.; Hong,L.;Tapriyal,D.
Energy Fuels 2012, 26,2528.
(5)
Lin,K.-Y. A.; Park,A.-H. A. Environ.Sci.Technol.2011, 45,
6
633.
(27) Shishkin,O.V.; Gorb,L.; Leszczynski,J. In Practical Aspects of
ComputationalChemistry:Methods,Concepts and Applications;Leszczynski,
J., Shukla,M. K., Eds.;Springer Netherlands:Dordrecht,2010,p 399.
(28) O’Brien,M. J.; Farnum,R.L.; Perry, R. J.; Genovese,S.E.
Energy Fuels 2014, 28,3326.
(6)
Perry, R. J.; Grocela-Rocha,T.A.; O'Brien,M. J.; Genovese,S.;
Wood,B. R.; Lewis,L.N.; Lam,H.;Soloveichik,G.;Rubinsztajn,M.;
Kniajanski,S.;Draper,S.; Enick,R. M.; Johnson,J.K.; Xie, H.-b.; Tapriyal,D.
ChemSusChem 2010, 3,919.
(
(
7)
8)
Perry, R. J.; O’Brien,M. J. Energy Fuels 2011, 25,1906.
Blasucci,V.; Dilek,C.; Huttenhower,H.;John,E.; Llopis-Mestre,
(29) Yang, Z.-Z.; Jiang,D.-e.;Zhu,X.; Tian,C.; Brown,S.; Do-Thanh,
C.-L.; He,L.-N.; Dai,S. Green Chem.2014, 16,253.
V.; Pollet,P.; Eckert,C. A.; Liotta,C.L. Chem.Commun. 2009,116.
9) Blasucci,V. M.; Hart,R.; Pollet,P.; Liotta,C. L.; Eckert,C. A.
Fluid Phase Equilib. 2010, 294,1.
(
This article is protected by copyright. All rights reserved.

ChemSusChem
10.1002/cssc.201601622
FULL PAPER
Entry for the Table of Contents (Please choose one layout)
Layout 2:
FULL PAPER
a
a
Deepika Malhotra, Phillip K. Koech,*
a
b
David J. Heldebrant, David C. Cantu,
Feng Zheng, Vassiliki-Alexandra
Glezakou, Roger Rousseau
a
b
b
Page No. – Page No.
Reinventing design principlesfor
developing low-viscositycarbon
dioxide binding organic liquids
Text for Table of Contents
(
2
CO BOLs) for flue gas clean up
This article is protected by copyright. All rights reserved.