Improving Cellulose Dissolution in Ionic Liquids by Tuning the Size of the Ions: Impact of the Length of the Alkyl Chains in Tetraalkylammonium Carboxylate

Article abstract of DOI:10.1002/cssc.201601830

Twenty ionic liquids based on tetraalkylammonium cations and carboxylate anions have been synthesized, characterized, and tested for cellulose dissolution. The amount of cellulose dissolved in these ionic liquids depends strongly on the size of the ions:

Full text of DOI:10.1002/cssc.201601830

Accepted Article
Title: Improving cellulose dissolution in ionic liquid by tuning the
size of the ions: impact of the length of the alkyl chains in
tetraalkylammonium carboxylate
Authors: Xiang Qian MENG, Julien DEVEMY, Vincent VERNEY,
Arnaud GAUTIER, Pascale HUSSON, and Jean-Michel
Andanson
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.201601830
Link to VoR: http://dx.doi.org/10.1002/cssc.201601830
A Journal of
www.chemsuschem.org

ChemSusChem
10.1002/cssc.201601830
Improving cellulose dissolution in ionic liquid by tuning the size of
the ions: impact of the length of the alkyl chains in
tetraalkylammonium carboxylate
Xiangqian MENG, Julien DEVEMY, Vincent VERNEY, Arnaud GAUTIER, Pascale
*
HUSSON, and Jean-Michel ANDANSON
Université Clermont Auvergne, CNRS, SIGMA Clermont, Institut de Chimie de Clermont-
.
Ferrand, F-63000 Clermont–Ferrand, France
*
E-mail: j-michel.andanson@univ-bpclermont.fr
Abstract
2
0 ionic liquids based on tetraalkylammonium cations and carboxylate anions have been
synthetized, characterized and tested for cellulose dissolution. The amount of cellulose
dissolved in these ionic liquids is depending strongly on the size of the ions: from 0 to 22 wt%
of cellulose can be dissolved at 90 °C. The best ionic liquids are the less viscous and these
ammonium carboxylate based ionic liquids can dissolve as much as the imidazolium based
ones. The viscosity of an ionic liquid can be decreased by the addition of DMSO as co-
solvent. After addition of co-solvent, similar amount of cellulose per ions is reached for most
ionic liquids. As observed by rheology, ionic liquids having the longest alkyl chains form a
gel when high amount of cellulose is dissolved which limits drastically their potential.
Molecular simulation and IR spectroscopy have also been used with the aim to understand
how molecular interactions differ between efficient and inefficient ionic liquids.
1
This article is protected by copyright. All rights reserved.

ChemSusChem
10.1002/cssc.201601830
Introduction
The objective of biorefinery is to transform renewable biomass into valuable fuels,
[
1]
chemicals and materials. One of the challenges which have to be overtaken before the
industrialization of the biorefinery concept is the conversion of the biomass in a liquid state
[
2]
and in particular the dissolution of the cellulose. This step is difficult because of the strong
hydrogen bond network (intra and intermolecular) as well as the hydrophobic interactions
between the sheets of the crystalline form of this biopolymer. The dissolution of cellulose has
been studied for over a century and several processes (like Viscose using CS or Lyocell using
2
[
3]
4
-Methylmorpholine N-oxide (NMMO)) have proved their efficiency. However, the
technologies developed so far have ecological or economical drawbacks (toxicity,
recyclability). In 2002, it was published for the first time that ionic liquids could dissolve
[4]
efficiently cellulose: up to 10 wt% of cellulose in a chloride based ionic liquid at 100 °C.
Afterward, many research groups have studied a large variety of ionic liquids in order to
[
5]
improve the cellulose dissolution. The process is depending not only on the temperature but
also on the cellulose (length of the polymer chains, crystallinity, impurities), on the amount of
water in the cellulose itself and in the ionic liquids (highly hygroscopic). The most
studied/efficient ionic liquids are composed of an imidazolium based cation with either a
chloride or an acetate anion. Up to 25 wt% of cellulose have been dissolved in these ionic
[5-6]
liquids at 100–120 °C.
This is comparable to the efficiency of the solvents used in
industry with the advantage of allowing an easier recycling of the solvent because of the
negligible vapor pressure of the ionic liquids. Chloride based ionic liquids tend to have a
relatively high melting temperature which might limit their applications. For example, the
[7]
melting point of the 1-butyl-3-methylimidazolium chloride ([Im ]Cl) is 69 °C. Carboxylate
1
4
based ionic liquids are liquid at lower temperatures but suffer stability issues and should be
[
8]
used below 120 °C while used for long process.
The search for new ionic liquids able to dissolve efficiently cellulose remains very active.
An elucidation of the dissolution process could largely help in the selection of the best
candidates. The role of the anion in the dissolution is now relatively well understood. It
interacts through strong H-bonds with the hydroxyl groups of the cellulose. Using COSMO-
RS (Conductor-like Screening Model for Real Solvents) Liu et al. have modeled the ability of
[
9]
hundreds of ionic liquids to dissolve cellulose. The calculated excess enthalpy was used to
estimate how an ionic liquid interacts with the cellulose. These authors have concluded that
the role of the cation is relatively small compared to the one of the anion. Molecular
simulations studies have shown that while the anion takes place around the hydroxyl groups
2
This article is protected by copyright. All rights reserved.

ChemSusChem
10.1002/cssc.201601830
of the cellulose, the cation fits the apolar part of the cellulose in order to obtain the best
[
10]
interactions between solvent and cellulose. Lindman et al. observed that only amphiphilic
solvents can dissolve cellulose meaning the hydrophobic interactions between the solvent and
[
11]
the cellulose are also necessary for the dissolution.
The addition of an aprotic and polar co-solvent like dimethylsulfoxide (DMSO), N,N-
dimethylformamide (DMF) or N-methylpyrrolidinone (NMP) to an efficient ionic liquid
[
12]
improves the kinetics of cellulose dissolution. Nevertheless, the exact role of the co-solvent
is still not perfectly understood. Interactions between mixtures of ILs (imidazolium
acetate)/DMSO and cellulose or cellobiose were studied through the measurement of
[
13]
mixing/dissolution enthalpy using calorimetry.
A slightly lower exothermicity was
measured with the addition of DMSO as a co-solvent. This reduction of the enthalpy indicates
less favorable interactions. Moreover, in a system cellulose + ionic liquid (imidazolium
acetate) / DMSO (80/20 wt/wt), no direct molecular interaction between DMSO and cellulose
[14]
was detected by molecular simulation. Ionic liquids are generally highly viscous and their
viscosity can increase by up to 3 orders of magnitude in the presence of 10 to 15 wt% of
[
15]
cellulose in it. The addition of DMSO as a co-solvent to an already highly efficient ionic
liquid (typically acetate anions) improves the dissolution by mainly decreasing the
[
16]
viscosity. The research for other co-solvents is still very active. Greener alternative like γ-
[
17]
valerolactone are explored.
Biphasic mixtures composed of DMI (1,3-dimethyl-2-
imidazolidinone), ILs and cellulose are also studied to simplify the separation and recycling,
[
18]
as demonstrated in a recent work.
Ammonium-based ionic liquids have demonstrated their ability to dissolve cellulose with
[
19]
the advantage to be relatively cheap. Several groups have evaluated the efficiency of this
family for cellulose dissolution with various results. Triethyloctylammonium chloride ([N2228
]
[
20]
Cl) can dissolve up to 15 wt% of cellulose,
diethyl-dimethylammonium acetate
[
21]
(
[N1122][OAc]) only 2 wt% at 100 °C, while in a third study several ionic liquids based on
tetraalkylammonium (([N1333], [N1133] and [N1144]) acetate, levulinate and itaconate were able
[17b]
to dissolve up to 10 wt% of cellulose at 90–110 °C.
As for the imidazolium, the addition
of a co-solvent to these ionic liquids increases the amount of cellulose dissolved up to 20 wt%
[
17b]
at 90 °C.
The co-solvent is also useful for ionic liquids presenting high melting points. For
example, in the case of tetrabutylammonium acetate ([N4444][OAc]), which has a melting
temperature around 100 °C, after the addition of a large amount of DMSO it has been found
[
22]
by Huang et al. that up to 8 wt% of cellulose could be dissolved. The effect of the presence
of oxygenated functions in the alkyl chain of the ammonium was also considered. Ammonium
3
This article is protected by copyright. All rights reserved.

ChemSusChem
10.1002/cssc.201601830
based ionic liquids with ether groups like [N122,1O2] (N,N-diethyl-N-(2-methoxyethyl)-N-
[
23]
methylammonium) can dissolve cellulose up to 12 wt% at 100 °C, while in another study,
[24]
it was found 18 wt% in [N11,2O2,2O2][OAc] at 80 °C.
In the present work, our objective is to study the effect of the ions’ size of ammonium
based ionic liquids on their ability to dissolve cellulose. For that purpose, 20 ionic liquids
containing a tetraalkylammonium cation and a carboxylate anion (acetate [OAc] and
propionate [OPr]) have been synthesized. The cations considered are triethylalkylammonium
([N222X]) and tripropylalkylammonium ([N333X]) where X=4, 6, 8, 10 and 12 (butyl, hexyl,
octyl, decyl and dodecyl). In this set of 20 ionic liquids, molecular interactions between the
ions and the cellulose are not expected to change substantially and only size of the ions should
explain the variation of cellulose solubility. As the addition of a co-solvent is known to
improve the dissolution of cellulose, the 20 ionic liquids have also been studied with addition
of a co-solvent (20 wt% DMSO / 80 wt% ionic liquid). In parallel to the dissolution
experiments, viscosity of solvents and rheology of (cellulose + solvent) mixtures were
measured with the aim to evaluate if transport properties could limit the dissolution. IR
spectroscopy and molecular simulation were used to evaluate the molecular interactions of
these systems.
Results and Discussion
The temperature generally used in studies for cellulose dissolution in ionic liquid is
[
5]
between 80 to 120 °C. However, we have noticed that ammonium carboxylate based ionic
liquids are not stable at high temperatures. During cellulose dissolution in [N2226][OAc] at
1
00 °C a strong coloration of the sample was observed within few hours which was not
observed at 80°C. To explore thermal stability of this ionic liquid, pure [N2226][OAc] was
1
heated at 100°C for 24 hours under inert atmosphere and the sample was then analyzed by H
NMR and IR spectroscopies. NMR analysis shows that 20–30% of ethyl and hexyl acetates
-1
(
IR: = 1730 cm ) are produced in a ratio 3/1 respectively accompanied by the same amount
of triethylamine. This is explained by the random displacement of the alkyl groups by the
nucleophilic acetate at high temperature. At 80°C, the same experiment leads to only traces
amount of decomposition (less than 1% after 24h and approximately 2% after 48h). Thus, our
study will mainly focus on dissolution of cellulose at 80 °C or lower.
4
This article is protected by copyright. All rights reserved.

ChemSusChem
10.1002/cssc.201601830
Temperature of dissolution
Figure 1: Comparison of the temperature of dissolution of 10 wt% of microcrystalline
cellulose in pure [N222X][OAc] (top) and in mixtures of [N222X][OAc]/DMSO (80/20 wt%)
(bottom).
The impact of the ion’s size is first studied by comparing the temperature of dissolution of
1
0 wt% of cellulose in the 20 ionic liquids. The evolutions of the brightness obtained for the 5
triethyl-alkyl-ammonium acetate ([N222X][OAc], X=4, 6, 8, 10 and 12) are given in Figure 1.
In the case of [N2224][OAc], the brightness is increasing during the experiment and cellulose is
not dissolved. The experiment was repeated with only 1 wt% of cellulose and even in that
case no dissolution could be observed. For the other 4 ionic liquids, all the cellulose is
dissolved at temperatures between 70 and 90 °C.
5
This article is protected by copyright. All rights reserved.

ChemSusChem
10.1002/cssc.201601830
The impact of the addition of a co-solvent can also be observed in Figure 1. Even with the
addition of 20 wt% of DMSO as co-solvent, [N2224][OAc] is still not able to dissolve any
cellulose. For the other 4 ionic liquids, the addition of DMSO helps to decrease the
temperature of dissolution (between 50 and 60°C instead of 70 – 90 °C without DMSO). In
fact, in all these cases the dissolution starts as soon as the experiment. In the case of the two
longer chains (X=10 and 12), it exists a shoulder (10% in intensity of the maximal brightness)
which is an indication that the last fragments of cellulose are more difficult to dissolve (origin
of the shoulder will be discussed in more details in the rheology section). To compare all 20
ionic liquids, with and without co-solvent, Tonset and T10% (°C) are given in Table 1. Tonset
corresponds to the beginning of the dissolution (as used in DSC (differential scanning
calorimetry to define the beginning of an event) while the T10% corresponds to the end to the
dissolution of 10 wt% of cellulose (temperature of the first image without any particle).
6
This article is protected by copyright. All rights reserved.

ChemSusChem
10.1002/cssc.201601830
Table 1: Temperature (°C) of dissolution of 10 wt% of microcrystalline cellulose with a
heating ramp of 0.25°C/min up to 90 °C. Tonset and T10% correspond to the beginning and the
end of the dissolution, respectively. Solvents with DMSO are composed of 80/20 (wt/wt) of
ionic liquid and DMSO. (-) no cellulose dissolution observed, (*) dissolution begins as soon
as experiment is started.
[
N222X
]
[N222X] [N333X] [N333X
]
Tonset (°C)
[
OAc]
-
51
51
57
[OPr]
-
[OAc] [OPr]
X=4
X=6
X=8
X=10
X=12
X=4
-
-
-
60*
60*
70*
67*
-
80
74
77
-
77
85
-
68*
-
-
-
X=6
X=8
35*
35*
34*
31*
34*
32*
31*
32*
43
38*
36*
36*
-
66
67
68
X=10
X=12
[
[
N222X
OAc]
-
]
[N222X] [N333X] [N333X
]
T10% (°C)
[OPr]
-
[OAc] [OPr]
X=4
X=6
-
-
-
-
73
79*
77*
90*
89*
-
X=8
79
80
89*
-
-
-
X=10
X=12
X=4
-
-
-
-
-
-
X=6
46*
52*
55*
47*
62*
48*
62*
59*
90
62*
69*
71*
-
X=8
X=10
X=12
84
78
-
The analysis of the dissolution temperatures on the 20 ionic liquids leads to the following
remarks:
With or without co-solvent, ionic liquids with the short butyl chain ([N2224][OAc],
N2224][OPr], [N3334][OAc] and [N3334][OPr]) are not able to dissolve cellulose. Even with
only 1 wt% of cellulose, no dissolution was observed.
Without co-solvent, the temperatures obtained for [N222X] are lower than for [N333X
cation. Moreover, temperature of dissolution in [N222X][OAc] are lower than in [N222X][OPr].
For all ionic liquids efficient for the cellulose dissolution, as already observed with
-
[
-
]
-
Figure 1 in the case of [N222X][OAc], the addition of co-solvent decreases the temperature of
7
This article is protected by copyright. All rights reserved.

ChemSusChem
10.1002/cssc.201601830
dissolution by more than 10 °C, in average around 20 °C. After addition of co-solvent, most
ionic liquids are starting to dissolve cellulose below 40 °C. For 6 ionic liquids, the presence of
DMSO allows the dissolution of 10 wt% of cellulose that was impossible in the pure ionic
liquid (case of [N333X][OAc] with X=6, 8, 10 and 12 of [N333X][OPr] with X=8 and 10).
Quantity of cellulose dissolved
Figure 2: Evolution of the brightness during dissolution of several amounts of
microcrystalline cellulose in [N2228][OAc] with and without co-solvent (80/20 wt%) with a
heating ramp of 0.25°C/min.
Dissolution of different quantities of cellulose has been realized in order to estimate the
maximal amount that can be dissolved in each ionic liquid. As example, few dissolution
experiments realized on [N2228][OAc] with and without co-solvent are given in Figure 2.
Without DMSO, the dissolution process occurs mainly between 50 and 70 °C. 10 wt% of
8
This article is protected by copyright. All rights reserved.

ChemSusChem
10.1002/cssc.201601830
cellulose is fully dissolved at 80 °C while 15 wt% is almost completely dissolved at 90 °C.
The brightness is stable and still over 50 with 19 wt% of cellulose at the end of the
experiment (90°C).
With DMSO the cellulose dissolves 20 °C below the pure ionic liquids. Up to 20 wt% of
cellulose can be dissolved before reaching 70 °C, but the brightness stays high for 22 and
2
5 wt% of cellulose. The large difference of brightness for only few percent of undissolved
cellulose illustrates the sensitivity of this technique to detect the last particles of cellulose. The
maximal amount of cellulose that can be dissolved at 80 °C is then 12 wt% in pure
[
N2228][OAc] (another experiment was done with this amount of cellulose but is not shown in
Figure 2 for clarity reason) and 20 wt% in [N2228][OAc]/DMSO.
With this procedure, the amount of dissolved cellulose can be underestimated up to 2 wt%
because not all the cellulose quantities were tested. After the analysis of the 20 ionic liquids
with this protocol, it appears that the highest amount of cellulose dissolved is 15 wt% at 80 °C
and 22 wt% at 90 °C in [N2226][OAc]. The results at 80 °C (heating rate of 0.25 °C/min) for
all ionic liquids with and without co-solvent are given in Figure 3.
Figure 3: Maximal amount of microcrystalline cellulose dissolved at 80 °C with a heating rate
of 0.25 °C/min for 20 ionic liquids without and with 20 wt% of DMSO as co-solvent.
[
N2224][OAc], [N2224][OPr], [N3334][OAc] and [N3334][OPr]) cannot dissolve any cellulose.
Without co-solvent, 5 ionic liquids are capable to dissolve at least 10 wt% of cellulose at
8
0 °C ([N2226][OAc], [N2228][OAc], [N22210][OAc], [N2226][OPr] and [N2228][OPr]). Without
+
co-solvent, none of the [N333X] based ionic liquids are capable of dissolving any cellulose.
9
This article is protected by copyright. All rights reserved.

ChemSusChem
10.1002/cssc.201601830
The addition of co-solvent increases the amount of cellulose dissolved, up to 20 wt% in
N2226][OAc]/DMSO and [N2228][OAc]/DMSO and 12 ionic liquids are capable of dissolving
[
at least 10 wt% of cellulose. A correlation exists between the amount of cellulose dissolved at
0 °C in pure ionic liquids and the temperature of dissolution; the ionic liquids capable to
8
dissolve at the lowest temperatures are also those dissolving the most.
Number of anion per glucose unit
The weight of cellulose dissolved per weight of solvent is not necessarily the most accurate
method to compare the efficiency of ions. The strongest interactions between the ionic liquids
and cellulose are H-bonds between the anions and the hydroxyl groups of the cellulose. The
ability of an ionic liquid to dissolve cellulose can be expressed by the APG ratio (moles of
Anions Per moles of Glucose units) at the highest amount of cellulose dissolved in each ionic
liquid. Lower is this ratio, better is the solvent. Every acetate anion can be acceptor of 2 H-
bonds while every glucose unit of the cellulose can be donor of 3 H-bonds with 3 hydroxyl
groups (with the exception of the end groups of the polymer chain). Thus, in theory, the
lowest APG ratio would be 1.5 (obtained when all anions interact with every hydroxyl group
of the cellulose). The APG ratio values obtained in the present work are given in Table 2.
Table 2: Number of anions per glucose unit (APG ratio) calculated from the dissolution
experiments presented in Figure 3. (-) less than 1 wt% of microcrystalline cellulose could be
dissolved. The number is in red when the ionic liquid is giving a better result than the
reference ionic liquid [Im ][OAc].
12
[
N222X] [N222X] [N333X] [N333X
]
[
OAc] [OPr] [OAc] [OPr]
X=4
X=6
X=8
X=10
X=12
X=4
-
-
-
-
-
-
-
-
-
-
-
-
4.2
4.8
5.4
-
6.3
5.6
-
-
-
-
-
-
X=6
X=8
X=10
X=12
2.4
2.1
2.2
2.7
2.7
2.3
2.1
3.8
9.5
2.6
2.4
3.5
-
5.0
2.3
4.3
1
0
This article is protected by copyright. All rights reserved.

ChemSusChem
10.1002/cssc.201601830
The APG ratio obtained without co-solvent for the ammonium based ionic liquids are
between 4.2 and 6.3. As comparison, the ability of a benchmark ionic liquid (1-ethyl-3-
methylimidazolium chloride acetate, [Im ][OAc]) has been evaluated for cellulose
1
2
dissolution with the same procedure. A value of 5.1 anion/glucose unit (corresponding to
7 wt% of cellulose) was obtained in this pure ionic liquid. Two of the ammonium based
1
ionic liquids ([N2226][OAc] and [N2228][OAc]) are thus more efficient than [Im ][OAc]. These
12
two ionic liquids are the two based on [N222X] which are liquids at room temperature. This
may indicate that the reason why imidazolium are good cations for cellulose dissolution is not
related only to the aromaticity or to the acidic proton of the imidazolium ring as proposed in
[
25]
literature.
The dissolution of cellulose in [Im ][OAc] with the addition of 20 wt% of DMSO was
1
2
also measured in the present work. The addition of co-solvent is almost not changing the mass
of cellulose dissolved (18 wt%) per mass of solvent (ionic liquid + DMSO). Thus it induces a
decrease by around 20% of APG (equal to 3.9 in [Im ][OAc]/DMSO). In presence of DMSO,
1
2
1
2 ammonium based ionic liquids are capable of dissolving higher amounts of cellulose
compared to the [Im ][OAc] + DMSO (in red in Table 2). This approach confirms roughly
1
2
the results observed when working with the mass of dissolved cellulose (Figure 3). However,
by using the APG ratio, a more precise description of the impact of alkyls’ sizes can be
obtained. For example, [N33310][OPr]/DMSO seems much less efficient than
[
N2226][OAc]/DMSO when using the wt% (15 vs 20 wt%). In fact the cellulose dissolution is
similar in both ionic liquids when looking at the APG ratio meaning at molecular level, the
number of ions interacting with cellulose is the same in both ionic liquids.
In summary, two ionic liquids ([N2226][OAc] and [N2228][OAc]) are capable to dissolve
high amounts of cellulose even without co-solvent at 80 °C. The addition of DMSO decreases
the dissolution temperature, allowing 12 ionic liquids to dissolve more cellulose than the
benchmark imidazolium based ionic liquid. Among these, the 10 most efficient are all
dissolving similar amounts of cellulose (corresponding to 2.5 APG). Ionic liquids with
[
NYYY6], [NYYY8] and [NYYY10] cations are the most efficient for cellulose dissolution. The
longest [NYYY12] are not as effective while the shortest [NYYY4] cannot dissolve any cellulose.
In the second part of this article, several properties of these ionic liquids will be measured
and correlated with their ability to dissolve cellulose. The main objective is to identify which
properties are 1) controlled by the size of the ions and 2) impacting their ability to dissolve
cellulose.
1
1
This article is protected by copyright. All rights reserved.

ChemSusChem
10.1002/cssc.201601830
Viscosity
The presence of 20 wt% of DMSO tends to improve the ability to dissolve cellulose
dissolution at lower temperatures and higher amounts of cellulose dissolved). One of the
(
main effects of the addition of a co-solvent in an ionic liquid is to decrease considerably its
viscosity. The viscosities of pure ammonium based ionic liquids have been measured as a
function of temperature and are given in the Figure 4. The viscosity of only 9 of the 20 ionic
liquids could be measured for practical reasons. As the quantity of water in the samples is
typically 0.5 wt%, we estimate that the measured viscosity is lower, by at least 20%, than the
[26]
viscosity of the dried ionic liquids. The fluctuations in the amount of water between ionic
liquids should induce errors of around 10 to 20% in between two ionic liquids. Even though,
comparisons and conclusions can be established.
Figure 4: Viscosity of 9 ammonium carboxylate based ionic liquids (with around 5000 ppm of
water). Symbols: circles for acetate, squares for propionate based ionic liquids, full symbols
for [N222X] and empty symbols for [N333X] cations. For the length of the longer alkyl chain of
the cation: red = hexyl, green = octyl, blue = decyl and black = dodecyl.
The less viscous ionic liquid is [N2226][OAc] with a viscosity of 1200 mPa.s at 25 °C,
[27]
much higher than the one of imidazolium based ionic liquids: 139 mPa.s for [Im ][OAc]
1
2
[28]
and 441 mPa.s for [Im ][OAc] at the same temperature. The 3 ionic liquids with the lowest
1
4
viscosities are [N2226][OAc] (88 mPa.s at 60°C), [N2226][OPr] (102 mPa.s at 60°C) and
N2228][OPr] (121 mPa.s at 60°C). The other 6 ionic liquids, with longer alkyl chains, have
much higher viscosities (between 178 and 250 mPa.s at the same temperature). This increase
[
1
2
This article is protected by copyright. All rights reserved.

ChemSusChem
10.1002/cssc.201601830
of the viscosity with the alkyl chain length was already observed on other ammonium ionic
[
29]
liquids.
The more viscous ionic liquids are also those not dissolving cellulose without co-solvent
indicating viscosity is probably a limiting parameter. The addition of 20 wt% of DMSO
decreases by around 80% the viscosity. In presence of DMSO all ionic liquids are largely less
viscous than any pure; the maximum viscosity is 31 mPa.s at 60 °C after addition of DMSO
(for [N3334][OAc]/DMSO) while the minimum viscosity of a pure ionic liquid is 88 mPa.s at
6
0°C. If viscosity would be the only parameter that explains fluctuation of the ability to
dissolve cellulose within similar ionic liquids, this would mean that after addition of DMSO,
all ionic liquids should have similar efficiency. This was observed for most ionic liquids (with
an APG ratio around 2.5 as can be seen in Table 2) but not in all cases. For example, [NYYY12
]
has the tendency to be less efficient and [NYYY4] based ionic liquid cannot dissolve any
cellulose. The fact that some ionic liquids are less effective ([NYYY12]) or unable ([NYYY4]) to
dissolve cellulose in presence of DMSO cannot be explained by their viscosity.
1
3
This article is protected by copyright. All rights reserved.

ChemSusChem
10.1002/cssc.201601830
Rheology
Figure 5: Complex viscosity modulus (η*) of systems composed of 10 wt% of
microcrystalline cellulose in [N2226][OAc]/DMSO (a) and [N33312][OAc]/DMSO (b) after
dissolution. (c) Evolution of the brightness of a sample containing 10 wt% of microcrystalline
cellulose in [N2226][OAc]/DMSO and [N33312][OAc]/DMSO as function of the temperature
with a heating rate of 0.25 °C/min.
To explore in more details the effect of transport limitations, rheological measurements at
different temperatures have been done on [N2226][OAc]/DMSO and [N33312][OAc]/DMSO
containing 10 wt% of pre-dissolved cellulose. [N2226][OAc]/DMSO is one of the efficient
solvent (APG ratio = 2.4) while [N33312][OAc]/DMSO is one of the poorly efficient solvent
1
4
This article is protected by copyright. All rights reserved.

ChemSusChem
10.1002/cssc.201601830
(APG ratio = 3.5). This later can dissolve cellulose but not as much as other
tetralkylammonium based ionic liquids. The evolution of the brightness during the dissolution
of 10 wt% of cellulose in these two ionic liquids using optical microscopy is presented in
Figure 5c. In Figure 5 (a and b), the moduli of the complex viscosity * is plotted versus the
frequency for both ionic liquids at different temperatures. In the case of [N2226][OAc]/DMSO
(7a) a unique viscoelastic behavior can be observed with viscosities ranging from about 1 to
1
00 Pa.s. [N33312][OAc]/DMSO presents a different viscoelastic behavior (7b). For
temperatures higher than 70 °C the same behavior as previously can be observed, but, for
temperatures lower than 70 °C a sharp transition in the viscoelastic behavior can be observed
(about two decades in viscosity) between an entangled viscous liquid behavior to a gel-solid
like behavior. At the lowest temperature the material behaves like a gel with an infinite
viscosity at zero frequency and even the addition of a co-solvent as DMSO does not allow
preventing the formation of a gel.
From Figure 5 (c), it can be seen that the two ionic liquids behave differently in brightness
analysis. In the case of [N33312], a shoulder appears in the brightness curve between 40 and
7
0°C (similarly to what was observed in Figure 1). This result is consistent with the formation
of a gel observed during the rheology experiments. The viscosity of the solution is very high
leading to great diffusion times. It has already been reported that imidazolium acetate based
[
30]
ionic liquids also induce a gel phase when more than 12 wt% of cellulose is dissolved.
When a gel is generated during the dissolution, the time needed to complete the dissolution
will be dramatically increased and therefore ionic liquids that avoid gelation should be
preferred.
Spectroscopy
When diffusion is not the limiting parameter in the dissolution process, as in the case of
[
NXXX4] based ionic liquids, a careful study of the interactions (between anions, cations and
cellulose) can be helpful to understand the solvents efficiency in cellulose dissolution.
1
5
This article is protected by copyright. All rights reserved.

ChemSusChem
10.1002/cssc.201601830
Figure 6: Part of the infrared spectra of [N2224][OAc] and [N2226][OAc]. Spectra have been
shifted for clarity.
A strong interaction between the anion and cation might explain why some ionic liquids
([NXXX4] in the present case) cannot dissolve cellulose. A competition to create H-bond with
the acetate anion might exist between cation and cellulose and a stronger interaction between
anion and cation could decrease the ability of the ionic liquid to interact with the cellulose and
to dissolve it. The strength of the interactions between anions and cations can be observed by
IR spectroscopy. Figure 6 presents the IR spectra of [N2224][OAc] and [N2226][OAc],
respectively a non-efficient and the most efficient ionic liquid. In dried ionic liquids (0.5 wt%
-
-1
-
H O), the positions of the COO asymmetric ( at 1580 cm ) and symmetric ( at 1365 cm
2
as
s
1
)
stretching modes are not depending on the cation. These two bands are an efficient probe of
the strength of the interactions with the carboxyl group. Only the symmetric stretching mode
-1
is shifted by 1 cm . As the comparison, with imidazolium cation ([Im ][OAc]), these two
1
4
-1
-1 [31]
modes are shifted by 20 cm at 1560 and 1384 cm . Imidazolium cations are interacting
stronger than ammonium with their acid proton of the aromatic ring. The specific interaction
of the imidazolium ring does not appear necessary to improve the properties of the solvent. In
fact, the absence of H-bond donor in ammonium cation could benefit to the interaction
between anion and cellulose since the anion is only able to make H-bonds with the hydroxyl
groups of the cellulose.
Similarly, when water is absorbed in the ammonium based ionic liquids, water molecules
are interacting via H-bonds with the acetate and both bands are also shifted; as example, in
[
N2226][OAc] with 5–10 wt% of water (enough for water molecules to interact with every
1
6
This article is protected by copyright. All rights reserved.

ChemSusChem
10.1002/cssc.201601830
-
-1
anion), the two COO stretching modes are at 1573 and 1383 cm . It appears then clearly that
both ammonium cations interact similarly and weakly with the carboxylate anion. The
strength of the interaction between ions cannot explain why ionic liquids with the smaller
cations cannot dissolve cellulose.
Molecular structure
Table 3: Percentage of hydroxyl groups of a cellohexaose (Cell.) involved as donor in a
hydrogen bond in 4 solvents: [N2224][OAc], [N2226][OAc] with and without DMSO calculated
by molecular simulation.
Donor
Cell.
Cell.
6 %
6 %
7 %
7 %
Acceptor
Anion
74 %
70 %
74 %
70 %
DMSO
[
[
[
[
N2224
]
]
-
N2226
-
N2224]/DMSO
N2226]/DMSO
0.2 %
0.4 %
The interactions between ions and cellulose were studied by molecular simulation with the
aim to clarify why the smaller [N2224] cannot dissolve cellulose as efficiently as [N2226] cation.
A cellulose oligomer (cellohexaose) has been simulated in 4 solvents [N2224][OAc] and
[
N2226][OAc] with and without DMSO. The structure of the solvents around the oligomer has
been carefully studied in particular with radial distribution functions. No observable
difference between the structures of the solvents could explain why ionic liquids with the
smaller [N2224] cation could be a worse solvent. The most important interaction between the
ions and cellulose is the H-bonds between the anion and the hydroxyl groups of the
[
32]
cellulose. The probabilities of the hydroxyls groups of the oligomer to form H-bonds with
the solvents are given in Table 3. Each hydroxyl group of the cellulose oligomer is donor of
an H-bond most of the time; between 76 and 81%. Generally the acceptor of the H-bond is the
acetate anion (70–74%), sometime the H-bond is intramolecular (6–7%) and the co-solvent is
almost never directly involved with the oligomer (0.2–0.4%). The number of H-bonds is even
higher in the case of [N2224][OAc] than for [N2226][OAc] indicating than [N2224][OAc] should
be a very efficient solvent for cellulosic polymer chains. With their molecular simulation
study Cho et al. explained that the cation interacts also with the cellulose and mainly via the
[
32]
side chain and the linker oxygens between the glucose cycles. Structure and interactions
1
7
This article is protected by copyright. All rights reserved.

ChemSusChem
10.1002/cssc.201601830
between the cation and the oligomer was scrutinized and no difference could be observed
between the two ionic liquids. These results from molecular simulation shows that
energetically [N2224][OAc] seems a better solvent for an oligomer of cellulose than
[
N2226][OAc].
Discussion
Size of the ions can have a dramatic impact on the ability of an ionic liquid to dissolve
cellulose. Ionic liquids studied in this work can be ranked in 4 groups:
-
efficient ionic liquids, capable to dissolve at least 10 wt% of cellulose even without co-
solvent: only based on [N222X] cations with X=6, 8 or 10. [N2226][OAc] is dissolving up to
5 wt% of cellulose at 80 °C (22 wt% at 90 °C and 20 wt% in presence of DMSO at 80 °C).
1
The ability to dissolve cellulose in these ammonium based ionic liquid is comparable to this
of the benchmark imidazolium based ionic liquid ([Im ][OAc]).
1
2
-
efficient ionic liquids only after addition of co-solvent. While in pure ionic liquids, only the
less viscous ionic liquids are capable to dissolve cellulose at 80 °C or below, after addition of
DMSO, 10 ionic liquids are dissolving similar amount of cellulose: around 2.5 anions per
glucose unit. The addition of 20 wt% of DMSO decreases enough the viscosity to overcome
this limitation of some ionic liquids.
-
poorly efficient ionic liquids which tend to dissolve less cellulose even with 20 wt% of
DMSO. These ionic liquids form a gel when dissolving cellulose even in presence of co-
solvent. The temperature should then be increased to suppress the gel structure and continue
the dissolution. This effect has particularly been observed with ionic liquids having the
longest alkyl chains: [NYYY12].
-
inefficient ionic liquids are generally based on the smallest [NYYY4] cations. These ionic
liquids are not able to dissolve any cellulose at temperature below 80 °C. The viscosity or
gelation cannot explain this impossibility to dissolve cellulose since even small amounts of
cellulose in presence of co-solvent cannot be dissolved. As shown by infrared spectroscopy,
the interactions between anion and cation are similar in [N2224][OAc] and in [N2226][OAc].
Moreover, molecular simulation illustrated that the direct molecular interactions responsible
of cellulose dissolution (namely H-bonds), are equivalent between cellulose oligomer and the
ions of [N2224][OAc] or [N2226][OAc]. Dissolution of crystalline cellulose is a complex
phenomenon and a cooperative interaction between many ions and cellulose has to exist in
order to start the cellulose dissolution.
1
8
This article is protected by copyright. All rights reserved.

ChemSusChem
10.1002/cssc.201601830
Conclusions
The ability to dissolve cellulose at 80 °C has been established for 20 tetraalkylammonium
carboxylate. The ionic liquids could be classified in 4 groups from highly efficient to
inefficient cellulose solvents demonstrating the effect of the size of the ions on the dissolution
ability. Viscosity, rheology, spectroscopy and molecular simulation were used to correlate the
ionic liquid's properties with their ability to dissolve cellulose. The inefficiency of the
smallest cations tested could also be due to an insufficient interaction with the hydrophobic
parts of the cellulose. Complementary studies will be necessary to test how the amphiphilic
structures of the cellulose and the ionic liquid associate with efficient and inefficient ionic
liquids (e.g. with [N2226][OAc] and [N2224][OAc]). In particular, more advanced molecular
[
10]
simulations with cellulose microfibrils
could describe the molecular organization at the
interface between the liquid and the undissolved polymer.
Experimental Section
Materials
Avicel PH101 with particle size of 50 m was supplied by Sigma and dried at 100 °C for
2
4 hours before use. DMSO (puriss, > 99.9%), tripropylamine (>98%), 1-bromobutane
(>98%), 1-bromooctane (>98%), 1-bromooctane (>98%) and 1-bromodecane (>98%) were
obtained from Fluka. Triethylamine (99%) was obtained from Acros while acetic (>99%) and
propionic (99%) acids were purchased from VWR and Merck, respectively. [Im ][OAc] was
1
2
obtained from Iolitec (>98%). Syntheses were carried in parallel (10) using a caroussel
reaction station (Radleys). pH was recorded on a Metler Toledo, seven compact.
Synthesis
Tetraalkylammonium bromide. In each reaction tube of the caroussel station were
introduced 10 mL of acetonitrile, 44.6 mmol of triethyl or tributyl amine and 37.1 mmol of
the desired bromoalcane. The temperature is increased to 85°C and the reaction mixture is
h
stirred 24–48 . The reaction mixture is purred in 150 mL of water and extracted with ether
(2X50mL). Evaporation of the aqueous phase affords a viscous oil which is left under reduced
pressure (1 mbar) for 48h to afford a white solid.
Tetraalkylammonium acetate and propionate. The tetraalkylammonium bromide were
dissolved in 150 mL of water and passed through a Dowex® 1X2 (50 grammes) column
charged with hydroxide ions followed by washing with 250 mL of water. The volume was
reduced to 150 mL under reduced pressure. The column was recharged with NaOH, then
1
9
This article is protected by copyright. All rights reserved.

ChemSusChem
10.1002/cssc.201601830
washed with water under pH reach 7, and the exchange of anion of tetraalkylammonium was
repeated. This process was performed at least 3 times. The final aqueous solution was
concentrated under reduced pressure to 150–200 mL, then neutralized to pH=5.2 (pHmeter)
with the appropriate carboxylic acid. Evaporation under vacuum of the solution furnished a
colorless liquid. Evaporation under vacuum (0.1 mbar) for at least 48h yields the desired ionic
liquid as a pale yellow product. Water content (measured by KF titration) in ionic liquid was
systematically lower than 1 wt% and generally around 5.000 ppm and only 10 ionic liquids
remain liquid at room temperature (Table 4) but all are liquid at 60°C.
Table 4: Phase of tetraalkylammonium
carboxylate ionic liquids at room temperature.
S=Solid, L=Liquid.
[
N222X
]
[N222X
]
[N333X
]
[N333X]
[
OAc]
[OPr]
[OAc]
[OPr]
X=4
X=6
S
L
L
S
S
S
S
S
S
S
S
L
L
L
L
S
L
L
L
L
X=8
X=10
X=12
1
Triethyl butyl ammonium acetate, [N2224][OAc] :RMN- H (400 MHz, D O):  (ppm) = 3.19 (q,
2
J=7.3Hz, 6H), 3.08 (m, 2H), 1.83 (s, 3H),1.57 (m, 2H), 1.30 (hex, J = 7.3Hz, 2H), 1.18 (3xt, J = 7.3
1
3
Hz, 9H), 0.88 (t, J = 7.3Hz, 3H). RMN- C (400 MHz, Dmso-d ):  (ppm) = 172.0, 55.6, 51.8, 26.3,
6
-1
2
2.9, 19.9, 13.5, 7.2. IR  (neat, cm ): 2982, 1584, 1483, 1462, 1368, 1180, 1028, 800.
Triethyl butyl ammonium proprionate, [N2224][OPr] : RMN- H (400 MHz, D O) :  (ppm) = 3.20
1
2
(
(
(
q, J=7.3Hz, 6H), 3.08 (m, 2H), 2.10 (q, J=7.7Hz, 2H),1.58 (m, 2H), 1.31 (hex, J = 7.5Hz, 2H), 1.19
1
3
3xt, J = 7.3 Hz, 9H), 0.98 (t, J = 7.7Hz, 3H), 0.88 (t, J = 7.3Hz). RMN- C (400 MHz, Dmso-d ): 
6
-1
ppm) = 175.0, 55.7, 52.0, 31.9, 22.9, 19.2, 13.7, 11.7, 7.1. IR  (neat, cm ): 2963, 1572, 1392, 1095,
01, 799.
9
1
Triethyl hexyl ammonium acetate, [N2226][OAc] :RMN- H (400 MHz, D O) :  (ppm) = 3.16 (q,
2
J=7.3Hz, 6H), 3.03 (m, 2H), 1.79 (s, 3H),1.55 (m, 2H), 1.23 (m, 6H), 1.15 (3xt, J = 7.3 Hz, 9H), 0.78
1
3
(
t, J = 7.3Hz, 3H). RMN- C (400 MHz, Dmso-d ) :  (ppm) = 172.0, 56.0, 51.9, 30.6, 26.3, 25.5,
6
-1
2
1.9, 20.9, 13.7, 7.2. IR  (neat, cm ) : 2963, 1584, 1370, 1165, 997, 893, 802.
Triethyl hexyl ammonium propionate, [N2226][OPr] : RMN- H (400 MHz, D O) :  (ppm) = 3.18 (q,
1
2
J=7.3Hz, 6H), 3.06 (m, 2H), 2.09 (q, J=7.7Hz, 2H),1.58 (m, 2H), 1.26 (m, 6H), 1.17 (3xt, J = 7.3 Hz,
1
3
9
1
1
H), 0.97 (t, J = 7.7Hz, 3H), 0.79 (t, J = 7.1Hz, 3H). RMN- C (400 MHz, Dmso-d ) :  (ppm) =
6
-1
75.4, 56.0, 51.9, 31.8, 30.6, 25.5, 21.9, 20.9, 13.7, 11.6, 7.2. IR (neat, cm ): 2963, 1578, 1460,
377, 1354, 851, 812.
2
0
This article is protected by copyright. All rights reserved.

ChemSusChem
10.1002/cssc.201601830
1
Triethyl octyl ammonium acetate, [N2228][OAc] : RMN- H (400 MHz, D O) :  (ppm) = 3.19 (q,
2
J=7.3Hz, 6H), 3.06 (m, 2H), 1.83 (s, 3H),1.58 (m, 2H), 1.28 (m, 10H), 1.18 (3xt, J = 7.3 Hz, 9H), 0.80
1
3
(
t, J = 7.1Hz, 3H). RMN- C (400 MHz, Dmso-d ):  (ppm) = 172.0, 56.0, 51.9, 31.1, 28.5, 28.4, 26.3,
6
-1
2
5.9, 22.0, 21.0, 13.9, 7.2. IR  (neat, cm ): 2928, 1584, 1458, 1371, 1161, 1005, 893, 800.
Triethyl octyl ammonium propionate, [N2228][OPr] : RMN- H (400 MHz, D O) :  (ppm) = 3.20 (q,
1
2
J=7.3Hz, 6H), 3.07 (m, 2H), 2.10 (q, J=7.7Hz, 2H),1.59 (m, 2H), 1.29 (m, 10H), 1.19 (3xt, J = 7.3 Hz,
1
3
9
1
1
H), 0.98 (t, J = 7.7Hz, 3H), 0.80 (t, J = 7.1Hz, 3H). RMN- C (400 MHz, Dmso-d ) :  (ppm) =
6
-1
75.1, 56.0, 51.9, 31.8, 31.2, 28.5, 28.4, 25.8, 22.0, 21.0, 13.9, 11.6, 7.2. IR  (neat, cm ): 2926, 1576,
458, 1379, 1354, 1282, 1176
1
Triethyl decyl ammonium acetate, [N22210][OAc] : RMN- H (400 MHz, D O) :  (ppm) = 3.21, 3.05
m, 2H), 1.83 (s, 3H), 1.58 (m, 2H), 1.29 (m, 14H), 1.20 (m, 9H), 0.83 (t, J = 7.1Hz, 3H). RMN- C
2
1
3
(
(
400 MHz, Dmso-d ) :  (ppm) = 172.0, 56.0, 51.9, 31.2, 28.9, 28.8, 28.7, 28.5, 26.1, 25.8, 22.0, 21.0,
6
-1
1
3.8, 7.2. IR  (neat, cm ): 2926, 1583, 1458, 1371, 1161, 1010, 893, 800
1
Triethyl decyl ammonium propionate, [N22210][OPr] : RMN- H (400 MHz, D O):  (ppm) = 3.21 (q,
2
J=7.3Hz, 6H), 3.07 (m, 2H), 2.11 (q, J=7.7Hz, 2H),1.59 (m, 2H), 1.30 (m, 14H), 1.20 (3xt, J = 7.3 Hz,
1
3
9
5
1
H), 0.99 (t, J = 7.7Hz, 3H), 0.82 (t, J = 7.1Hz, 3H). RMN- C (400 MHz, Dmso-d ):  (ppm) = 175.1,
6
-1
6.0, 51.9, 31.8, 31.3, 28.9, 28.8, 28.7, 28.5, 25.9, 22.1, 21.0, 13.8, 11.6 , 7.2. IR  (neat, cm ): 2926,
580, 1458, 1379, 1352, 1280, 1190, 1161, 1013, 854, 808
1
Triethyl dodecyl ammonium acetate, [N22212][OAc] : RMN- H (400 MHz, D O) :  (ppm) = 3.17 (q,
2
J=7.3Hz, 6H), 3.02 (m, 2H), 1.80 (s, 3H),1.54 (m, 2H), 1.25 (m, 18H), 1.16 (m, 9H), 0.78 (t, J =
1
3
7
2
9
.1Hz, 3H). RMN- C (400 MHz, Dmso-d ) :  (ppm) = 172.0, 56.0, 51.9, 31.3, 29.1, 29.0, 28.9, 28.8,
6
-1
8.7, 28.6, 26.2, 25.8, 22.1, 21.0, 13.9, 7.2. IR  (neat, cm ): 2918, 1585, 1470, 1381, 1030, 1008,
00, 795.
1
Triethyl dodecyl ammonium propionate, [N22212][OPr] : RMN- H (400 MHz, D O) :  (ppm) = 3.17
2
(
q, J=7.3Hz, 6H), 3.01 (m, 2H), 2.05 (q, J=7.7Hz, 2H),1.54 (m, 2H), 1.25 (m, 18H), 1.16 (m, J = 7.3
1
3
Hz, 9H), 0.94 (t, J = 7.7Hz, 3H), 0.79 (t, J = 7.1Hz, 3H). RMN- C (400 MHz, Dmso-d ) :  (ppm) =
6
1
75.1, 56.0, 51.9, 31.8, 31.3, 29.1, 29.0, 28.9, 28.8, 28.7, 28.5, 25.9, 22.1, 21.0, 13.9, 11.6, 7.2. IR 
-1
(
neat, cm ): 2926, 1580, 1458, 1379, 1352, 1280, 850, 806.
1
Tripropyl butyl ammonium acetate, [N3334][OAc] : RMN- H (400 MHz, D O) :  (ppm) = 2.77 (m,
2
1
3
8H), 1.52 (s, 3H), 1.30 (m, 8H), 0.97 (hex, J = 7.3Hz, 2H), 0.56 (t, J = 7.3Hz, 12H). RMN- C (400
-1
MHz, Dmso-d ):  (ppm) = 172.0, 59.2, 57.6, 26.3, 23.1, 19.2, 14.8, 13.4, 10.5. IR  (neat, cm ):
6
2
967, 2941, 2915, 2876, 1586, 1474, 1368, 1316, 991, 891, 760
1
Tripropyl butyl ammonium propionate, [N3334][OPr] : RMN- H (400 MHz, D O) :  (ppm) = 3.03
2
(
m, 8H), 2.04 (q, J=7.7Hz, 2H), 1.56 (m, 8H), 1.23 (hex, J = 7.3Hz, 2H), 0.92 (t, J = 7.7Hz, 3H), 0.81
13
(
t, J = 7.3Hz, 12H). RMN- C (400 MHz, Dmso-d ) :  (ppm) = 175.2, 59.2, 57.7, 31.9, 23.1, 19.2,
6
-1
1
7
4.9, 13.4, 11.6, 10.5. IR  (neat, cm ): 2967, 2936, 2878, 1582, 1460, 1377, 1352, 1283, 964, 853,
58.
1
Tripropyl hexyl ammonium acetate, [N3336][OAc] : RMN- H (400 MHz, D O) :  (ppm) = 3.02 (m,
2
8
H), 1.76 (s, 3H), 1.54 (m, 8H), 1.19 (m, 6H), 0.81 (t, J = 7.3Hz, 9H), 0.74 (t, J = 7.1Hz, 3H). RMN-
1
3
C (400 MHz, Dmso-d ):  (ppm) = 172.1, 59.2, 57.8, 30.7, 26.2, 25.5, 21.9, 21.1, 14.8, 13.8, 10.5. IR
6
-1

(neat, cm ): 2967, 2934, 2876, 2859, 1586, 1466, 1370, 1317, 962, 891, 760
1
Tripropyl hexyl ammonium propionate, [N3336][OPr] : RMN- H (400 MHz, D O) :  (ppm) = 3.02
2
(
m, 8H), 2.03 (q, J=7.7Hz, 2H), 1.54 (m, 8H), 1.19 (m, 6H), 0.91 (t, J = 7.7Hz, 3H), 0.81 (t, J = 7.3Hz,
13
9
2
1
H), 0.74 (t, J = 7.1Hz, 3H). RMN- C (400 MHz, Dmso-d ):  (ppm) = 175.1, 59.2, 57.8, 31.8, 30.7,
6
-1
5.5, 21.9, 21.1, 14.8, 13.7, 11.6, 10.5. IR  (neat, cm ): 2963, 2928, 2876, 1582, 1460, 1377, 1352,
281, 963, 853, 762.
2
1
This article is protected by copyright. All rights reserved.

ChemSusChem
10.1002/cssc.201601830
1
Tripropyl octyl ammonium acetate, [N3338][OAc] : RMN- H (400 MHz, D O) :  (ppm) = 3.03 (m,
2
8
H), 1.78 (s, 3H), 1.56 (m, 8H), 1.20 (m, 10H), 0.83 (t, J = 7.3Hz, 9H), 0.75 (t, J = 7.0Hz, 3H). RMN-
1
3
C (400 MHz, Dmso-d ) :  (ppm) = 172.0, 59.2, 57.8, 31.1, 28.5, 28.4, 26.2, 25.9, 22.0, 21.1, 14.8,
6
-1
13.9, 10.5. IR  (neat, cm ): 2963, 2926, 2876, 2855, 1586, 1462, 1369, 1319, 990, 893, 760
Tripropyl octyl ammonium propionate, [N3338][OPr] :
1
RMN- H (400 MHz, D O) :  (ppm) = 3.03 (m, 8H), 2.03 (q, J=7.7Hz, 2H), 1.55 (m, 8H), 1.19 (m,
1
2
1
3
0H), 0.93 (t, J = 7.7Hz, 3H), 0.83 (t, J = 7.3Hz, 9H), 0.77 (t, J = 7.1Hz, 3H). RMN- C (400 MHz,
Dmso-d ):  (ppm) = 176.0, 60.1, 58.7, 32.7, 32.1, 29.4, 29.3, 26.8, 22.9, 22.0, 15.8, 14.8, 12.5, 11.4 .
6
-1
IR  (neat, cm ): 2963, 2926, 2876, 2859, 1582, 1460, 1377, 1350, 1281, 970, 853, 760
1
Tripropyl decyl ammonium acetate, [N33310][OAc] : RMN- H (400 MHz, D O):  (ppm) = 3.05 (m,
2
8
H), 1.78 (s, 3H), 1.58 (m, 8H), 1.21 (m, 14H), 0.85 (t, J = 7.3Hz, 9H), 0.79 (t, J = 7.0Hz, 3H). RMN-
1
3
C (400 MHz, Dmso-d ):  (ppm) = 172.0, 59.2, 57.8, 31.3, 28.9, 28.8, 28.7, 28.5, 26.2, 25.8, 22.1,
6
-1
2
8
1.1, 14.8, 13.9 (CH ), 10.5. IR  (neat, cm ): 2965, 2928, 2878, 2857, 1586, 1462, 1370, 1317, 962,
93, 760
3
1
Tripropyl decyl ammonium propionate, [N33310][OPr] : RMN- H (400 MHz, D O):  (ppm) = 3.05
2
(
m, 8H), 2.04 (q, J=7.7Hz, 2H), 1.57 (m, 8H), 1.21 (m, 14H), 0.93 (t, J = 7.7Hz, 3H), 0.85 (t, J =
1
3
7
3
2
.3Hz, 9H), 0.78 (t, J = 7.1Hz, 3H). RMN- C (400 MHz, Dmso-d ):  (ppm) = 175.2, 59.2, 57.8,
6
-1
1.7, 31.3, 28.9, 28.8, 28.7, 28.5, 25.9, 22.1, 21.1, 14.9, 13.8, 11.6, 10.5. IR  (neat, cm ): 2963, 2926,
876, 2855, 1582, 1460, 1377, 1352, 1283, 964, 853, 756.
1
Tripropyl dodecyl ammonium acetate, [N33312][OAc] : RMN- H (400 MHz, D O):  (ppm) = 3.06
2
(
m, 8H), 1.79 (s, 3H), 1.58 (m, 8H), 1.23 (m, 18H), 0.86 (t, J = 7.3Hz, 9H), 0.79 (t, J = 7.0Hz, 3H).
1
3
RMN- C (400 MHz, Dmso-d ):  (ppm) = 172.0, 59.2, 57.8, 31.3, 29.1, 29.0, 28.9, 28.8, 28.7, 28.5,
2
1
6
-
1
6.1, 25.9, 22.1, 21.1, 14.8, 13.8, 10.5. IR  (pur, cm ): 2965, 2928, 2878, 2857, 1586, 1466, 1371,
317, 966, 893, 760
1
Tripropyl dodecyl ammonium propionate, [N33312][OPr] : RMN- H (400 MHz, D O):  (ppm) =
2
3
7
3
2
.06 (m, 8H), 2.04 (q, J=7.7Hz, 2H), 1.57 (m, 8H), 1.22 (m, 18H), 0.93 (t, J = 7.7Hz, 3H), 0.86 (t, J =
1
3
.3Hz, 9H), 0.79 (t, J = 7.1Hz, 3H).RMN- C (400 MHz, Dmso-d ):  (ppm) = 175.1, 59.2, 57.8, 31.7,
6
-1
1.3, 29.1, 29.0, 28.9, 28.8, 28.7, 28.5, 25.9, 22.1, 21.1, 14.8, 13.8, 11.6, 10.5.IR  (neat, cm ) : 2959,
924, 2874, 2855, 1582, 1460, 1379, 1352, 1283, 966, 855, 762.
Spectroscopy: For organic compounds characterization after synthesis, IR spectra were
recorded on a Shimadzu Fourier Transform Infrared Spectrophotometer FTIR-8400S,
equipped with a PIKE MIRacle Attenuated Total Reflectance (ATR) accessory (ATR crystal
plate: germanium). For the molecular interactions study, IR spectra were recorded with a
Nicolet 380 FT-IR spectrometer (DTGS detector) equipped with ATR single reflection
-1
diamond from Specac. IR spectra were averaged on 32 scans and with a resolution of 4 cm .
1
NMR spectra were recorded in Fourier Transform mode with a Bruker AVANCE 400 ( H at
13
4
00 MHz, C at 100 MHz), at 298K. Data are reported as chemical shifts (δ) in ppm.
1
13
Residual solvent signals were used as internal references ( H, C).
2
2
This article is protected by copyright. All rights reserved.

ChemSusChem
10.1002/cssc.201601830
Microscopy: In situ dissolution was followed by observing the disappearance of cellulose
particles (partially crystalline) with time using a Leica DM2500M optical microscope. During
the dissolution the sample is placed in between a glass slide and a glass cover in the hot stage
(
Linkam LTS420) in order to control temperature and atmosphere (under N flow) of the
2
[
16]
sample similar to our previous work. Since the sample is not stirred during the experiment
in the microscope we paid a particular attention to start the dissolution experiment with a
homogeneous distribution of the cellulose particles. Cross-polarized light was used to
improve the illumination of the sample by increasing the contrast between the solvent (black)
and the cellulose particles (white). Images of 2048  1536 pixels are taken regularly during
2
the dissolution experiments and represent a surface of around 4 mm of the sample and circa
th
1
/100 of the total sample. A typical sample is around 40 mg which corresponds to a
thickness of 100 μm. The percentages of cellulose given in this work are calculated as the
ratio of (mass of cellulose) by (mass cellulose + mass ionic liquid + mass DMSO). As
illustration of a typical experiment, a sample containing 10 wt% cellulose in [N2226][OAc]
was heating at a rate of 0.25 °C/min from 30 to 80 °C. An image is taken every 2 minutes (or
0
.5 °C) and few images of this experiment are given in Figure 7. All the cellulose is dissolved
before the sample reaches 80 °C and most of the particles are disappearing between 50 and
0 °C. The full dissolution of cellulose was confirmed by X-ray diffraction and the absence of
6
chemical reaction between cellulose and ionic liquid was verified by IR.
2
3
This article is protected by copyright. All rights reserved.

ChemSusChem
10.1002/cssc.201601830
Figure 7: Images of an in situ dissolution of 10 wt% of microcrystalline cellulose in
N2226][OAc] from 30 °C to 80 °C with a heating rate of 0.25 °C/min. The images’ size is
[
around 2.3 mm x 1.7 mm.
The brightness of each image taken during a dissolution experiment is measured and its
evolution is used to identify the temperature/time corresponding to the total dissolution of the
cellulose and to compare different experiments. As example, the evolution of brightness
during 3 experiments with 10 wt% of cellulose in [N2226][OAc] is given in Figure 8. The
brightness of the sample is systematically normalized to 100 for each experiment which does
not necessary corresponds to the first image of the experiment. When the cellulose is fully
dissolved the brightness of the sample is constant with a value close to 0. The brightness of
the sample increases during the first few minutes. This corresponds mainly to a swelling of
[
16, 33]
the cellulose particles already referred in literature.
Because of this, it is not possible to
simply correlate the brightness of an image with the quantity of undissolved cellulose.
However we can estimate that the value of the brightness is an overestimation of the quantity
of undissolved cellulose. For example, with the heating rate of 2 °C/min, at 80 °C, a
brightness of 20 is measured meaning that at least 80% of the initial cellulose has been
dissolved.
We also observe that the temperature of dissolution is depending on the heating rate. At
2
°C/min, the cellulose is not completely dissolved before 90 °C, at 0.25 °C/min it is fully
dissolved at 70 °C and at 0.01 °C/min the temperature is reduced to 52 °C (see Figure 8). At
the slowest heating rate, the entire dissolution takes 1.5 days while at 0.25 °C/min it takes less
than 3 hours. A heating ramp of 0.25 °C/min was used with temperatures increasing from 30
to 90 °C in order to realize an experiment in 4 hours (except in the case of ionic liquids solids
at ambient temperature for which the dissolution experiments start at higher temperatures).
The dissolution temperature was evaluated with a first set of experiments with a fixed
quantity of cellulose (10 wt%). Experiments with different quantities of cellulose were also
conducted in order to estimate the maximum quantity that can be dissolved in the ionic
liquids.
2
4
This article is protected by copyright. All rights reserved.

ChemSusChem
10.1002/cssc.201601830
Figure 8: Evolution of the brightness during the dissolution of 10 wt% of microcrystalline
cellulose in [N2226][OAc] as observed in an optical microscope with 0.01, 0.25 and
2
.00 °C/min heating ramps. Lines are connecting every experimental point (every 0.5 °C).
Symbols are only here to help the reader.
Characterization of regenerated cellulose: In few cases, cellulose powder was precipitated
from ionic liquids by addition of water and then washed 3 times. The regenerated powders are
as white as the untreated cellulose. They were used to confirm the dissolution of the cellulose
(strong decrease of the crystallinity of the cellulose as observed by X-ray diffractograms
obtained with a Philips X-Pert Pro using Cu-Kα radiation). The absence of chemical reaction
of the native cellulose with the ionic liquid during the dissolution was also established
(absence of C=O bands in the infrared spectra).
Viscosity: The viscosity of the solvents was measured using an Anton Paar Lovis 2000ME
microviscometer based on the falling ball principle. A capillary tube with an internal diameter
[
26]
of 1.8 mm was considered optimal for viscosity up to 300 mPa.s. The density of the fluid is
necessary for the calculation of the viscosity from falling ball experiments. The density of
-3
[
N2226][OAc] (0.96063 g.cm at 25 °C, with a linear evolution from 25 to 80 °C : density
T /°C) = 0.97436-0.0005527xT(/°C)) was measured using a vibrating tube densimeter DMA
(
5
000M from Anton Paar. Densities for the other solvents were calculated from the density of
3
[34]
[
N2226][OAc] and the molar volume of CH group of alkyl chain of 16.92 cm /mol.
2
Viscosities were measured on ionic liquids with around 0.5 wt% of water. It was considered
that the measured viscosity is undervaluing the viscosity of pure ionic liquid by at least
[
26]
2
0%.
2
5
This article is protected by copyright. All rights reserved.

ChemSusChem
10.1002/cssc.201601830
Rheology: 10 wt% of cellulose has been dissolved in 0.5 g of ionic liquid within few hours
at 80 °C. Rheology measurements have been performed within 24 hours after the dissolution
of the cellulose in the ionic liquids in order to avoid a strong decrease of the polymer
[
35]
length.
Melt viscoelastic experiments have been carried out with an ARES mechanical
spectrometer TA Instruments in oscillatory frequency mode using parallel plates (25 mm
diameter) under air atmosphere. In all cases strain amplitude has been checked to be sure to
remain within the linear viscoelastic range. Temperatures of experiments ranged from 90 °C
till to 30 °C. A sinusoidal strain is applied to the sample and a resulting sinusoidal stress
signal is recorded. Its deconvolution between the in phase (=0) and the out of phase (=/2)
components allow the determination of the loss and storage viscosities.
Molecular Simulation: Molecular structures of four solvents ([N2224][OAc] and
[
N2226][OAc], with and without addition of DMSO) around one cellohexaose (6 glucoses
units) has been calculated at 127 °C. Each simulation box was built with 1500 ions pairs, 1
cellohexaose and 0 or 637 DMSO molecules (corresponding to 20 wt% of DMSO).
Simulations were performed with a time step of 2 fs in an NPT ensemble and with a cutoff of
[
36]
1
2 Å using LAMMPS molecular dynamics code. All molecules were simulated using all-
[
37]
[38]
atom flexible models derived from the OPLS-AA (DMSO,
cellohexaose
and ionic
[
39]
liquids ). For each solvent, 10 initial configurations were created every 200 ps in a first
simulation at 427 °C. With each of these configurations, 8 cycles of (1) 400 ps at 427 °C, (2)
4
00 ps of equilibration at 127 °C and finally (3) 1 ns of acquisition were computed. The
structure and interactions between a cellohexaose molecule and each solvent were then
studied with a total trajectory of 80 ns using 80 different initial configurations in order to
probe all configurations of the oligomer. The number of hydrogen bonds between anions and
cellohexaose were calculated using the usual geometrical definition (distance O…H < 2.5Å
and OOH angle < 30°).
Acknowledgements
We thank GDR 3585 (LIPS: Ionic Liquids and Polymers) for its financial support. The
authors acknowledge Prof. Agilio Padua for his contribution in the molecular simulations.
Keyword: cellulose, ionic liquid, dissolution, polarizing optical microscopy, rheology
2
6
This article is protected by copyright. All rights reserved.

ChemSusChem
10.1002/cssc.201601830
References
[
1] A. J. Ragauskas, C. K. Williams, B. H. Davison, G. Britovsek, J. Cairney, C. A. Eckert,
W. J. Frederick, J. P. Hallett, D. J. Leak, C. L. Liotta, J. R. Mielenz, R. Murphy, R.
Templer, T. Tschaplinski, Science 2006, 311, 484-489.
[
[
2] A. Stark, Energy & Environmental Science 2011, 4, 19-32.
3] a) D. Klemm, B. Heublein, H.-P. Fink, A. Bohn, Angew. Chem., Int. Ed. Engl. 2005,
4
4, 3358-3393; b) T. Liebert, in Cellulose Solvents: For Analysis, Shaping and
Chemical Modification, Vol. 1033, American Chemical Society, 2010, pp. 3-54; c) G.
Mancini, S. Papirio, P. N. L. Lens, G. Esposito, Energy Fuels 2016, 30, 1892-1903.
4] R. P. Swatloski, S. K. Spear, J. D. Holbrey, R. D. Rogers, J. Am. Chem. Soc. 2002, 124,
[
4
974-4975.
[
[
[
[
5] K. C. Badgujar, B. M. Bhanage, Bioresour. Technol. 2015, 178, 2-18.
6] A. Brandt, J. Grasvik, J. P. Hallett, T. Welton, Green Chem. 2013, 15, 550-583.
7] M. Kick, P. Keil, A. König, Fluid Phase Equilib. 2013, 338, 172-178.
8] a) M. T. Clough, K. Geyer, P. A. Hunt, S. Son, U. Vagt, T. Welton, Green Chem. 2015,
1
7, 231-243; b) M. T. Clough, K. Geyer, P. A. Hunt, J. Mertes, T. Welton, Phys. Chem.
Chem. Phys. 2013, 15, 20480-20495.
[
[
[
9] Y.-R. Liu, K. Thomsen, Y. Nie, S.-J. Zhang, A. S. Meyer, Green Chem. 2016, 6246-
6
254
10] a) B. D. Rabideau, A. Agarwal, A. E. Ismail, J. Phys. Chem. B 2014, 118, 1621-1629;
b) B. Mostofian, J. C. Smith, X. Cheng, Cellulose 2014, 21, 983-997.
11] a) B. Lindman, G. Karlström, L. Stigsson, J. Mol. Liq. 2010, 156, 76-81; b) B.
Medronho, A. Romano, M. G. Miguel, L. Stigsson, B. Lindman, Cellulose 2012, 19,
5
81-587.
[
[
12] R. Rinaldi, Chem. Commun. (Cambridge, U. K.) 2011, 47, 511-513.
13] a) J. M. Andanson, A. A. H. Padua, M. F. Costa Gomes, Chem. Commun. (Cambridge,
U. K.) 2015, 51, 4485-4487; b) H. F. N. de Oliveira, R. Rinaldi, ChemSusChem 2015,
8
, 1577-1584.
[
14] a) S. Velioglu, X. Yao, J. Devémy, M. G. Ahunbay, S. B. Tantekin-Ersolmaz, A.
Dequidt, M. F. Costa Gomes, A. A. H. Pádua, J. Phys. Chem. B 2014, 118, 14860-
1
4869; b) F. Huo, Z. Liu, W. Wang, J. Phys. Chem. B 2013, 117, 11780-11792.
15] M. Gericke, K. Schlufter, T. Liebert, T. Heinze, T. Budtova, Biomacromolecules 2009,
0, 1188-1194.
16] J.-M. Andanson, E. Bordes, J. Devemy, F. Leroux, A. A. H. Padua, M. F. C. Gomes,
Green Chem. 2014, 16, 2528-2538.
17] a) E. Gale, R. H. Wirawan, R. L. Silveira, C. S. Pereira, M. A. Johns, M. S. Skaf, J. L.
Scott, ACS Sustainable Chemistry & Engineering 2016; b) F. Boissou, A. Muhlbauer,
K. De Oliveira Vigier, L. Leclercq, W. Kunz, S. Marinkovic, B. Estrine, V. Nardello-
Rataj, F. Jerome, Green Chem. 2014, 16, 2463-2471.
[
[
[
1
[
[
18] H. F. N. de Oliveira, M. T. Clough, R. Rinaldi, ChemSusChem 2016, 9, 3324-3329.
19] a) L. Chen, M. Sharifzadeh, N. Mac Dowell, T. Welton, N. Shah, J. P. Hallett, Green
Chem. 2014, 16, 3098-3106; b) A. George, A. Brandt, K. Tran, S. M. S. N. S. Zahari,
D. Klein-Marcuschamer, N. Sun, N. Sathitsuksanoh, J. Shi, V. Stavila, R. Parthasarathi,
S. Singh, B. M. Holmes, T. Welton, B. A. Simmons, J. P. Hallett, Green Chem. 2015,
1
7, 1728-1734.
20] M. Kostag, T. Liebert, O. A. El Seoud, T. Heinze, Macromol. Rapid Commun. 2013,
4, 1580-1584.
[
3
[
[
21] B. Zhao, L. Greiner, W. Leitner, RSC Advances 2012, 2, 2476-2479.
22] Y.-B. Huang, P.-P. Xin, J.-X. Li, Y.-Y. Shao, C.-B. Huang, H. Pan, ACS Sustainable
Chem. Eng. 2016, 4, 2286-2294.
2
7
This article is protected by copyright. All rights reserved.

ChemSusChem
10.1002/cssc.201601830
[
23] K. Ohira, Y. Abe, M. Kawatsura, K. Suzuki, M. Mizuno, Y. Amano, T. Itoh,
ChemSusChem 2012, 5, 388-391.
[
[
24] Z. Chen, S. Liu, Z. Li, Q. Zhang, Y. Deng, New J. Chem. 2011, 35, 1596-1606.
25] A. Xu, L. Cao, B. Wang, J. Ma, Advances in Materials Science and Engineering 2015,
2
015, 4.
26] J. M. Andanson, X. Meng, M. Traïkia, P. Husson, J. Chem. Thermodyn. 2016, 94, 169-
76.
27] M. C. Castro, H. Rodríguez, A. Arce, A. Soto, Ind. Eng. Chem. Res. 2014, 53, 11850-
1861.
[
[
1
1
[
[
28] M. Kanakubo, T. Makino, T. Umecky, J. Mol. Liq. 2016, 217, 112-119.
29] a) M. H. Ghatee, M. Bahrami, N. Khanjari, J. Chem. Thermodyn. 2013, 65, 42-52; b) N.
Muhammad, M. I. Hossain, Z. Man, M. El-Harbawi, M. A. Bustam, Y. A. Noaman, N.
B. Mohamed Alitheen, M. K. Ng, G. Hefter, C.-Y. Yin, J. Chem. Eng. Data 2012, 57,
2
191-2196.
[
[
30] H. Song, Y. Niu, Z. Wang, J. Zhang, Biomacromolecules 2011, 12, 1087-1096.
31] B. A. Marekha, M. Bria, M. Moreau, I. De Waele, F.-A. Miannay, Y. Smortsova, T.
Takamuku, O. N. Kalugin, M. Kiselev, A. Idrissi, J. Mol. Liq. 2015, 210, Part B, 227-
2
37.
[
[
32] H. M. Cho, A. S. Gross, J.-W. Chu, J. Am. Chem. Soc. 2011, 133, 14033-14041.
33] L. C. Fidale, N. Ruiz, T. Heinze, O. A. E. Seoud, Macromol. Chem. Phys. 2008, 209,
1
240-1254.
[
[
34] K. Machanová, A. Boisset, Z. Sedláková, M. Anouti, M. Bendová, J. Jacquemin, J.
Chem. Eng. Data 2012, 57, 2227-2235.
35] a) O. M. Gazit, A. Katz, ChemSusChem 2012, 5, 1542-1548; b) Y. Ahn, S.-Y. Kwak,
Y. Song, H. Kim, Phys. Chem. Chem. Phys. 2016, 18, 1460-1469.
36] S. Plimpton, J. Comput. Phys. 1995, 117, 1-19.
[
[
37] W. L. Jorgensen, D. S. Maxwell, J. Tirado-Rives, J. Am. Chem. Soc. 1996, 118, 11225-
1
1236.
38] G. A. Kaminski, R. A. Friesner, J. Tirado-Rives, W. L. Jorgensen, J. Phys. Chem. B
001, 105, 6474-6487.
39] J. N. Canongia Lopes, J. Deschamps, A. A. H. Pádua, J. Phys. Chem. B 2004, 108,
038-2047.
[
[
2
2
2
8
This article is protected by copyright. All rights reserved.