New High-Performance Porous Ionic Liquids for Low Pressure CO2 Capture Jocasta Avila, Luiz Fernando Lepre, Catherine Santini, Martin Tiano, Sandrine Denis-Quanquin, Kai Chung Szeto, Agilio Padua, Margarida Costa Gomes
Submitted date: 15/12/2020 • Posted date: 17/12/2020 Licence: CC BY-NC-ND 4.0 Citation information: Avila, Jocasta; Lepre, Luiz Fernando; Santini, Catherine; Tiano, Martin; Denis-Quanquin, Sandrine; Szeto, Kai Chung; et al. (2020): New High-Performance Porous Ionic Liquids for Low Pressure CO2 Capture. ChemRxiv. Preprint. https://doi.org/10.26434/chemrxiv.13382741.v1
Porous ionic liquids are non volatile, versatile materials that associate porosity and fluidity. New porous ionic liquids, based on the ZIF-8 metal-organic framework and on phosphonium acetate or levulinate salts, were prepared and show an increased capacity to absorb carbon dioxide at low pressures. Porous suspensions based on phosphonium levulinate ionic liquid absorb reversibly 103% more carbon dioxide per mass than pure ZIF-8 per mass at 1bar and 303K. We show how the rational combination of MOFs with ionic liquids can greatly enhance low pressure CO2 absorption, paving the way toward a new generation of high-performance, readily available liquid materials for effective low pressure carbon capture.
File list (2)
low pressure CO2 capture
Jocasta Avila,† L. Fernando Lepre,† Catherine S. Santini,‡ Martin Tiano,†
Sandrine Denis-Quanquin,† Kai Chung Szeto,‡ Agilio A. H. Padua,† and
Margarida Costa Gomes∗,†
†Laboratoire de Chimie de l’ENS Lyon, CNRS and Université de Lyon, 46 allée d’Italie,
69364 Lyon, France
‡University of Lyon, CPE Lyon, CNRS, UMR 5265, Chemistry, Catalysis, Polymers and
Processes (C2P2), 43 Bvd. Du 11 Novembre 1918, F-69616 Villeurbanne, France
E-mail: margarida.costa-gomes@ens-lyon.fr
Porous ionic liquids are non volatile, versatile materials that associate porosity and
fluidity. New porous ionic liquids, based on the ZIF-8 metal-organic framework and on
phosphonium acetate or levulinate salts, were prepared and show an increased capacity
to absorb carbon dioxide at low pressures. Porous suspensions based on phosphonium
levulinate ionic liquid absorb reversibly 103% more carbon dioxide per mass than pure
ZIF-8 per mass at 1 bar and 303K. We show how the rational combination of MOFs
with ionic liquids can greatly enhance low pressure CO2 absorption, paving the way
toward a new generation of high-performance, readily available liquid materials for
effective low pressure carbon capture.
Materials combining the properties of liquids and porous solids, such as fluidity and perma-
nent porosity, are receiving growing attention from academia and industry since the discovery
of porous liquids in 2007.1 These polyfunctional materials open a wide range of possibili-
ties in materials and process development, including potential applications in separations,
catalysis and drug delivery.2
Porous liquids can be obtained by functionalizing porous solid materials or by dissolving
them in sterically hindered solvents, leading to type I and type II porous liquids, respec-
tively.2–5 An alternative approach consists in suspending porous solid materials in voluminous
solvents to form type III porous liquids.6–10 These last were successfully prepared by mixing
metal-organic frameworks (MOF) with ionic liquids to form stable suspensions that remain
liquid at ambient conditions. These suspensions have been designated as porous ionic liq-
uids as they maintain typical features of both the porous solid and the liquid salt.6,11 So
far, voluminous phosphonium-based ionic liquids, such as trihexyltetradecylphosphonium
bis(trifluoromethylsulfonyl)imide, [P6,6,6,14][NTf2], and trihexyltetradecylphosphonium chlo-
ride, [P6,6,6,14][Cl], have succeeded in stabilizing and maintaining the permanent porosity of
both ZIF-8 and HKUST-1 (a zinc methylimidazolate and a copper benzene-1,3,5-carboxylate
MOF, respectively) forming porous ionic liquids characterized by a significantly increased,
yet non-selective, physisorption of carbon dioxide, nitrogen and methane.6,11
One of the most appealing features of ionic liquids is the possibility of combining anions
and cations, including reactive groups, to tune their physical and chemical properties.12 Ionic
liquids are practically non-volatile and so their use can decrease the costs of gas separation
processes, namely for solvent regeneration, while also reducing the emissions of volatile
and toxic chemicals.13 Ionic liquids containing amino-functionalized cations14 or aprotic-
heterocyclic15 or carboxylate anions13,16,17 can react with carbon dioxide and have been
successfully tested as high capacity liquid gas sorbents for carbon capture processes.18
We have prepared new porous ionic liquids based on voluminous phosphonium cations
2
and carboxylate anions and tested them as liquid sorbents for carbon dioxide at low pres-
sures. So far, studies involving chemisorption of gases using fluid porous ionic liquids have
not been reported. We believe that if permanent porosity can be added to ionic liquids
capable of reversibly reacting with carbon dioxide,13,16,17 their capacity could be further in-
tensified. Imidazolium acetate ionic liquids are known for their ability to chemically absorb
relatively large amounts of carbon dioxide through a reversible reaction between the gas and
the imidazolium cation.16,17 Phosphonium acetate ionic liquids can also absorb carbon diox-
ide chemically,19 so phosphonium carboxylate salts are probably good candidates to make
reactive, porous ionic liquids. In this work we verify the presence of permanent porosity in
suspensions of ZIF-8, a previously tested MOF in different phosphonium carboxylate ionic
liquids, and we study the role played by the cation and the anion in the carbon dioxide
absorption mechanism.
The ionic liquids tetrabutylphosphonium acetate, [P4,4,4,4][OAc], and tetrabutylphosphonium
levulinate, [P4,4,4,4][Lev], were both synthesized and characterized as described in the Suple-
mentary Information. The ionic liquid trihexyltetradecylphosphonium bis(trifluoromethyl
sulfonylamide), [P6,6,6,14][NTf2], was purchased from Iolitec >98% pure. Tetrabutylphospho-
nium hydroxide[P4,4,4,4][OH], was purchased as a 40% wt. solution in water from Sigma
Aldrich. Before use, the [P4,4,4,4][OAc], [P4,4,4,4][Lev] and [P6,6,6,14][NTf2] ionic liquids were
degassed for at least 72 h under a primary vacuum and kept dried and degassed before the
preparation of the ionic liquids mixtures and of the porous liquids. Acetic acid ≥ 99% and
levulinic acid ≥ 97% were both purchased from Sigma Aldrich and used without further
purification. ZIF-8, a 2-methylimidazole zinc salt (ZnIm2) with a pore aperture of 3.4Å and
a pore diameter of 11.6Å — Basolite Z1200 (BASF product 691348) — was obtained from
3
Sigma Aldrich, batch STBG590V. The solid was passed through a 11 µm pore Nylon Nitex
sieve to reduce the distribution of solid particle sizes in the porous liquid. Carbon dioxide,
CO2 4.5, was purchased from Messer with a mole fraction purity of 99.995% and used as
received.
Sample preparation
The ionic liquid samples used are listed in Table 1. The ionic liquid mixtures were prepared by
weighting the components using a New Classic MS Mettler Toledo balance with an accuracy
of ±0.01mg inside a glove box GP(campus) Jacomex.
Table 1: Sample nomenclature and composition of the ionic liquids and the ionic liquids mixtures
Sample Chemicals mIL1 (g) mIL2 (g) χ(IL2)
IL1 [P6,6,6,14][NTf2] - - - IL2 [P4,4,4,4][OAc] - - - IL3 [P4,4,4,4][Lev] - - - IL10.75IL20.25 [P6,6,6,14][NTf2]0.75[P4,4,4,4][OAc]0.25 4.04216 0.56097 0.24958 IL10.5IL20.5 [P6,6,6,14][NTf2]0.5[P4,4,4,4][OAc]0.5 3.08371 1.25800 0.49435 IL10.25IL20.75 [P6,6,6,14][NTf2]0.25[P4,4,4,4][OAc]0.75 1.93509 2.42761 0.75040
The porous liquids, suspensions of ZIF-8 in the IL1, IL2, IL10.75IL20.25, IL10.5IL20.5,
IL10.25IL20.75 and IL3, were prepared at room temperature by weighing the components
using a New Classic MS Mettler Toledo balance with an accuracy of ±0.01mg, and by
stirring the mixture at 430–540min−1 during 5min. The porous liquids were degassed under
primary vacuum for up to 24 h before use. The porous liquids prepared are listed in Table
2, the uncertainty on the suspended solid concentration being estimated as ±0.0005%.
Density and viscosity
Density and viscosity were measured in a coupled Anton Paar density meter (DMA 5000
M) and viscometer (LOVIS 2000 ME) in the temperature range 293–353K at atmospheric
4
Table 2: Sample nomenclature and composition of the porous liquids
Sample Chemicals mIL (g) mMOF (g) MOF %w/w
PoIL1-Z5 IL1+ZIF-8 2.01215 0.09983 4.9614 Po(IL10.75IL20.25)− Z5 IL10.75IL20.25 + ZIF-8 1.91275 0.09564 5.0001 Po(IL10.5IL20.5)− Z5 IL10.5IL20.5 + ZIF-8 1.99626 0.09977 4.9978 Po(IL10.25IL20.75)− Z5 IL10.25IL20.75 + ZIF-8 1.99327 0.10014 5.0239 PoIL3-Z5 IL3+ZIF-8 2.51755 0.12335 4.8996
pressure. The densimeter’s measuring element is a U-shaped vibrating-tube, that is elec-
tronically excited to oscillate at its characteristic frequency whose value changes depending
on the density of the filled sample. The true density of the sample is determined via a
precise measurement of the characteristic frequency. The calibration of the equipment was
performed before the measurements with two substances of the precisely known densities,
air and the Anton Paar density standard ultra-pure water. Even if the densimeter is only
recommended for the measurement of homogeneous fluid samples, we believe that in the
present case, the density of the stable suspensions is also precise to within 5× 10−5 g cm−3
as a function of the temperature controlled to within 0.001 C.
The viscometer is based on Hoeppler falling ball principle, measuring the rolling time of
a ball through the liquid confined in a glass capillary. In order to cover the samples viscosity
in the worked temperature range, two capillaries with different diameter, 1.8 and 2.5 mm,
were calibrated with two different standard oils, APN26 and APN415 respectively, and two
balls with known material, geometry and density. The precision of this module is 0.05% in
the viscosity and 0.02 C in the temperature. Both the densimeter and viscometer cells were
filled with ca. 1 mL of sample and the measurements were carried out simultaneously, using
the temperature table scan mode.
5
Gas absorption
Gas solubilities were measured by a gravimetric method using an Intelligent Gravimetric
Analyzer (IGA001) from Hiden Analytical in the 0.5–5 bar pressure range at 303K. The
gravimetric experiments were carried out as extensively described previously.11 Essentially,
the samples are loaded into the microbalance and are degassed at secondary vacuum up to
24 h before starting several cycles of absorption/desorption for each temperature, which are
automatic controlled by the equipment software. The mass of gas absorbed, mg, at each pres-
sure and temperature was obtained from the raw weight data, mreading, using equation (1).
mreading = ms +mg +mEP g −
∑ i
mi
ρs(Ts) ρg(Ts, p) (1)
where ms is the mass of degassed sample, mEP g the effect due to adsorbed gas on the balance
components (determined by performing a blank measurement), and the sums over the i and
j components account for the respective buoyancy effects, on the sample and counterweight
sides, respectively. Components may be at different temperatures as specified in Table
S1. The effect of the gas dissolved for the calculation of the buoyancy effect (last term in
equation (1)) was considered negligible.
A blank measurement performed with an empty cell but in the same pressure and tem-
perature conditions as the measured samples is used to determine the necessary corrections:
mEP reading = m0 +mEP
(2)
where m0 is initially recorded by the IGA balance software during the empty run, and zeroed
by the software (tare) before starting the isotherms with sample.
6
The quantity of gas absorbed at each temperature and pressure per mass of liquid, b(p, T )
is calculated as:
b(p, T ) = (mg/Mw)
NMR measurements
NMR experiments were carried out in the pure ionic liquids, [P4,4,4,4][OAc] and [P4,4,4,4][Lev],
before and after contact with CO2. The pure ionic liquid was first degassed and dried
under primary vacuum (< 0.1 mbar) during 72 hours at 353K for [P4,4,4,4][OAc] and at
313K for [P4,4,4,4][Lev]. The liquids were then transferred to a high pressure NMR tube and
degassed again over 48h at the same temperatures. Prior to degassing, sealed glass capillaries
containing DMSO-d6 for [P4,4,4,4][OAc] and C6D6 for [P4,4,4,4][Lev] was inserted in the NMR
high-pressure tube to act as internal references, respectively. 1H, 31P, 13C, COSY, HSQC, HMBC NMR spectra were measured in a Bruker Avance 400
MHz spectrometer with a Prodigy probe at 342K for [P4,4,4,4][OAc] with or without CO2,
and at 323K for [P4,4,4,4][Lev] with or without CO2. After recording the spectra for the pure
ionic liquids under vacuum (p ∼ 0.1 bar), the same samples were left under a pressure of 10
bar of CO2 at 342K for [P4,4,4,4][OAc] and 20 bar of CO2 at 313K for [P4,4,4,4][Lev] for several
days until no pressure drop was observed after closing the gas cylinder. A digital pressure
gauge with an accuracy of ±0.4 bar was used to control the gas pressure. 1H, 31P and 13C
NMR spectra of the CO2 solutions were then collected using the same conditions as that
used for the pure ionic liquids.
Infrared measurements
The infrared spectra of [P6,6,6,14][NTf2], [P4,4,4,4][OAc] and its mixtures were collected in the
attenuated total reflection mode (ATR) of a Perkin Elmer Spectrum 65 FT-IR spectrometer.
A droplet of each sample (IL1, IL10.75IL20.25, IL10.5IL20.5 and IL10.25IL20.75) and a small
7
amount of solid (IL2) was placed on top of the ATR crystal, and the measurements were
performed in order to accumulate 64 scans in the wave number range of 515–4000 cm−1 with
2 cm−1 of resolution.
The infrared spectra of [P4,4,4,4][Lev] under CO2 pressure were obtained in an integrated
system comprising mass flow controllers (Brooks), FT-IR adapted high temperature, airtight
reaction chamber (Harrick Scientific) with a 4-way valve and a pressure regulating valve.
The reaction chamber was equipped with ZnSe windows and fitted into the Praying Mantis
optical unit also provided by Harrick. A film of the ionic liquid was deposited onto a
silicon wafer in an Ar-filled glovebox. All the lines were extensively purged before CO2
was introduced in the chamber. The FTIR spectra were collected in a Thermo Scientific
FTIR 6700 spectrophotometer in diffuse reflectance mode equipped with a MCT detector.
A spectrum comprises 64 scans and was recorded continuously at 4 cm−1 resolution at room
temperature up to 10 bar.
Results and Discussion
The CO2 absorption capacity of the synthesized tetrabutylphosphonium acetate, [P4,4,4,4][OAc],
was gravimetrically measured at 343K. [P4,4,4,4][OAc] is a solid at room temperature having
a melting point at 327.5K.20 The experimental results depicted in Figure 1 reveal a consid-
erable hysteresis between absorption and desorption isotherm cycles probably due to a slow
diffusivity of the gas. Average values with the corresponding uncertainties were considered
within this work. The experimental data are listed in detail in Supplementary Tables S2-S3.
Table S2 shows the raw data, whereas Table S3 shows the average between absorption an
desorption cycles for each isotherm with the respective uncertainty in temperature, pressure
and gas solubility.
The isotherm in Figure 1 shows that CO2 absorption capacity is not linearly dependent
with pressure, suggesting that the gas is chemically absorbed by [P4,4,4,4][OAc]. Indeed,
8
Figure 1: CO2 Absorption by [P4,4,4,4][OAc] in the pressure range of 0–5 bar at 343K. .
chemical reactions between CO2 and tetraalkylphosphonium-based ionic liquids have already
been reported when phosphonium cations are paired with basic anions such as azolides19 or
phenolates.21 In the presence of CO2 these basic anions are capable of stabilizing the most
acidic proton in the α − carbon of the tetraalkylphosphonium cation and a proton transfer
can occur. The corresponding acid of the anion is formed as well as an ylide species that is
subsequently trapped with CO2 (Figure 2). The ylide generation from phosphonium-based
ionic liquids has been used as Wittig reagents22 or reaction catalysts.23
Concerning the acetate anion, it is long known that it is basic enough to abstract the acid
proton from imidazolium cations and lead to a chemical reaction between imidazolium-based
acetate ionic liquids and CO2.16,17 Therefore, one could expect that the basicity of the acetate
anions can also lead to a carboxylation at the α − carbon atom of tetraalkylphosphonium
acetate cations.
In order to study the CO2 absorption mechanism in [P4,4,4,4][OAc], NMR spectroscopic
experiments were performed before and after CO2 absorption. The 31P NMR spectra in
9
Figure 2: Mechanism of CO2 absorption by a tetraalkylphosphonium acetate ionic liquid.
Figure 3 shows two different phosphonium environments after CO2 absorption at 10 bar and
342K. The new peak at 30.75 ppm reveals that CO2 reacts with the phosphonium cation
and a new species is generated. In fact, Gohndrone et al.19 observed a similar peak for the
ionic liquid [P6,6,6,14][2-CNpyr] after CO2 absorption at 3 bar and 333K. The peaks at ca.
37.2 ppm and at ca. 44.3 ppm are present both before and after contact with CO2 and can
be assigned to the presence of oxide impurities. The 1H NMR spectrum in Figure 3 shows a
new broad peak at ca. 16.1 ppm, corresponding to the acidic proton (COOH) of a carboxylic
acid. Precise analysis of 1H, 13C NMR spectra, together with HSQC and COSY 2D NMR
experiments, confirmed the formation of a zwitterionnic adduct by the condensation of CO2
onto [P4,4,4,4], as well as the protonation of acetate anion in acetic acid. (Details are given
in Supplementary Information).
Since [P4,4,4,4][OAc] is solid at room temperature, mixtures with another phosphonium
based ionic liquid, [P6,6,6,14][NTf2], were prepared in order to be able to work with a liquid in
a wider range of temperature. The addition of [P4,4,4,4][OAc] (IL2) to [P6,6,6,14][NTf2] (IL1)
lead to homogeneously stable mixtures at room temperature in the range of [P4,4,4,4][OAc]
mole fraction compositions, 0.25 ≤ χ(IL2) ≥ 0.75. The mixtures were characterized by
determining their densities and viscosities that are reported in Tables S4-S5 and Figure S3.
The interactions between the components of the mixture and their deviations from ideal
behavior were studied using infrared spectroscopy and by determining their excess molar
properties (Figure S3-S5). No strong, specific interaction was found in the mixtures and no
effect of the proximity of the ions due an efficient packing in the mixtures could be detected
(details are included in Supplementary Information).
10
Figure 3: 1H (left) and 31P (right) NMR spectra of [P4,4,4,4][OAc] before (black line) and after (red line) CO2 absorption at 342K. The 1H highlighted peak at ca. 16 ppm corresponds to the acidic proton (COOH) of acetic acid (H3CCOOH) produced.
Since all the mixtures are liquid at room temperature, the gas absorption measurements
could be performed at 303K. The results are presented in Figure 4 and Supplementary
Tables S2-S3. All the mixtures are able to absorb significant larger quantities of carbon
dioxide than pure [P6,6,6,14][NTf2], the absorption capacity increasing with the increase of
concentration in phosphonium acetate. As previously observed for [P4,4,4,4][OAc] (Figure 1),
CO2 absorption does not vary linearly with pressure for all mixtures, pointing towards a
chemical reaction of the gas with the mixtures even in the sample with lower [P4,4,4,4][OAc]
mole fraction, IL10.75IL20.25.
We have followed our previous approach to prepare porous ionic liquids based on the
11
Figure 4: Absorption of CO2 by IL1, IL10.75IL20.25, IL10.5IL20.5 and IL10.25IL20.75 in the pressure range of 0–5 bar at 303K.
[P6,6,6,14][NTf2] + [P4,4,4,4][OAc] mixtures6,11 and ZIF-8. The suspensions were characterized
by measuring their gas absorption capacity and by comparing it to that of the liquid6 (rep-
resented in Figure 5 and reported in Tables S2-S3). The additional gas capacity attributed
to the different MOF concentration in the porous ionic liquids was calculated as previously:6
nMOF g = nPL
g − nIL g (4)
where, nMOF g , nPL
g and nIL g stand for the quantity of gas absorbed by the MOF, the porous
liquid and the ionic liquid in the sample, respectively. In a similar way, the relative amount
of gas adsorbed by the MOF in the porous liquids relative to that adsorbed in the pure MOF
(%β) is calculated as:
12
Table 3: MOF content in the porous ionic liquid, quantity of gas absorbed, expected gas absorption for the MOF content, percentage of MOF active to adsorb gas in the porous ionic liquid. All data is at 5 bar and T = 303K and %β is related to the experimental absorption, bExp.
Sample MOF bExp bCal %β (%w/w) (mmol g−1) (mmol g−1)
CO2
PoIL1-Z5 4.9614 0.0995 0.1308 76 Po(IL10.75IL20.25)− Z5 5.0001 0.0210 0.1319 16 Po(IL10.5IL20.5)− Z5 4.9978 0.0642 0.1319 49 Po(IL10.25IL20.75)− Z5 5.0239 0.0168 0.1325 13 PoIL3-Z5 4.8996 0.1373 0.1292 100 ZIF-8 100 2.6372 - 100
Figure 5: Absorption of CO2 in the pressure range of 0–5 bar at 303K by IL1 (black), IL10.75IL20.25 (dark blue), IL10.5IL20.5 (dark green), IL10.25IL20.75 (dark orange) compared to the porous liquids formed by the addition of ZIF-8 in IL1 and in IL1+IL2 mixtures, PoIL1-Z5 (grey), Po(IL10.75IL20.25) − Z5 (light blue), Po(IL10.5IL20.5) − Z5 (light green), Po(IL10.25IL20.75)− Z5 (light orange) and ZIF-8 (open symbols and dashed line).
13
We have shown in our previous work11 that the relative increase in the gas physical
absorption of porous ionic liquids compared to those of the salts that constitute them depends
on the relative amount of MOF suspended, thus proving that the pores of the solid remain free
in the suspension. As highlighted in Table 3 (and in Figure 5), this is also observed herein for
PoIL1-Z5 which presents an increase in CO2 absorption corresponding to approximately 76%
of the MOF’s capacity. However, in the suspensions formed by the ionic liquids mixtures,
Po(IL1χIL21−χ), the relative increase on CO2 absorption compared to that of the IL1χIL21−χ
mixtures is much lower than the expected capacity based on the quantity of added MOF.
A maximum %β of ca. 49% is found for Po(IL10.5IL20.5) indicating that the ZIF-8 pores
are not fully available to host the gas.11 As shown before, ionic liquids based on tetrabutyl
phosphonium cations are voluminous enough to stay out of the ZIF-8 pores. The lower than
expected CO2 absorption in Po(IL1χIL21−χ) can only be explained by the presence of acetic
acid, a product of the chemical reaction between [P4,4,4,4][OAc] and CO2, which is small
enough to enter the pores of ZIF-8.
In order to circumvent other species than the gas from entering the pores in the suspension
while maintaining reactivity towards carbon dioxide, we have prepared porous ionic liquids
based on salts with larger carboxylate anions. We have maintained the sufficiently volumi-
nous cation (P + 4,4,4,4) and replaced acetate (CH3COO–) by levulinate (CH3CO(CH2)2COO–),
a considerably larger anion. We obtained a liquid salt at room temperature, [P4,4,4,4][Lev],
with a density and a viscosity slightly lower than that of IL1 and that of IL1χIL21−χ mixtures
with χ > 0.5 ( reported in Table S4 and Table S5, respectively).
As in the case of acetate, the levulinate anion also has a carboxylic group able to accept
a proton from the α−carbon of the tetrabutylphosphonium cation. In presence of CO2, a
reaction mechanism similar to that observed for [P4,4,4,4][OAc] (Figure 2) could be expected
for [P4,4,4,4][Lev]. We have identified other possible reaction paths for [P4,4,4,4][Lev] in presence
of CO2 as shown in Figure 6 with carboxylation occurring in the acidic protons on positions
3 and 5 of the levulinate anion.
14
Figure 6: Chemical structure of the [P4,4,4,4][Lev], (1), along with carboxylated species formed after the absorption of CO2, (2), (3), (3’), (4) and (4’). Numbers in red highlight the acidic protons position of the anion levulinate.
The absorption of CO2 by [P4,4,4,4][Lev] was followed by NMR and FTIR spectroscopy.
As with [P4,4,4,4][OAc], the 31P NMR spectrum shows a new peak, corresponding to a new
phosphonium species (Figure 7). In addition, three different peaks in the 20–15 ppm region of 1H NMR spectrum show that at least three new carboxylic functions were formed in presence
of CO2. Precise analysis of 1H and 13C NMR spectra, together with HSQC, HMBC and
COSY 2D NMR experiments led us to consider the formation of several compounds. These
compounds correspond to the expected product (2) in Figure 6, formed by the carboxylation
15
Figure 7: 1H (left) and 31P (right) NMR spectra of [P4,4,4,4][Lev] before (black line) and after (red line) CO2 absorption at 323K. The 1H highlighted peaks at ca. in both spectra corresponds to the acidic protons (COOH) produced.
16
of one α−carbon of the phosphonium cation, but also to the products of carboxylation of
the anion in positions 3 and 5. We suggest the formation of Lev-3-CO2-keto (3) and Lev-5-
CO2-enol (4’) in Figure 6, from CO2 addition at the C-3 position of the anion, and from CO2
addition at the C-5 position followed by the enolisation of the carbonyl group, respectively
(details are included in Supplementary Information). 1H NMR spectra of levulinic acid and levulinate anions after carboxylation in different
positions were obtained from ab initio calculations (using Gaussian 09, Revision D.01) at the
HF/6-311+G(2d/p) level using the GIAO method in an implicit solvent of relative permit-
tivity 20.7. The geometries were previously optimized at the B3LYP/6-31G(d) level, in the
same implicit solvent. The nature of the three acid proton peaks observed experimentally in
the chemical shift range of 15–20 ppm (as seen in Figure 7) could be confirmed as depicted
in Figure S25. The different positions of the carboxylation were considered as well as the
keto-enolic equilibrium. Even if the simulation did not take into account the cation, [P4,4,4,4] +
and the dissolved CO2, it clearly shows the existence of intramolecular hydrogen bonding
involving the COOH and OH groups in the anion. The presence of this hydrogen bonds
leads to the de-shielding of the protons involved and confirms the possibility of carboxylation
of different sites of the anion.
FTIR spectroscopy of [P4,4,4,4][Lev] with different CO2 pressures and contact times be-
tween the gas and the ionic liquid confirmed the apparition of new O-H, C=C and C=O
bonds, in accordance with the formation of Lev-3-CO2-keto and Lev-5-CO2-enol considered
after NMR analysis (details in Supplementary Information).
The chemical absorption of carbon dioxide by [P4,4,4,4][Lev] is clearly shown in the isotherm
at 303K depicted in Figure 8, as the gas absorption does not vary linearly with pressure. The
capacity of [P4,4,4,4][Lev] to reversibly absorb CO2 is remarkable, especially at low pressures
(bellow 2 bar) where it exceeds that of the pure MOF by more than 100% at 303K.
When 5% w/w of ZIF-8 is suspended in [P4,4,4,4[Lev] to form the porous liquid PoIL3-Z5,
the CO2 absorption capacity is further enlarged, its relative increase being correlated with
17
Figure 8: Absorption of CO2 by IL3, PoIL3-Z5 and ZIF-8 in the pressure range of 0–5 bar at 303K.
the MOF concentration. As seen in Table 3, all the pores of the solid in this suspension are
available to absorb gas, PoIL3-Z5 presenting a %β of ca. 100%. This observation confirms
that the molar volume of the species in the reaction medium is important, the levulinic acid
as well as the CO2-anion aducts formed being voluminous enough to not enter the pores of
the suspended solid.
As previously reported, the volume of the ion pairs used to form type III porous ionic
liquids with ZIF-8 should be high enough to prevent pore obstruction in the liquid.6,11
Herein, we have shown that the volume of the species formed when these materials are used
as reaction media must also be considered. The species in solution must be sufficiently
voluminous to remain out of the pores in the liquid that are then free to adsorb gas. The
MOF must also be chosen carefully as shown in our previous work6 to allow the formation
of stable suspensions and to maintain permanent porosity in the liquid.
The porous liquid formed by the [P4,4,4,4][Lev] and ZIF-8 is, up to now, the most per-
18
formant fluid porous ionic liquid for absorbing CO2 at low pressure with a capacity of
1.5mmol g−1 at 303K and 2 bar, similar to those of currently used sorbent materials such
as ethanolamines (≈ 1.8mmol g−1 for N-ethyl-diethanolamine at 40 C and 1 bar24) or other
solid amine sorbents (2.6 mmol g−1 for mesoporous silica decorated with polyethylenimine
at 30 C and 1 bar24).
Conclusion
We have shown that it is possible to form porous ionic liquids that capture large quantities
of carbon dioxide reversibly and at low pressure. When sufficiently voluminous ionic liquids
are chosen, carbon dioxide absorption is intensified by the presence of permanent pores
in the liquid. The ionic liquids can be engineered to chemically absorb carbon dioxide
namely through the use of phosphonium carboxylate salts. The mechanisms of the CO2
chemical absorption were analysed in detail and point towards the use of these ionic liquids
as a promising route to prepare efficient and stable porous liquids for low pressure carbon
capture.
We present herein a real alternative to the use of aqueous amines for low pressure CO2
capture in absorption processes. The use of porous ionic liquids present the advantages
of using a non-volatile, thermally stable and non-corrosive liquid absorbent that does not
require a solvent and with performances at low pressures similar to those of the currently
used aqueous amines. Furthermore, the preparation of porous ionic liquids is easy and can
be done from off-the-shelf well characterised and readily available metal organic frameworks
and ionic liquids. The approach described makes use of the most attractive features of both
MOFs and ionic liquids: in the former, the ability to choose the metal, the organic linker
and the pore size, and in the latter the tuning of the physical and chemical properties of the
liquid salt through proper pairing of anions and cations, including reactive groups.
19
Acknowledgement
M.C.G. and J.A. thank IDEX-LYON for financial support (Programme Investissements
d’Avenir ANR-16-IDEX-0005).
Supporting Information Available
Experimental procedures, characterization and spectral data (ATR FTIR and NMR). Nu-
merical data in absorption measurements, density and viscosity.
20
References
(1) O’Reilly, N.; Giri, N.; James, S. L. Porous liquids. Chem. - A Eur. J. 2007, 13, 3020–
3025.
(2) Giri, N.; Del Pópolo, M. G.; Melaugh, G.; Greenaway, R. L.; Rätzke, K.; Koschine, T.;
Pison, L.; Gomes, M. F.; Cooper, A. I.; James, S. L. Liquids with permanent porosity.
Nature 2015, 527, 216–220.
(3) Zhang, J.; Chai, S. H.; Qiao, Z. A.; Mahurin, S. M.; Chen, J.; Fang, Y.; Wan, S.;
Nelson, K.; Zhang, P.; Dai, S. Porous liquids: A promising class of media for gas
separation. Angew. Chemie - Int. Ed. 2015, 54, 932–936.
(4) Zhang, F.; Yang, F.; Huang, J.; Sumpter, B. G.; Qiao, R. Thermodynamics and Kinetics
of Gas Storage in Porous Liquids. J. Phys. Chem. B 2016, 120, 7195–7200.
(5) Kearsey, R. J.; Alston, B. M.; Briggs, M. E.; Greenaway, R. L.; Cooper, A. I. Accelerated
robotic discovery of type II porous liquids. Chem. Sci. 2019,
(6) Costa Gomes, M.; Pison, L.; ervinka, C.; Padua, A. Porous Ionic Liquids or Liquid
Metal–Organic Frameworks? Angew. Chemie - Int. Ed. 2018, 57, 11909–11912.
(7) Li, P.; Chen, H.; Schott, J. A.; Li, B.; Zheng, Y.; Mahurin, S. M.; Jiang, D. E.; Cui, G.;
Hu, X.; Wang, Y.; Li, L.; Dai, S. Porous liquid zeolites: Hydrogen bonding-stabilized
H-ZSM-5 in branched ionic liquids. Nanoscale 2019, 11, 1515–1519.
(8) Shan, W.; Fulvio, P. F.; Kong, L.; Schott, J. A.; Do-Thanh, C. L.; Tian, T.; Hu, X.;
Mahurin, S. M.; Xing, H.; Dai, S. New Class of Type III Porous Liquids: A Promis-
ing Platform for Rational Adjustment of Gas Sorption Behavior. ACS Appl. Mater.
Interfaces 2018, 10, 32–36.
(9) Bavykina, A.; Cadiau, A.; Gascon, J. Porous liquids based on porous cages, metal
21
organic frameworks and metal organic polyhedra. Coord. Chem. Rev. 2019, 386, 85–
95.
(10) Liu, S.; Liu, J.; Hou, X.; Xu, T.; Tong, J.; Zhang, J.; Ye, B.; Liu, B. Porous Liquid:
A Stable ZIF-8 Colloid in Ionic Liquid with Permanent Porosity. Langmuir 2018, 34,
3654–3660.
(11) Avila, J.; Cervinka, C.; Dugas, P.-Y.; Padua, A.; Costa Gomes, M. Porous ionic liquids:
structure, stability and gas absorption mechanisms. Submitted
(12) Sheridan, Q. R.; Schneider, W. F.; Maginn, E. J. Role of Molecular Modeling in the
Development of CO2-Reactive Ionic Liquids. Chem. Rev. 2018, 118, 5242–5260.
(13) Shiflett, M. B.; Drew, D. W.; Cantini, R. A.; Yokozeki, A. Carbon dioxide capture
using ionic liquid 1-butyl-3-methylimidazolium acetate. Energy and Fuels 2010, 24,
5781–5789.
(14) Bates, E. D.; Mayton, R. D.; Ntai, I.; Davis, J. H. CO 2Capture by a Task-Specific
Ionic Liquid. J. Am. Chem. Soc. 2002, 124, 926–927.
(15) Gurkan, B. E.; de la Fuente, J. C.; Mindrup, E. M.; Ficke, L. E.; Goodrich, B. F.;
Price, E. A.; Schneider, W. F.; Brennecke, J. F. Equimolar CO 2Absorption by Anion-
Functionalized Ionic Liquids. J. Am. Chem. Soc. 2010, 132, 2116–2117.
(16) Gurau, G.; Rodríguez, H.; Kelley, S. P.; Janiczek, P.; Kalb, R. S.; Rogers, R. D.
Demonstration of chemisorption of carbon dioxide in 1,3-dialkylimidazolium acetate
ionic liquids. Angew. Chemie - Int. Ed. 2011, 50, 12024–12026.
(17) Stevanovic, S.; Podgoršek, A.; Pádua, A. A.; Costa Gomes, M. F. Effect of water on
the carbon dioxide absorption by 1-alkyl-3- methylimidazolium acetate ionic liquids. J.
Phys. Chem. B 2012, 116, 14416–14425.
22
(18) Bui, M. et al. Carbon capture and storage (CCS): The way forward. Energy Environ.
Sci. 2018, 11, 1062–1176.
(19) Gohndrone, T. R.; Bum Lee, T.; Desilva, M. A.; Quiroz-Guzman, M.; Schneider, W. F.;
Brennecke, J. F. Competing reactions of CO2 with cations and anions in azolide ionic
liquids. ChemSusChem 2014, 7, 1970–1975.
(20) Yang, Q.; Xu, D.; Zhang, J.; Zhu, Y.; Zhang, Z.; Qian, C.; Ren, Q.; Xing, H. Long-Chain
Fatty Acid-Based Phosphonium Ionic Liquids with Strong Hydrogen-Bond Basicity and
Good Lipophilicity: Synthesis, Characterization, and Application in Extraction. ACS
Sustain. Chem. Eng. 2015, 3, 309–316.
(21) Lee, T. B.; Oh, S.; Gohndrone, T. R.; Morales-Collazo, O.; Seo, S.; Brennecke, J. F.;
Schneider, W. F. CO2 Chemistry of Phenolate-Based Ionic Liquids. J. Phys. Chem. B
2016, 120, 1509–1517.
(22) Firaha, D. S.; Gibalova, A. V.; Hollóczki, O. Basic Phosphonium Ionic Liquids as Wittig
Reagents. ACS Omega 2017, 2, 2901–2911.
(23) Macarie, L.; Simulescu, V.; Ilia, G. Phosphonium-Based Ionic Liquids Used as Reagents
or Catalysts. ChemistrySelect 2019, 4, 9285–9299.
(24) Dutcher, B.; Fan, M.; Russell, A. G. Amine-based CO2 capture technology development
from the beginning of 2013-A review. ACS Appl. Mater. Interfaces 2015, 7, 2137–2148.
23
low pressure CO2 capture
Jocasta Avila,† L. Fernando Lepre,† Catherine S. Santini,‡ Martin Tiano,†
Sandrine Denis-Quanquin,‡ Kai Chung Szeto,‡ Agilio Padua,† and Margarida
Costa Gomes∗,†
†Laboratoire de Chimie de l’ENS Lyon, CNRS and Université de Lyon, 46 allée d’Italie,
69364 Lyon, France
‡University of Lyon, CPE Lyon, CNRS, UMR 5265, Chemistry, Catalysis, Polymers and
Processes (C2P2), 43 Bvd. Du 11 Novembre 1918, F-69616 Villeurbanne, France
E-mail: margarida.costa-gomes@ens-lyon.fr
Supplementary Information
Ionic liquids synthesis and characterization
The [P4,4,4,4][OAc] and [P4,4,4,4][Lev] were both synthetized via simple neutralization
reactions of tetrabutylphosphonium hydroxide, [P4,4,4,4][OH], with the respective carboxylic
acids, acetic and levulinic acids, yielding >94 %. The structures of both ionic liquids were
confirmed by NMR (1H, 13C and 31P) and by mass spectroscopy, then, the density and
viscosity were measured. All the detailed characterization results are in the supplementary
information.
Tetrabutylphosphonium acetate, [P4,4,4,4][OAc]. The [P4,4,4,4][OH], aqueous so-
lution 1 eq. (25.01 g; 36.17 mmol) was introduced into a 250 mL round three-necked flask
coupled to a reflux condenser and a thermometer, followed by the addition of a slight
Figure S1: Schematic neutralization reactions of tetrabutylphosphonium hydroxide, [P4,4,4,4][OH], with acetic and levulinic acids to produce the ionic liquids [P4,4,4,4][OAc] and [P4,4,4,4][Lev], respectively.
excess of acetic acid 1.02 eq. (2.217 g; 36.92 mmol diluted in 10 mL of water) using an
isobaric dropping funnel. The neutralization reaction was allowed to occur during 3 h
under stirring inside a water bath at room temperature. Upon mixing both components,
an increase of 10 C was observed in the solution by the thermometer. After neutralization,
a colorless solution is obtained and the water was removed by lyophilization at 217.15 K
and 1× 10−4 bar, remaining a white solid. The reactional scheme is shown in Fig. S1. 1H NMR (300 MHz, CDCl3, 298 K) δ/ppm: 0.90 (t, 12H, (CH2CH2CH3)4); 1.44 (m, 16H,
(CH2CH2CH3)4);2.24 (m, 8H, P(CH2)4); 1.84 (s, 3H, CH3COO). 13C (300 MHz, CDCl3,
298 K) δ/ppm: 13.39 ((CH2CH2CH3)4); 18.20 ((CH2CH2CH3)4); 18.83 ((CH2CH2CH3)4);
23.99 (P CH2); 23.64 (CH3COO); 176.88 (COO).31P NMR (121 MHz, 298 K) δ/ppm:
33.09 (s, P(CH2CH2CH2CH3)4). MS ES+ m/z (%Rel. Intensity): 259 M+ (100). Calc. for
[P4,4,4,4]+ 259.2555, found 259.3. MS ES– m/z (%Rel. Intensity): 377 M– (100). Calc. for
[P4,4,4,4][CH3COO] –2 377.2821, found 377.3. Yeld 95 %.
Tetrabutylphosphonium levulinate, [P4,4,4,4][Lev]. [P4,4,4,4][OH] aqueous so-
lution 1 eq. (25.01 g; 36.17 mmol) was neutralized with the addition of a slight ex-
cess of levulinic acid 1.02 eq. (2.217 g; 36.92 mmol diluted in 10 mL of water) as de-
2
scribed before. Upon mixing both components, an increase of 10 C was observed
in the solution as well as in the synthesis of [P4,4,4,4][OAc]. After neutralization, a
suspension is obtained. The excess of water was first removed in a rotary evapora-
tor at 323 K and 5× 10−2 bar, followed by the evaporation under vacuum at 333 K
and 1× 10−5 bar. After the water removal, a translucent dark brown liquid was ob-
tained. The suspension previously observed is no longer present. The reactional
scheme is shown in Fig. S1. 1H NMR (300 MHz, CDCl3, 297 K) δ/ppm: 0.95 (t, 12H,
(CH2CH2CH3)4); 1.48 (t, 8H, (CH2CH2CH3)4); 1.52 (t, 8H, (CH2CH2CH3)4); 2.38 (m,
8H, P(CH2)4); 2.71 (t, 2H, CH3COCH2CH2COO); 2.44 (t, 2H, CH3COCH2CH2COO);
2.14 (s, 3H, CH3COCH2CH2COO). 13C NMR (75 MHz, CDCl3, 297 K) δ/ppm: 13.50
(4C, (CH2CH2CH3)4); 18.32 (4C, (CH2CH2CH3)4); 18.94 (4C, (CH2CH2CH3)4); 23.87 (4C
(P CH2)); 176.79 (1C, CH3COCH2CH2COO); 29.92 (1C, CH3COCH2CH2COO); 40.91
(1C, CH3COCH2CH2COO); 210.06 (1C, CH3COCH2CH2COO); 32.46 (1C, CH3COCH2CH2COO). 31P NMR (121 MHz, CDCl3, 297 K) δ/ppm: 33.09 (s, P(CH2CH2CH2CH3)4). MS ES+
m/z (%Rel. Intensity): 259 M+ (100). Calc. for [P4,4,4,4]+ 259.2555, found 259.2. MS ES–
m/z (%Rel. Intensity): 489 M– (100). Calc. for [P4,4,4,4][CH3COCH2CH2COO] –2 489.3344,
found 489.3. Yield 94 %.
Sorption balance measurements
Table S1: IGA 001 microbalance components contributing to the buoyancy force.
Subscript Item Mass/g Material Density /g cm−3 T /K s dry sample ms Described in Table 1-2 ρs Tsample a interacting gas ma CO2 ρa Tsample i1* sample container 0.6143 stainless steel 7.9 Tsample i2* wire 0.0644 tungsten 21.0 Tsample i3 chain 0.1934 gold 19.8 318 j1 counter-weight 0.7870 stainless steel 7.9 308 j2 hook 0.0054 tungsten 21.0 308 j3 chain 0.1498 gold 19.8 318
* The gas absorption measurements of the sample IL2, [P4,4,4,4[OAc], were performed using i1 and i2 with a mass of 0.6032 g and 0.0650 g, respectively.
Figure S2: Schematic diagram of the IGA 001 gravimetric microbalance. The symbols i and j, described in Table S1, represent the sample side and the counterweight side components of the microbalance, respectively. FW and FB are the forces experienced by the sample side of the microbalance and the arrows indicate the direction of these forces: the weight due to gravity (FW ) and the buoyance due the fluid displacement (FB).1,2
4
Gas absorption data
Table S2: Absorption and desorption of CO2 by IL1, IL10.75IL20.25, IL10.5IL20.5, IL10.25IL20.75, Po(IL10.75IL20.25)−Z5, Po(IL10.5IL20.5)−Z5, Po(IL10.25IL20.75)−Z5, PoIL1- Z5, PoIL3-Z5 and ZIF-8 as function of pressure from 0–5 bar at 303 K, and by IL2 as function of pressure from 0–5 bar at 343 K. The exact concentrations of the ionic liquids mixtures and the ZIF-8 in the porous liquids are given in Table 1-2 of the manuscript.
T
CO2 — Absorption
IL1 IL2 IL3 303.191 0.0000 0.0000 339.449 0.0000 0.0000 304.834 0.0000 0.0000 303.153 0.4988 0.0270 343.000 0.2484 0.1693 303.148 0.4994 0.9072 303.148 0.7492 0.0400 343.183 0.4921 0.2961 303.158 0.9982 1.1419 303.162 0.9986 0.0528 343.009 0.9903 0.4306 303.153 1.9993 1.3710 303.138 2.4978 0.1307 343.077 2.9829 0.6615 303.158 2.9976 1.5371 303.148 4.9970 0.2617 343.072 4.9929 0.8544 303.143 4.9980 1.7686
IL10.75IL20.25 IL10.5IL20.5 IL10.25IL20.75
304.801 0.0000 0.0000 303.848 0.0000 0.0000 304.863 0.0000 0.0000 303.153 0.4998 0.0757 303.143 0.4984 0.2719 303.153 0.4998 0.6321 303.162 0.9977 0.1199 303.153 0.9972 0.3653 303.143 0.9987 0.7798 303.167 1.9973 0.1943 303.148 1.9978 0.4911 303.158 1.9988 0.9391 303.153 2.9975 0.2641 303.153 2.9987 0.6035 303.158 2.9981 1.0671 303.158 3.9982 0.3317 303.143 3.9993 0.6802 303.148 3.9983 1.1805 303.148 4.9982 0.3963 303.148 4.9984 0.7815 303.158 4.9969 1.2838
Po(IL10.75IL20.25)− Z5 Po(IL10.5IL20.5)− Z5 Po(IL10.25IL20.75)− Z5 303.196 0.0000 0.0000 304.820 0.0000 0.0000 303.596 0.0000 0.0000 303.158 0.4987 0.0528 303.153 0.4987 0.3000 303.138 0.5001 0.6525 303.153 0.9987 0.0897 303.148 0.9986 0.4071 303.148 0.9977 0.8099 303.134 1.9993 0.1668 303.148 1.9990 0.5277 303.143 1.9986 0.9500 303.138 2.9976 0.2456 303.138 2.9986 0.6293 303.134 2.9983 1.0572 303.143 3.9989 0.3240 303.148 3.9982 0.7204 303.138 3.9990 1.1518 303.145 4.9985 0.3975 303.158 4.9981 0.8066 303.153 4.9975 1.2361
PoIL1-Z5 PoIL3-Z5 ZIF-8 303.153 0.0000 0.0000 304.563 0.0000 0.0000 303.153 0.0000 0.0000 303.143 0.4994 0.0350 303.153 0.9990 1.1388 303.153 0.4990 0.2917 303.138 0.7478 0.0527 303.148 1.9978 1.5050 303.148 0.7484 0.4251 303.143 0.9983 0.0704 303.138 2.9976 1.6886 303.148 0.9982 0.5508 303.153 2.4987 0.1754 303.153 4.9978 1.9712 303.158 2.4967 1.3526 343.159 4.9977 0.1574 303.138 4.9991 2.6372
CO2 — Desorption
IL1 IL2 IL3 303.148 4.9970 0.2617 343.072 4.9929 0.8544 303.143 4.9979 1.7686 303.138 2.4978 0.1310 343.067 2.9979 0.7947 303.148 2.9982 1.5938 303.153 0.9986 0.0528 343.038 0.9976 0.6787 303.148 1.9996 1.4391 303.138 0.7485 0.0404 343.130 0.5004 0.5946 303.148 0.9973 1.1814
5
T
mmol g−1
303.153 0.4993 0.0274 343.101 0.2437 0.5022 303.153 0.4991 0.8926 303.281 0.0000 0.0000 340.350 0.0000 0.0000 303.158 0.0000 0.0000
IL10.75IL20.25 IL10.5IL20.5 IL10.25IL20.75
303.150 4.9973 0.3963 303.134 4.9984 0.7815 303.134 4.9984 1.2838 303.148 3.9999 0.3437 303.143 3.9990 0.7243 303.143 3.9990 1.0428 303.167 2.9989 0.2838 303.138 2.9971 0.6560 303.138 2.9971 0.9727 303.134 1.9974 0.2189 303.129 1.9986 0.5765 303.129 1.9986 0.8912 303.153 0.9980 0.1447 303.153 0.9986 0.4551 303.153 0.9986 0.7672 303.143 0.4992 0.0801 303.148 0.4984 0.3549 303.148 0.4984 0.6653 303.148 0.0000 0.0000 303.158 0.0000 0.0000 303.158 0.0000 0.0000
Po(IL10.75IL20.25)− Z5 Po(IL10.5IL20.5)− Z5 Po(IL10.25IL20.75)− Z5 303.162 4.9985 0.3975 303.162 4.9981 0.8066 303.153 4.9975 1.2361 303.153 3.9977 0.3387 303.153 3.9989 0.7375 303.148 3.9995 1.1493 303.138 2.9987 0.2741 303.138 2.9982 0.6485 303.153 2.9975 1.0399 303.153 1.9998 0.2053 303.153 1.9979 0.5291 303.158 1.9997 0.9130 303.148 0.9991 0.1313 303.148 0.9973 0.4091 303.153 0.9980 0.7558 303.148 0.4972 0.0865 303.148 0.4988 0.2758 303.134 0.4992 0.6280 303.205 0.0000 0.0000 303.205 0.0000 0.0000 303.148 0.0000 0.0000
PoIL1-Z5 PoIL3-Z5 ZIF-8 303.143 4.9984 0.3482 303.153 4.9978 1.9712 303.138 4.9991 2.6372 303.148 2.4980 0.1771 303.134 2.9985 1.7069 303.134 2.4986 1.3577 303.138 0.9982 0.0733 303.158 1.9997 1.4684 303.153 0.9993 0.5605 303.143 0.7480 0.0547 303.138 0.9987 1.1328 303.148 0.7474 0.4293 303.148 0.4991 0.0370 303.143 0.0000 0.0000 303.153 0.4991 0.2943 303.258 0.0000 0.0000 303.148 0.0000 0.0000
Table S3: Average absorption of CO2 by IL1, IL10.75IL20.25, IL10.5IL20.5, IL10.25IL20.75, Po(IL10.75IL20.25)−Z5, Po(IL10.5IL20.5)−Z5, Po(IL10.25IL20.75)−Z5, PoIL1-Z5, PoIL3-Z5 and ZIF-8 as function of pressure from 0–5 bar at 303 K, and by IL2 as function of pressure from 0–5 bar at 343 K. The uncertainties of temperature, pressure and gas absorption are also depicted in the table.
CO2 — Average absorption
mmol g−1
IL1 303.236± 0.064 0.0000± 0.0000 0.0000± 0.0000 303.153± 0.000 0.4990± 0.0003 0.0272± 0.0003 303.143± 0.007 0.7488± 0.0005 0.0402± 0.0002 303.158± 0.006 0.9982± 0.0005 0.0528± 0.0000 303.138± 0.000 2.4982± 0.0006 0.1308± 0.0002 303.148± 0.000 4.9970± 0.0000 0.2617± 0.0000
PoIL1-Z5 303.206± 0.074 0.0000± 0.0000 0.0000± 0.0000
6
mmol g−1
303.146± 0.004 0.4992± 0.0002 0.0360± 0.0014 303.141± 0.004 0.7479± 0.0001 0.0537± 0.0015 303.141± 0.004 0.9982± 0.0001 0.0718± 0.0021 303.151± 0.004 2.4984± 0.0005 0.1763± 0.0012 303.143± 0.000 4.9984± 0.0000 0.3482± 0.0000
IL2 339.900± 0.637 0.0000± 0.0000 0.0000± 0.0000 343.051± 0.071 0.2460± 0.0033 0.3357± 0.2354 343.157± 0.037 0.4962± 0.0059 0.4453± 0.2111 343.024± 0.021 0.9940± 0.0052 0.5546± 0.1755 343.072± 0.007 2.9904± 0.0107 0.7281± 0.0942 343.072± 0.000 4.9929± 0.0000 0.8544± 0.0000
IL10.75IL20.25
303.975± 1.169 0.0000± 0.0000 0.0000± 0.0000 303.148± 0.007 0.4995± 0.0004 0.0779± 0.0031 303.158± 0.006 0.9979± 0.0002 0.1323± 0.0175 303.151± 0.023 1.9974± 0.0001 0.2066± 0.0174 303.160± 0.010 2.9982± 0.0010 0.2740± 0.0139 303.153± 0.007 3.9990± 0.0012 0.3377± 0.0085 303.148± 0.000 4.9982± 0.0000 0.3963± 0.0000
Po(IL10.75IL20.25)− Z5 303.201± 0.006 0.0000± 0.0000 0.0000± 0.0000 303.153± 0.007 0.4979± 0.0011 0.0697± 0.0238 303.151± 0.004 0.9989± 0.0003 0.1105± 0.0294 303.144± 0.013 1.9996± 0.0004 0.1861± 0.0272 303.138± 0.000 2.9982± 0.0008 0.2599± 0.0201 303.148± 0.007 3.9983± 0.0009 0.3313± 0.0104 303.154± 0.012 4.9985± 0.0000 0.3975± 0.0000
IL10.5IL20.5
303.503± 0.488 0.0000± 0.0000 0.0000± 0.0000 303.146± 0.004 0.4984± 0.0000 0.3134± 0.0587 303.153± 0.000 0.9979± 0.0010 0.4102± 0.0635 303.139± 0.013 1.9982± 0.0006 0.5338± 0.0604 303.146± 0.011 2.9979± 0.0012 0.6298± 0.0371 303.143± 0.000 3.9992± 0.0002 0.7022± 0.0312 303.141± 0.010 4.9984± 0.0000 0.7815± 0.0000
Po(IL10.5IL20.5)− Z5 304.041± 1.102 0.0000± 0.0000 0.0000± 0.0000 303.151± 0.004 0.4988± 0.0001 0.2879± 0.0171 303.141± 0.010 0.9979± 0.0009 0.4081± 0.0014 303.151± 0.004 1.9984± 0.0008 0.5284± 0.0010 303.143± 0.007 2.9984± 0.0003 0.6389± 0.0135 303.146± 0.004 3.9986± 0.0005 0.7290± 0.0121
7
IL10.25IL20.75
304.011± 1.206 0.0000± 0.0000 0.0000± 0.0000 303.151± 0.004 0.4991± 0.0010 0.6487± 0.0234 303.148± 0.007 0.9986± 0.0001 0.7735± 0.0089 303.144± 0.021 1.9987± 0.0002 0.9152± 0.0339 303.148± 0.014 2.9976± 0.0007 1.0199± 0.0668 303.146± 0.004 3.9987± 0.0005 1.1116± 0.0974 303.146± 0.017 4.9976± 0.0011 1.2838± 0.0000
Po(IL10.25IL20.75)− Z5 303.872± 1.024 0.0000± 0.0000 0.0000± 0.0000 303.136± 0.003 0.4997± 0.0006 0.6403± 0.0174 303.151± 0.004 0.9979± 0.0002 0.7828± 0.0383 303.151± 0.011 1.9991± 0.0008 0.9315± 0.0262 303.144± 0.013 2.9979± 0.0006 1.0486± 0.0123 303.143± 0.007 3.9992± 0.0003 1.1505± 0.0018 303.153± 0.000 4.9975± 0.0000 1.2361± 0.0000
IL3 303.996± 1.185 0.0000± 0.0000 0.0000± 0.0000 303.153± 0.007 0.4992± 0.0002 0.8999± 0.0103 303.153± 0.007 0.9977± 0.0006 1.1617± 0.0280 303.151± 0.004 1.9994± 0.0002 1.4050± 0.0481 303.153± 0.007 2.9979± 0.0004 1.5655± 0.0401 303.143± 0.000 4.9980± 0.0000 1.7686± 0.0000
PoIL3-Z5 303.853± 1.004 0.0000± 0.0000 0.0000± 0.0000 303.146± 0.011 0.9988± 0.0002 1.1358± 0.0042 303.153± 0.007 1.9987± 0.0014 1.4867± 0.0259 303.136± 0.003 2.9981± 0.0006 1.6978± 0.0129 303.153± 0.000 4.9978± 0.0000 1.9712± 0.0000
ZIF-8 303.151± 0.004 0.0000± 0.0000 0.0000± 0.0000 303.153± 0.000 0.4990± 0.0001 0.2930± 0.0019 303.148± 0.000 0.7479± 0.0007 0.4272± 0.0029 303.151± 0.004 0.9987± 0.0008 0.5593± 0.0018 303.146± 0.017 2.4976± 0.0014 1.3551± 0.0036 303.138± 0.000 4.9991± 0.0000 2.6372± 0.0000
8
T
293.148 1.074089 -0.005 - - - 293.148 0.974800 -0.02 298.154 1.070444 -0.003 - - - 298.154 0.971497 0.001 303.154 1.066782 0.001 - - - 303.154 0.968351 0.003 313.154 1.059510 0.004 - - - 313.154 0.962038 0.009 232.154 1.052253 0.006 - - - 323.154 0.955778 0.01 333.154 1.045026 0.005 333.155 0.916385 -0.0003 333.154 0.949563 0.006 343.155 1.037838 -0.00006 343.154 0.910443 0.0007 343.155 0.943378 -0.002 353.154 1.030677 -0.008 353.154 0.904519 -0.0003 353.154 0.937217 -0.01
IL10.75IL20.25 IL10.5IL20.5 IL10.25IL20.75
293.145 1.055153 -0.006 293.146 1.03064 -0.001 293.149 0.994392 -0.003 298.154 1.051540 -0.0004 298.154 1.02720 -0.001 298.154 0.99112 -0.0002 303.146 1.047972 0.002 303.154 1.023751 0.0004 303.154 0.987866 0.0008 313.154 1.040864 0.002 313.154 1.020319 0.0001 313.154 0.981372 0.002 232.154 1.033742 0.004 232.154 1.016886 -0.00002 323.154 0.974893 0.001 333.154 1.026655 0.003 333.154 1.009987 0.003 333.155 0.968402 0.0005 343.155 1.019590 -0.0004 343.154 0.996266 0.001 343.153 0.961942 0.0007 353.154 1.012537 -0.005 353.154 0.989436 -0.003 353.154 0.955491 -0.002
Po(IL10.75IL20.25)− Z5 Po(IL10.5IL20.5)− Z5 Po(IL10.25IL20.75)− Z5 293.154 1.052959 -0.005 293.148 1.037482 -0.007 293.150 1.001631 -1.0 298.154 1.078146 -0.0009 298.154 1.034129 -0.004 298.154 0.998456 -1.0 298.154 1.052959 -0.002 303.154 1.030783 -0.0009 303.154 0.995288 -1.0 313.154 1.039372 0.003 313.154 1.024027 0.01 313.154 0.988973 -1.0 323.154 1.032614 0.004 323.154 1.017479 0.003 323.154 0.982661 -1.1 333.154 1.025862 0.004 333.155 1.010823 0.006 333.154 0.976339 -1.1 343.155 1.019141 0.00099 343.155 1.004243 0.001 343.154 0.970043 -1.1 353.154 1.012469 -0.007 353.155 0.997725 0.01 353.155 0.963794 -1.2
PoIL1-Z5 PoIL3-Z5 293.148 1.081692 -0.003 293.149 0.980066 -0.006 298.154 1.078146 -0.0009 298.154 0.977027 -0.0004 303.154 1.074610 0.001 303.154 0.973972 0.008 313.154 1.067575 0.001 313.154 0.967842 0.03 323.154 1.060539 0.002 323.154 0.961761 0.04 333.154 1.053488 0.004 333.154 0.957410 -0.1 343.155 1.046491 0.0006 343.155 0.949827 0.04 353.154 1.039517 -0.005 353.154 0.944049 0.02
10
Table S5: Experimental viscosities of the IL1, IL2, IL3, IL10.75IL20.25, IL10.5IL20.5 and IL10.25IL20.75 in the temperature range of 293–353 K. The deviations indicated are relative to the Vogel-Tammann-Fulcher (VTF) functions with coefficients listed in S7
T
% IL1 IL2
293.148 488.4 0.2 - - - 298.154 363.3 -0.3 - - - 303.154 275.0 -0.9 - - - 308.154 210.9 -1.2 - - - 313.154 155.4 3.8 - - - 323.154 99.89 0.8 - - - 333.154 65.97 -0.4 333.155 67.101 0.04 343.155 45.84 -2.9 343.154 43.134 0.04 353.154 32.92 -5.0 353.154 29.102 0.03
IL3 IL10.75IL20.25
- - - 293.148 459.6 0.7 298.154 253.2 0.03 298.154 337.8 -2.0 303.154 186.9 0.05 303.154 249.2 -2.6 308.154 140.9 -0.09 308.154 172.4 5.0 313.154 108.7 -0.8 313.154 133.4 3.0 323.154 65.20 1.8 323.154 84.11 -0.4 333.150 42.92 0.6 333.154 55.41 -3.0 343.154 29.64 -1.0 343.155 38.29 -6.0 353.154 21.28 -2.6 353.154 27.46 -9.0
IL10.5IL20.5 IL10.25IL20.75
293.145 593.0 0.3 293.154 719.1 -1.0 - - - 298.154 503.4 -0.02
303.146 303.3 -4.0 303.154 366.6 -1.0 308.154 207.0 2.7 308.154 270.0 -1.8 313.154 151.9 4.0 313.154 163.3 -2.8 323.154 88.14 5.0 323.154 97.83 13.0 333.154 59.53 -2.0 333.154 62.63 7.0 338.155 51.99 10.0 343.155 42.12 -0.5 348.155 36.43 14.0 353.154 29.29 -8.0
11
Table S6: ParametersA0 andA1 from linear functions used to fit the experimental densities, ρ = A0+A1T , as a function of temperature from 293–353 K and the corresponding absolute average deviation (AAD)
Sample A0/g cm−3 A1/g cm−3 K−1 AAD/% IL1 1.2862 −7.2384× 10−4 0.004 IL2 1.1141 −5.9333× 10−4 0.0005 IL3 1.1580 −6.2546× 10−4 0.007 IL10.75IL20.25 1.2632 −7.0990× 10−4 0.003 IL10.5IL20.5 1.2320 −6.8690× 10−4 0.001 IL10.25IL20.75 1.1846 −6.4881× 10−4 0.002 Po(IL10.75IL20.25)− Z5 1.2508 −6.7513× 10−4 0.003 Po(IL10.5IL20.5)− Z5 1.2318 −6.6300× 10−4 0.005 Po(IL10.25IL20.75)− Z5 1.1865 −6.3085× 10−4 1.05 PoIL1-Z5 1.2878 −7.0307× 10−4 0.002 PoIL3-Z5 1.1546 −5.9561× 10−4 0.003
Table S7: Parameters, A, B and T0, from VTF functions used to fit experimental viscosity as a function of temperature, η = Ae(B/(T −T0), and the corresponding absolute average deviation (AAD).
Sample A/mPa s B/K T0/K AAD/% IL1 0.0118 1825.30 121.43 1.8 IL2 0.0326 1241.10 170.50 0.04 IL3 0.0287 1314.33 153.48 0.9 IL10.75IL20.25 0.0277 1367.04 152.55 3.5 IL10.5IL20.5 0.0453 1170.50 169.71 5.3 IL10.25IL20.75 0.0004 2938.60 88.93 4.0
12
Characterization of the ionic liquid mixtures, IL1χIL1IL2χIL2
Figure S3: Top: experimental densities (ρ) and viscosities (η) of [P6,6,6,14][NTf2] (IL1 – black), IL10.75IL20.25 (blue), IL10.5IL20.5 (green), IL10.25IL20.75 (orange) and [P4,4,4,4][OAc] (IL2 – red) as a function of the temperature, with the respective linear and VTF fittings. Bottom: excess molar volume (V E) and viscosity deviations (η) as a function of IL2 mole fraction in the IL1+IL2 mixtures for two temperatures, 342 and 353 K.
Figure S3 (top left) and Table S4 present the measured densities of the mixtures. As the
phosphonium acetate ionic liquid3 has a lower density when compared to the [P6,6,6,14][NTf2]2,4,
the density of the mixtures decreases with the increase of acetate mole fraction, and decreases
with the temperature as expected. On the other hand, the viscosity, shown in Figure S3 (top
right) and Table S5, increases with the addition of the phosphonium acetate salt, except for the
0.25 molar fraction mixture for which the viscosity slightly decreases. This unexpected behaviour
might be related to the water content of the samples which could not be accurately measured
for these systems via coulometric titration since carboxylates and ketones reacts with the Karl
Fischer reagent releasing water. As expected, the viscosity decreases with the temperature in all
samples.
13
The excess molar volume, V E, were determined from the density values according to Equation
4:
) (S1)
where ρmix and ρIL are the densities of the mixture and of the pure ionic liquids IL1 and IL2,
respectively, χIL is the molar fraction of each ionic liquid in the mixture andMIL is the molecular
weight. The excess molar volumes at different temperatures, represented in Figure S3 (bottom
left), are positive in the whole range of IL2 molar fraction. This result suggests the existence
of weak interactions upon mixing, since the mixture volume is larger than the volume of its
components.5 V E increases with the increase in temperature and the maximum expansion being
slighly deviated towards richer compositions in IL2 (mole fractions of 0.3 and 0.5 at 342 and 353
K, respectively).
η = η − ∑
χiηi (S2)
where χi and ηi are the mole fractions and dynamic viscosities of the componentes in the mixture,
IL1 and IL2, respectively. The viscosity deviations at different temperatures, depicted in Figure
S3 (bottom right), are negative in the whole range of IL2 molar fraction. The viscosity of a
mixture depends of the entropy of solution that is related to the liquid structure. However, the
viscosity deviations depend on intermolecular interactions, size and shape of the molecules.6,7
Negative deviations are related to weak/non-specific intermolecular interactions, arising from
the size and shape of the mixture components.6,7 In all plots, η decreases with the increase
in temperature and the minimum deviation is found between 0.25 and 0.45 IL2 mole fraction.
Which might be an effect of the cation asymmetry of IL1 inhibiting the efficient packing of
components of the mixture, since none specific interactions are evidenced.
The FTIR ATR spectra of cation and anion vibrations, depicted in Figure S4, do not show
any specific interactions in the mixtures involving [P6,6,6,14][NTf2] and [P4,4,4,4][OAc], however, a
small shift towards higher energy (blue shift) is observed in the [NTf2]– anion vibrational modes
as the [P4,4,4,4][OAc] concentration increases, as shown in Figure 2 for S-N-S angle deformation
and in Supplementary Figure S5 for CF3 and SO2 asymmetric stretching and SO2 symmetric
stretching, νa(S N S), νa(CF3), νa(SO2) and νs(SO2), respectively. The same trending is
14
Figure S4: ATR spectra of IL1, IL10.75IL20.25, IL10.5IL20.5, and IL10.25IL20.75. Left: ν(S N S), the S N S bending of the NTf –2 , right: ν(COO), the COO stretching of the OAc– anion. The grey lines guide the eyes to observe the wave number shifts.
observed in the [OAc]– anion vibrational mode, ν(COO) stretching, which shifts towards higher
energy as the concentration of [P6,6,6,14][NTf2] increases in the mixtures.
These FTIR ATR results suggest that the distance between both anions with respect to the
cations is slightly longer in the mixtures than in both pure ionic liquids, where the [NTf2]− anions
are closer to the cations in the 0.25 IL2 mole fraction and the [OAc]– anions are closer to the
cations in the 0.75 IL2 mole fraction mixtures. The cation polar vibrational modes, P-C stretch-
ing and CH2 out of the plane stretching (Figure S5), also show a small shift with the increase of
the phosphonium acetate concentration, this time towards lower energies (red shift). This result
might be related to the [P4,4,4,4]+ cation contribution, which presents less energetic vibration
modes due to its higher symmetry. The terminal groups of the alkyl chains are not sensitive to
the small differences in distance between cation and anion in the mixtures (Figure S5), depending
only in the increase of the [P4,4,4,4]+ concentration. The FTIR ATR spectra bands, indicated in
the text and in the Figures S4 and S5, refer to the vibrational modes with major contribution to
the vibrational band. The ionic liquid [P4,4,4,4][OAc] spectra bands were assigned according to
Meyer et al.8 and [P6,6,6,14][NTf2] according to Cieniecka et al.9, Dharaskar et al.10 and Vitucci
et al.11 for trihexyltetradecylphosphonium bis(trifluoromethylsulfonyl)imide, trihexyltetrade-
cylphosphonium bis(2,4,4-trimethyl pentyl)phosphinate and N- trimethyl-N-hexylammonium
bis(trifluoromethanesulfonyl)imide, respectively. In conclusion, the analysis of the vibrational
spectra neither show any strong and specific interaction in the mixtures nor an effect of the
proximity of the ions due an efficient packing in the mixtures, being in agreement with the
15
viscosity and molar volume excess properties.
Figure S5: ATR spectra of IL1, IL10.75IL20.25 , IL10.5IL20.5 , and IL10.25IL20.75: (Top left) νs(SO2), the S=O symmetric stretching and the asymmetric CF3 stretching of the [NTf2]– anion; (Top right) νa(SO2), the double degenerated S=O asymmetric stretching of the [NTf2]– anion, νs(CH3) and νa(CH3) the symmetric and asymmetric CH3 stretching of the [OAc]– anion, respectively; (Bottom left) ν(P C), the P-C stretching and ν(CH2) of the [P4444]+ cation; (Bottom right) νs(CH3), νa(CH3), νs(CH2) and νa(CH2), the symmetric and asymmetric CH3 and CH2 stretching of both the [P4444]+ cation and [OAc]– anion, respectively. The grey lines guide the eyes to observe the wave number shifts.
16
NMR spectra before and after the CO2 absorption
Figure S6: Chemical structure of the [P4,4,4,4][OAc] (left) before and (right) after the absorption of CO2.
[P4,4,4,4][OAc] before and after the CO2 absorption
[P4,4,4,4][OAc] (DMSO-d6 as reference): Structure [P4,4,4,4][OAc] in Fig. S6 1H NMR : 2.24-2.16 (m, 8H, A), 1.21 (s, 3H, 2), 1.21-1.16 (m, 8H, B), 1.08-1.03 (m, 8H,
C), 0.53 (m, 12H, D) 13C NMR : 171.9, 25.6, 23.5 (d, J = 15.1 Hz), 23.3 (d, J = 4.6 Hz), 20.0 (d, J = 47.3 Hz),
12.9 31P NMR : 33.4
[P4,4,4,4][OAc] after CO2 absorption (DMSO-d6 as reference): Structure [P4,4,4,4 CO2][HOAc]
in Fig. S6 1H NMR : 16 (bs, 1H), 2.44-2.38 (m, [P4,4,4,4 CO2][HOAc](A’)), 2.17-2.10 (m, [P4,4,4,4][OAc](A)),
2.05-1.96 (m, [P4,4,4,4 CO2][HOAc](A)), 1.54-1.44 (m, [P4,4,4,4–CO2][HOAc](B’)), 1.33 (s, [P4,4,4,4][OAc](2)
and [P4,4,4,4 CO2][HOAc](2’)), 1.23-1.14 (m, [P4,4,4,4][OAc](C) and [P4,4,4,4 CO2][HOAc](C)) ,
1.11-1.02 (m, [P4,4,4,4][OAc](B) and [P4,4,4,4 CO2][HOAc](B) and (B’) ), 0.55-0.51 ((m, [P4,4,4,4][OAc](D)
and [P4,4,4,4 CO2][HOAc](D) and (D’)) 13C NMR :172.5, 165.8 (d, J = 2.7 Hz, [P4,4,4,4 CO2][HOAc]), 41.7 (d, J = 40.2 Hz,
[P4,4,4,4 CO2][HOAc](A’)), 30.0 ([P4,4,4,4 CO2][HOAc](B’)), 23.9-23.5 (massif), 23.5 (d, J = 15
Hz), 23.2 (d, J = 4.6 Hz), 18.7 (d, J = 47.6 Hz, [P4,4,4,4 CO2][HOAc](A)), 18.0 (d, J = 47.4 Hz,
[P4,4,4,4][OAc](A)), 13.2, 12.83 31P NMR : 33.3 ([P4,4,4,4][OAc]), 30.75 [P4,4,4,4 CO2][HOAc]
17
Figure S7: 1H NMR of [P4,4,4,4][OAc] before (black line, bottom) and after (red line, top) CO2 absorption at 342 K. 18
Figure S8: 13C NMR of [P4,4,4,4][OAc] before (black line, bottom) and after (red line, top) CO2 absorption at 342 K. 19
Figure S9: 31P NMR of [P4,4,4,4][OAc] before (black line, bottom) and after (red line, top) CO2 absorption at 342 K. 20
[P4,4,4,4][Lev] before and after the CO2 absorption
[P4,4,4,4][Lev] (benzene-d6 as reference): Structure (1)[P4,4,4,4][Lev] in Fig. S10 1H NMR : 2.69-2.57 (m, 10H, 1(A) and 1(3)), 2.33-2.30 (t, J = 6.7 Hz, 2H, 1(2)), 2.20 (s,
3H, 1(5)), 1.69 (m, 8H, 1(B)), 1.59 (m, 8H, 1(C)), 1.05 (t, J = 6.8 Hz, 12H, 1(D)). 13C NMR : 208.8 (1(4)), 174.1 (1(1)), 41.3 (1(3)), 34.0 (1(2)), 30.1 (1(5)), 24.5 (d,
J = 15.4 Hz, 1(B)), 24.2 (d, J = 4.5 Hz, 1(C)), 18.8 (d, J = 47.5 Hz, 1(A)), 13.9 (1(D)). 31P NMR : 33.68.
[P4,4,4,4][Lev] after CO2 absorption (benzene-d6 as reference):
Figure S10: Chemical structure of the [P4,4,4,4][Lev] along with carboxylated species formed after the absorption of CO2.
21
Considering that, in the condition of preparation of the NMR sample, the CO2 adsorption was
limited, we describe herein the spectra in two parts. First, the spectra of the ionic liquids, slightly
modified by the presence of dissolved CO2 ("main signals"), and second, the "new peaks", showing
only the signals showing formation of new species from carboxylation of the phosphonium cation
but also from carboxylation of the levulinate anion, albeit they can only be very partially
assigned, and their integrations are not relevant. In particular, the news peaks in the 210 -
197 ppm region of 13C NMR spectrum attest the presence of several products which contain
a carbonyl function, proving that the levulinate anion is carboxylated at different positions.
New peaks in the 180 - 160 ppm region are due to new carboxylate functions. The multiplet at
2.95-2.87 ppm in 1H NMR spectrum is undoubtedly assigned to the position A’ of structure 2
(noted below 2(A’)), since the edited HSQC spectrum shows a new C-H corresponding to the
α−carbon of the phosphonium cation.
Other relevant peaks are emphasized, and some assignments are proposed based on 2D-spectra
(COSY, edited HSQC and HMBC experiments) analysis.
Main peaks : 1H NMR : 2.66–2.58 (m, 10H, 1(A) and 1(3)), 2.40-2.36 (t, J = 6.7 Hz, 2H, 1(2)), 2.21 (s,
3H, 1(5)), 1.69 (m, 8H, 1(B)), 1.59 (m, 8H, 1(C)), 1.05 (t, J = 7.2 Hz, 12H, 1(D)). 13C NMR : 208.6 (1(4)), 174.6 (1(1)), 41.0 (1(3)), 33.3 (1(2)), 30.0 (1(5)), 24.4 (d,
J = 15.4 Hz, 1(C)), 24.1 (d, J = 4.5 Hz, 1(B)), 18.8 (d, J = 47.5 Hz, 1(A)), 13.8 (1(D)). 31P NMR : 33.59.
New peaks : 1H NMR : 6.2 (s, CH)), 4.49-4.46 (large s, CH, 3(3) ?), 3.70-3.64 (m, CH, 3(3) ?), 3.35-3.30
(m, CH2, 4’(3)), 2.95-2.87 (m, CH, 2(A’)). 13C NMR : 207.3, 201.0, 197.4 (4’(4)), 182.5, 179.0, 176.1, 176.0 (4’), 171.3, 170.2, 166.8,
123.3 (CH), 88.3 (CH,3(3) ?), 60.5 (CH, 3(3) ?), 35.5 (CH2, 4’(3)), 33.3 (4’(2)) 28.9, 28,2, 27.5
(d, J = 45.4 Hz, CH, 2(A’)), 23.1, 19.8, 17.8, 17.6. 31P NMR : 31.1.
22
Figure S11: 1H NMR of [P4,4,4,4][Lev] before (black line, bottom) and after (red line, top) CO2 absorption at 323 K. 23
Figure S12: 13C NMR of [P4,4,4,4][Lev] before (black line, bottom) and after (red line, top) CO2 absorption at 323 K. 24
Figure S13: 31P NMR of [P4,4,4,4][Lev] before (black line, bottom) and after (red line, top) CO2 absorption at 323 K. 25
Figure S14: 2D edited HSQC NMR spectrum of [P4,4,4,4][OAc] at 342 K using DMSO-d6 as reference. Signals are assigned according to the proposed structure.
Figure S15: 2D edited HSQC NMR spectrum of [P4,4,4,4][OAc] after CO2 absorption at 342 K using DMSO-d6 as reference. Signals are assigned according to the proposed structure.
26
Figure S16: COSY NMR spectrum of [P4,4,4,4][Lev] at 323 K using Benzene d6 as reference. Signals are assigned according to the proposed structure 1 in Fig. S10
Figure S17: 2D edited HSQC NMR spectrum of [P4,4,4,4][Lev] at 323 K using Benzene d6 as reference. Signals are assigned according to the proposed structure 1 in Fig. S10.
27
Figure S18: COSY spectrum of [P4,4,4,4][Lev] after CO2 absorption at 323 K using Ben- zene d6 as reference. Signals are assigned according to the proposed structure 1 in Fig. S10.
Figure S19: Detailed COSY spectrum of [P4,4,4,4][Lev] after CO2 absorption at 323 K using Benzene d6 as reference. Signals are assigned according to the proposed structure 4’ in Fig. S10.
28
Figure S20: 2D edited HSQC NMR spectrum of [P4,4,4,4][Lev] after CO2 absorption at 323 K using Benzene d6 as reference. Signals are assigned according to the proposed structure 1 in Fig. S10.
Figure S21: Detailed 2D edited HSQC NMR spectrum of [P4,4,4,4][Lev] after CO2 absorption at 323 K using Benzene d6 as reference. Signals are assigned according to the structures 2, 3’, and 4’ in Fig. S10.
29
30
Figure S22: Simulated 1H NMR spectra of levulinic acid and different carboxylated structures of the Lev– anion. Lev represents the linear levulinic acid, lev2 the levulinic acid with intramolecular hydrogen bonding. The following structures represent the levulinic carboxylated species along with their keto-enolic equilibrium and different intermolecular interactions. Protons highlighted in the tables correspond to hydrogen’s from the: (blue) carboxylic acid ( COOH) groups; (violet) COOH involved in hydrogen bonding intramolecular interactions; (green) hydroxyl (O H) groups involved in hydrogen bonding intramolecular interactions. E represents the energy difference between the lev2 and lev1 in the first case, E = Elev2 − Elev1, and the difference between the enol and keto forms for the following cases, E = Eenol − Eketo .
31
FTIR spectra of [P4,4,4,4][Lev] under CO2 pressure
Figure S23: Infrared spectra of [P4,4,4,4][Lev] under CO2 pressure over the exposure time at 303 K.FTIR spectra of [P4,4,4,4][Lev] under CO2 pressure over the exposure time at 303 K from 780–3200 cm−1, where the intensities along the time are offset to 0.6, 1.2, 1.8, 2.4, 3.0, 3.6, 4.2 and 4.8, respectively.
The infrared technique was also very sensitive to detect the changes in the vibrational
spectra of the [P4,4,4,4][Lev] with the increase of the CO2 pressure and the contact time. The
most sensitive bands are highlighted in Fig. S23 and shown in detail in Fig. S24. We can
clearly observe a broadening in the ν(C O) stretching mode assigned to the carbonyl and the
carboxylate groups of the anion, in the region of 1500–1900 cm−1. This feature is due to small
differences in the vibrational frequency of the different carboxylated anion species, 3, 3’, 4,
and 4’, and the superposition with the ν(C C) double bond stretching present in the enol
structures 3’ and 4’. Also, new bands became evident, such as: the broad band in the region of
2550–2800 cm−1 assigned to a superposition of the ν(O H) stretching of the new carboxylic and
the enol O H bonds; the increase of the complexity of the spectrum between 1400–1500 cm−1
32
Figure S24: Infrared spectra of [P4,4,4,4][Lev] under CO2 pressure over the exposure time at 303 K.
33
can be assigned to the mixture of carboxylated phosphonium cation and several carboxylated
levulinate anion in equilibrium, since this region of the IR spectra comprehends the ν(P C)
stretching mode of the cation and the ν(CH2) bending/deformation of both cation and anion;
the band at ca. 1270 cm−1 assigned to the enol ν(C O) stretching mode; the band at ca.
792 cm−1 can be assigned to other vibrations involving the anion ν(C C) double bond and the
different structures of the phosphonium cation; the new bands in the range of 1975–2125 cm−1
were considered as an artefact of the experiment and so were not considered in the analysis. In
conclusion, modifications of the IR spectrum confirm the presence of several new carboxylated
products that are compatible to the NMR spectra interpretations.
34
References
(1) Lepre, L. F.; Szala-Bilnik, J.; Pison, L.; Traïkia, M.; Pádua, A. A.; Ando, R. A.; Costa
Gomes, M. F. Can the tricyanomethanide anion improve CO2 absorption by acetate-based
ionic liquids? Phys. Chem. Chem. Phys. 2017, 19, 12431–12440.
(2) Avila, J.; Cervinka, C.; Dugas, P.-Y.; Padua, A.; Costa Gomes, M. Porous ionic liquids:
structure, stability and gas absorption mechanisms. Submitted
(3) Otero, I.; Lepre, L. F.; Dequidt, A.; Husson, P.; Costa Gomes, M. F. How Does the Addition
of a Third Ion Affect the Molecular Interactions and the Thermodynamic Properties of
Acetate-Based Ionic Liquids? J. Phys. Chem. B 2017,
(4) Costa Gomes, M.; Pison, L.; ervinka, C.; Padua, A. Porous Ionic Liquids or Liquid
Metal–Organic Frameworks? Angew. Chemie - Int. Ed. 2018, 57, 11909–11912.
(5) Mahajan, A. R.; Mirgane, S. R. Excess molar volumes and viscosities for the binary
mixtures of n-octane, n-decane, n-dodecane, and n-tetradecane with octan-2-ol at 298.15
K. J. Thermodyn. 2013, 1 .
(6) Kauzman, W.; Eyring, H. The Viscous Flow of Large Molecules. J. Am. Chem. Soc. 1940,
62, 3113–3125.
(7) Clará, R. A.; Gómez Marigliano, A. C.; Sólimo, H. N. Density, viscosity, vapor-liquid equi-
librium, excess molar volume, viscosity deviation, and their correlations for the chloroform
+ 2-butanone binary system. J. Chem. Eng. Data 2006, 51, 1473–1478.
(8) Meyer, R.; kyu Ha, T.; Frei, H.; Günthard, H. Acetic acid monomer: Ab initio study,
barrier to proton tunnelling, and infrared assignment. Chem. Phys. 1975,
(9) Cieniecka-Rosonkiewicz, A.; Pernak, J.; Kubis-Feder, J.; Ramani, A.; Robertson, A. J.;
Seddon, K. R. Synthesis, anti-microbial activities and anti-electrostatic properties of
phosphonium-based ionic liquids. Green Chem. 2005, 7, 855–862.
(10) Dharaskar, S. A.; Wasewar, K. L.; Varma, M. N.; Shende, D. Z.; Tadi, K. K.; Yoo, C. K.
Synthesis, characterization, and application of novel trihexyl tetradecyl phosphonium bis
35
Technol. 2014, 123, 1–10.
(11) Vitucci, F. M.; Trequattrini, F.; Palumbo, O.; Brubach, J. B.; Roy, P.; Paolone, A. Infrared
spectra of bis(trifluoromethanesulfonyl)imide based ionic liquids: Experiments and DFT
simulations. Vib. Spectrosc. 2014, 74, 81–87.
36