Flame-retardant thermoplastics derived from plant cell wall polymers by single ionic liquid substitution

Article abstract of DOI:10.1039/c8nj04797a

Three components of plant cell walls - cellulose, hemicellulose and lignin - were converted into flame-retardant thermoplastics by adducting only a single ionic liquid species via covalent bonds. They showed thermoplasticity and formed thin films by hot pressing. They also showed flame retardancy and self-extinguished the fire during burning. The properties of the samples depend on the cation species of ionic liquids adducted and thus are controllable. In the present study, more than 66% of the hydroxyl groups present on the polymers were maintained after derivatisation; they thus have the potential for further functionalisation for moulding, practical use and so on, in addition to flame retardancy and thermoplasticity.

Full text of DOI:10.1039/c8nj04797a

NJC
View Article Online
View Journal
PAPER
Flame-retardant thermoplastics derived from
plant cell wall polymers by single ionic liquid
Cite this:DOI: 10.1039/c8nj04797a
substitution†
Ryunosuke Nishita,a Kosuke Kuroda, *a Shohei Ota,a Takatsugu Endo,b
a
Shiori Suzuki,a Kazuaki Ninomiyac and Kenji Takahashi
Three components of plant cell walls—cellulose, hemicellulose and lignin—were converted into flame-
retardant thermoplastics by adducting only a single ionic liquid species via covalent bonds. They showed
thermoplasticity and formed thin films by hot pressing. They also showed flame retardancy and self-
extinguished the fire during burning. The properties of the samples depend on the cation species of ionic
liquids adducted and thus are controllable. In the present study, more than 66% of the hydroxyl groups
present on the polymers were maintained after derivatisation; they thus have the potential for further
functionalisation for moulding, practical use and so on, in addition to flame retardancy and thermoplasticity.
Received 21st September 2018,
Accepted 20th December 2018
DOI: 10.1039/c8nj04797a
rsc.li/njc
flame retardants also play a role as plasticisers.10 However, such
flame retardants easily bleed out because they are only mixed
Introduction
Cellulose, hemicellulose, and lignin are produced as the main with the polymers.
components of cell walls and have accumulated in large quantities
Introduction of flame retardants via covalent bonds is one of
on earth. They have attracted attention as candidates to produce the methods to prevent bleeding out of the flame retardants.
new materials for the replacement of petroleum-based materials. For example, phosphoric acid-type flame retardants can be
Research studies regarding cellulose especially have progressed in covalently bonded to the hydroxyl groups of carbohydrates
the past few decades and cellulose has been converted into and have been reported to improve the flame retardancy of
thermoplastics through acylation.1–4 On the other hand, flame wood powders.8 In addition, Aoki et al. produced a flame-
retardancy is well known as an important property of thermo- retardant thermoplastic from cellulose in two steps: adduction
plastics, but cellulosic materials are known to be flammable of a phosphoric acid-type flame retardant to the residual
materials. Therefore, there is a strong demand for flame- hydroxyl groups of a cellulose-derived thermoplastic, cellulose
retardant thermoplastics derived from cellulose.
propionate (CP).11 The cellulose plastic synthesised via this
Addition of flame retardancy to plant-derived polymers has method, however, cannot be further functionalised because
been reported.5,6 For example, boric acid flame retardants are almost all of the hydroxyl groups have been substituted by
added to cellulose plastics.7 The flame retardants form foam the flame-retardant group (phosphoric acid) and plasticiser-
layers on the surfaces of the materials to prevent heat and air group (acyl group) and thus there are no more reactive groups.
from passing to the interior part of the materials when the Therefore, further control of other properties of the cellulose
plastics are heated.5 The phosphoric acid-type flame retardants plastic is important for moulding and practical use (e.g. physico-
also form char layers composed of polyphosphate on the surfaces chemical properties such as solubility and viscosity, and
of the materials, in addition to radical trap by PO-radicals in the mechanical properties such as tensile strength, bending strength
gas phase.8,9 It is noted that some of the phosphoric acid-type and elastic modulus). For this reason, the current cellulose-
derived flame-retardant thermoplastics have limited application,
and they are not recognised as a suitable alternative to
petroleum-derived plastics. If a certain number of hydroxyl
groups could be maintained on cellulose after the addition of
thermoplasticity and flame retardancy, the cellulose-derived
flame-retardant thermoplastic can be further functionalised by
substitution with other kinds of functional groups and thus may
have the potential to match or surpass petroleum-derived plastics.
We here propose the introduction of a single substituent that can
a Faculty of Biological Science and Technology, Institute of Science and Engineering,
Kanazawa University, Kakuma-Machi, Kanazawa 920-1192, Japan.
E-mail: kkuroda@staff.kanazawa-u.ac.jp
b Faculty of Science and Engineering, Doshisha University, 1-3 Tatara Miyakodani,
Kyotanabe-Shi, Kyoto 610-0394, Japan
c Institute for Frontier Science Initiative, Kanazawa University, Kakuma-Machi,
Kanazawa 920-1192, Japan
† Electronic supplementary information (ESI) available. See DOI: 10.1039/
c8nj04797a
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2019
New J. Chem.

View Article Online
Paper
NJC
impart both thermoplasticity and flame retardancy to cellulose,
to maintain the hydroxyl groups after derivatisation.
The present method can be applied to other polymers derived
from plant cell walls, namely, hemicellulose and lignin, because
We focused on ionic liquids (ILs) as substituents, to impart they also have hydroxyl groups. In this study, all three kinds of
thermoplasticity and flame retardancy to cellulose simulta- ‘‘flame-retardant cell-wall-derived thermoplastics’’ were devel-
neously. Ionic liquids are salts that are liquid below 100 1C. oped, while maintaining a certain number of hydroxyl groups.
Ionic liquids have been reported to function as plasticisers or
flame retardants when mixed with polymers.12 For example,
cellulose does not show a softening point because cellulose
chains are strongly bound by hydrogen bonding. Ionic liquids
act as plasticisers when mixed with cellulose, through disrupting
hydrogen bonds and increasing the free volume to make free
Results and discussion
Flame-retardant cell-wall-derived thermoplastics with an
imidazolium-based ionic liquid
space for the cellulose molecules to move.13 Different ionic The residual weight of [C2mim][(MeO)(H)PO2] at 400 1C upon
liquid species are known to impart flame retardancy by mixing thermogravimetric analysis (TGA) under an air atmosphere was
with plastics or impregnating wood.14–17 Based on these reports, preliminarily checked as an index of flame retardancy8 (Fig. S1,
suitably designed ionic liquids are expected to function as both ESI†). The residual weight was 41%, while [C2mim]Br (the same
plasticisers and flame retardants. Furthermore, ionic liquids do cation and bromide anion) showed only 2%. The residual weight
not bleed out if they adduct to cellulose as substituents.
is attributed to char formation of the [(MeO)(H)PO2] anion. This
In this study, we selected 1-ethyl-3-methylimidazolium methyl- result shows that the flame retardancy of [C2mim][(MeO)(H)PO2]
phosphonate ([C2mim][(MeO)(H)PO2], see Fig. 1), as it can is mainly based on the [(MeO)(H)PO2] anion (although actually
dissolve cellulose18–25 and the [(MeO)(H)PO2] anion can be the [C2mim] cation was also found to involve flame retardancy in
bound by the hydroxyl group of cellulose by an esterification a burning test (details later)).
reaction at high temperatures.26 Substitution of the hydroxyl
Cellulose, xylan and lignin were respectively added to [C2mim]-
group was further expected to show another effect in addition [(MeO)(H)PO2] and were substituted with [C2mim][(MeO)(H)PO2]
to the influence of just mixing ILs; the decrease in hydroxyl ([C2mim]+cellulose, [C2mim]+xylan with [C2mim]+lignin; struc-
groups leads to further disruption of the hydrogen bond network tures are shown in Fig. 1). Lignin required a longer reaction
of cellulose, resulting in a decrease in the softening point.27 time (3 h), because the reaction hardly proceeded within 1 h.
From a flame retardancy perspective, [C2mim][(MeO)(H)PO2] was
In the 1H NMR spectrum of [C2mim]+cellulose, peaks derived
expected to show not only the flame-retarding effect of the ionic from [C2mim] cations were observed at 1.37, 3.84, 4.19, 7.75, 7.83
liquid itself but also a flame-retardant effect as a phosphoric and 9.43 ppm, while peaks derived from [(MeO)(H)PO2] anions
acid-type compound, as well as other phosphorus-containing were observed at 6.09 and 7.05 ppm (Fig. S2, ESI†). Furthermore,
ILs.14,15 It has also been reported that if phosphoric acid-type the signal attributed to the methyl group of the [(‘‘Me’’O)(H)PO2]
flame retardants are covalently introduced at the 6 position of anion (around 3.2 ppm) was not observed in the spectrum
cellulose, the process of char formation is further improved.5,11 of [C2mim]+cellulose. These results strongly suggest that
This method has two more advantages: single step reaction the [(MeO)(H)PO2] anion was covalently introduced into the
(the previous flame-retardant thermoplastic9 was formed in two cellulose, accompanied by removal of the methyl group of the
steps) and control of the degree of substitution. The latter [(‘‘Me’’O)(H)PO2] anion. [C2mim]+xylan and [C2mim]+lignin were
enables the precise control of the resulting properties of the confirmed to have been synthesised along with [C2mim]+cellulose
plastics, because it has been reported that some of the propionyl (see Fig. S2, ESI†).
groups already introduced are eliminated when introducing a
Introduction of [C2mim][(MeO)(H)PO2] into each of the
phosphoric acid-type flame retardant into the CP in the case of three components of plant cell walls was also confirmed by
two-step production.11
FT-IR spectroscopy (Fig. S3, ESI†). The peaks derived from the
P–O–C bond at 821 cmꢀ1, PQO bond at 1211 cmꢀ1 and CQN
bond at 1571 cmꢀ1 were detected in [C2mim]+cellulose but not in
the unreacted polymer.28,29 The peaks derived from these bonds
were also detected in [C2mim]+xylan and [C2mim]+lignin (see
Fig. S3, ESI†). These results indicated that [C2mim][(MeO)(H)PO2]
was introduced to each component of plant cell walls.
The substitution ratios of hydroxyl groups of cellulose, xylan
and lignin with [C2mim][(MeO)(H)PO2] were detected by 31P NMR
spectroscopy. Their substitution ratio was 33, 32 and 14%,
respectively, against the number of hydroxyl groups. Namely, they
maintained 67, 68 and 86% of the hydroxyl groups, respectively,
after derivatisation.
Characteristics of the samples are shown in Table 1. All three
samples were able to form thin films when subjected to hot
pressing (Fig. 2). The films of [C2mim]+cellulose and [C2mim]+xylan
Fig. 1 The reaction scheme for substitution of cellulose, xylan, and lignin
by [C2mim][(MeO)(H)PO2], and the structures of the ionic liquid-substituted
cellulose, xylan and lignin.
New J. Chem.
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2019

View Article Online
NJC
Paper
Table 1 Characteristics of plant-based polymers before and after derivatisation with the ionic liquids
Thermal
Softening decomposition
Fire
suppression
Conversion Film
Film
Miscibility
Residual weight Char
Cation
—
ratio
formation character in water
point (1C) temperature (5%) (at 400 1C)
formation ability
Cellulose
propionate
89%
|
Soft
Not miscible
160
301 1C
12%
Slightly
formed
Completely
burned
d
b
b
Cellulose
—
—
Not formed —b
Soft
Not formed —b
Not miscible
Miscible
Not miscible
Not miscible
Not misciblec
—
313 1C
230 1C
234 1C
248 1C
243 1C
11%
52%
57%
62%
62%
—
—
[C2mim]+ 33%
|
160
140
140
120
| (foam) |
+
b
b
[P6,6,6,6
[P8,8,8,8
[P6,6,6,14
]
]
(33%)a
(33%)a
(33%)a
—
—
|
|
+
|
Hard
Hard
|
|
+
]
|
d
b
b
Xylan
—
—
Not formed —b
Not miscible
Miscible
—
235 1C
228 1C
228 1C
232 1C
223 1C
39%
48%
47%
57%
59%
—
—
[C2mim]+ 32%
|
Soft
170
130
130
120
| (foam) |
+
b
b
b
[P6,6,6,6
[P8,8,8,8
[P6,6,6,14
]
]
(32%)a
(32%)a
(32%)a
Not formed —b
Not formed —b
Not misciblec
Not misciblec
Not misciblec
—
—
—
—
+
b
+
]
|
Hard
|
|
d
b
b
Lignin
—
—
Not formed —b
Miscible
Miscible
Not miscible
Not miscible
—
—
256 1C
269 1C
287 1C
250 1C
225 1C
67%
74%
61%
71%
71%
—
—
d
[C2mim]+ 14%
|
|
|
|
Hard
| (foam) |
+
[P6,6,6,6
[P8,8,8,8
[P6,6,6,14
]
]
(14%)a
(14%)a
(14%)a
Hard
Hard
Hard
190
140
|
|
|
|
|
|
+
+
]
Slightly miscible —d
a
b
They have the same conversion ratio as the [C2mim]+polymers because we conducted only exchange of the free cation species. Not evaluated
c
d
because the samples did not form thin films. The film was not miscible with water but broken and dispersed in water. Not detected.
Thermal properties of the polymers regarding flame retardancy
were investigated (see Table 1). The residual weight of each
sample at 400 1C upon TGA under an air atmosphere was checked
as an index of flame retardancy8 (see Fig. S4, ESI†). As mentioned
above, phosphoric acid-type flame retardants are effective for
blocking oxygen and heat by generation of polyphosphate as char
when combusted.8 The residual weight of untreated cellulose at
400 1C was 11%. The residual weight of CP, which is a typical
cellulose plastic, at 400 1C was 12%, equivalent to that of untreated
Fig. 2 Thin films of [C2mim]+cellulose (left), [C2mim]+xylan (centre) and
[C2mim]+lignin (right) after hot pressing.
were flexible and transparent, while the film of [C2mim]+lignin was cellulose. By contrast, the residual weight of [C2mim]+cellulose
hard and not transparent. We developed thermoformable films was 52%, suggesting flame retardancy. Furthermore, the
from the three types of polymers derived from plant cell walls with residual weight of [C2mim]+cellulose (52%) is larger than each
[C2mim][(MeO)(H)PO2] at a low substitution ratio (r33%).
residual weight of untreated cellulose (11%) and that of
Thermal properties of the polymers regarding thermo- [C2mim][(MeO)(H)PO2] (41%). The increase should be attributed
plasticity were investigated (Table 1). The softening points to the effect of immobilised phosphoric acid on cellulose. The
of the samples were briefly checked while elevating the immobilised phosphoric acid on cellulose promotes carbonisation
temperature. Both [C2mim]+cellulose and [C2mim]+xylan were of the polysaccharides when exposed to high temperatures.5,8,9
remarkably softened at about 160 and 170 1C, respectively. On The residual weight at 400 1C of the flame-retardant cellulose
the other hand, the clear softening point of [C2mim]+lignin previously reported (i.e. the cellulose derivative prepared by intro-
was not observed (namely, it was still hard up to 250 1C). duction of the phosphoric acid-type flame retardant into CP
However, it seems that the fluidity of the molecules of through two steps, as mentioned in the introduction section)
[C2mim]+lignin was improved by heating because [C2mim]+lignin was 30–40%.9 Based on these results, [C2mim]+cellulose showed
formed a thin film upon hot pressing. TGA was conducted not only thermoplasticity and flame retardancy while maintaining
under air to investigate the thermal decomposition tempe- many hydroxyl groups, but also a better flame retardancy
rature (the temperature when the residual weight is 95%). compared with the flame-retardant plastic previously reported.
[C2mim]+cellulose and [C2mim]+xylan showed decomposition Just as a reference, the TGA curve of [C2mim]+cellulose mea-
temperatures of 230 and 235 1C (full TGA curves in Fig. S4, sured under N2 is shown in Fig. S5 (ESI†). The residual weight
ESI†). Thus, these samples showed the potential for use as of [C2mim]+xylan under air was 48% at 400 1C, whereas that
thermoplastics, because the thermal decomposition tempera- of untreated xylan was 39%, indicating an improvement in
ture was higher than the softening points (160 1C and 170 1C). flame retardancy. The improved value was similar to that of
[C2mim]+lignin showed
269 1C, which was slightly higher than those of polysaccharide 74% at 400 1C, whereas that of untreated lignin was 67%. The
derivatives. small difference was presumably due to the flame retardancy of
a
decomposition temperature of [C2mim]+cellulose. The residual weight of [C2mim]+lignin was
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2019
New J. Chem.

View Article Online
Paper
NJC
unreacted lignin. Together, these results confirm that each of [C2mim][(MeO)(H)PO2] at a low substitution ratio (r33%).
the three components of plant cell walls substituted with Both flame retardancy and thermoplasticity at the low sub-
[C2mim][(MeO)(H)PO2] shows flame retardancy upon TGA.
stitution ratio are clearly attributed to functionalisation by
It is noted here that the thermal decomposition temperature of only a single species, [C2mim][(MeO)(H)PO2]. On the other
un-derivatised cellulose was 306 1C, and that of [C2mim]+cellulose hand, [C2mim]+lignin did not show a clear macroscopic
was 230 1C. A decrease in the decomposition temperature of softening behaviour during the experiment for checking the
[C2mim]+cellulose can be explained by the decomposition and softening point, although it could form a thin film with hot
simultaneous char formation temperature of [C2mim][(MeO)(H)PO2] pressing.
(see also Fig. S1, ESI†).30,31 The thermal decomposition temperature
of un-derivatised xylan was 235 1C, and there was almost no change
Flame-retardant cell-wall-derived thermoplastics with
phosphonium-based ionic liquids
after substitution with [C2mim][(MeO)(H)PO2] (228 1C). This
was probably because the original decomposition temperature We considered whether lignin could be macroscopically softened
of untreated xylan was near the thermal decomposition tem- further through more precise design of ionic liquids. In this study,
perature of the phosphorus component. The decomposition we focused on phosphonium cations with long alkyl chains. The
temperature of lignin also showed a slight difference before phosphonium cations have high molecular weights and bulky
and after derivatisation (256 and 269 1C). The reason for the structures; thus, they were expected to have large free volumes
decomposition temperature not decreasing to around 230 1C and increase the free volume of lignin through substitution.
may be due to the low conversion ratio (14%).
Consequently, derivatisation with the phosphonium cations was
The samples were then burned (Fig. 3 shows the photos of expected to increase the thermal fluidity of the lignin molecules
the samples after burning, and movies regarding CP and and decrease the softening point.11 Thus, the softening points of
[C2mim]+cellulose are shown in Movies 1–4, ESI,†). When the the lignin derivatised with long-alkyl-chain phosphonium cations
film of CP was in contact with the flame of an alcohol lamp, it were investigated, with a focus on the alkyl chain length of the
burned completely with dripping and with slight char formation. phosphonium cations (length itself and its symmetry/asymmetry)
On the other hand, when three [C2mim]+ samples were in and the consequent size of the free volume. Furthermore,
contact with the flame of the alcohol lamp, combustion started derivatisation with the phosphonium cations was also expected
but a char layer immediately formed and self-extinguished the to drastically control various properties of the biopolymers. For
fire. We concluded that this was caused by the flame-retardant example, substitution with phosphonium cations would make
effect of [C2mim][(MeO)(H)PO2], because this behaviour is con- the polymers hydrophobic while the imidazolium-type polymers
sistent with that of phosphoric acid-type flame retardants.
The formed char layers on the samples have a foamed
were hydrophilic.
The cations of [C2mim]+cellulose, [C2mim]+xylan, and
structure; a foamed char layer is generally quite effective to [C2mim]+lignin were exchanged with tetrahexylphosphonium
show flame-retardancy because it shuts heat out. The scanning ([P6,6,6,6]+), tetraoctylphosphonium ([P8,8,8,8]+) and trihexy-
electron microscope (SEM) images of the samples showed the (tetradecyl)phosphonium ([P6,6,6,14]+) cations (the structures
fine cells divided by thin char layers (Fig. S6, ESI†). The foamed are shown in Fig. 1). Both the [P8,8,8,8] cation and [P6,6,6,14
]
char layer is characteristic of intumescent flame-retardants and cation have 32 carbon atoms but they are symmetric and
formed by simultaneous char formation and gas generation. asymmetric, respectively. The fractional free volumes of the
The gas is presumably generated by decomposition of the ionic liquids have not been reported and they was calculated in
imidazolium cation since it contains N atoms like in melamine, this study. The fractional free volumes of [C2mim][(MeO)(H)PO2],
which is a popular nitrogen-type flame-retardant. The char [P6,6,6,6][(MeO)(H)PO2], [P8,8,8,8][(MeO)(H)PO2] and [P6,6,6,14]-
layer was additionally confirmed to include P atoms by using [(MeO)(H)PO2]) were 0.311, 0.363, 0.369, and 0.367, respectively.
SEM-energy dispersive X-ray spectroscopy (EDX) (Fig. S7, ESI†): The higher fractional free volumes of phosphonium salts may
the char layer includes polyphosphate. Utilisation of typical ILs indicate the potential to decrease the softening points of polymers,
as intumescent flame retardants has not been reported, to the including lignin. The abbreviations used for the derivatives are
best of our knowledge.
[P8,8,8,8]+cellulose and so on.
We here developed flame-retardant thermoplastics from the
three types of polymers derived from plant cell walls with [P8,8,8,8
Characteristics of the samples are shown in Table 1.
+
]
and [P6,6,6,14]+cellulose were able to form thin films
with hot pressing and [P6,6,6,14]+cellulose was brittle (the photos
are shown in Fig. S8, ESI†). On the other hand, [P6,6,6,6]+cellulose
was not able to form films with hot pressing. In the case of xylan,
[P6,6,6,14]+xylan was able to form thin films with hot pressing,
while it was brittle, but [P6,6,6,6]+ and [P8,8,8,8]+xylan were not.
+
In the case of lignin, [P6,6,6,6]+, [P8,8,8,8
]
and [P6,6,6,14]+lignin
were all able to form thin films with hot pressing, while the
[P8,8,8,8]+lignin film was brittle. All samples substituted with
long-alkyl-chain phosphonium cations were harder (i.e., less
flexible) than the [C2mim]+ samples. In summary, [P8,8,8,8]+ and
Fig. 3 Char formation of the thin films of [C2mim]+cellulose (left),
[C2mim]+xylan (centre) and [C2mim]+lignin (right) after burning and extin-
guishing of the fire.
New J. Chem.
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2019

View Article Online
NJC
Paper
+
[P6,6,6,14]+cellulose; [P6,6,6,14]+xylan and [P6,6,6,6]+, [P8,8,8,8
[P6,6,6,14]+lignin formed thin films.
]
and
The residual weight of the samples at 400 1C was determined
by using TGA (see Table 1 and Fig. S4, ESI†). The residual
Thermal properties with respect to thermoplasticity were weights of [P6,6,6,6]+, [P8,8,8,8]+ and [P6,6,6,14]+cellulose were 57, 62
investigated (see Table 1). While [C2mim]+cellulose was mark- and 62%, respectively, suggesting a higher flame retardancy
edly softened at about 160 1C, [P6,6,6,6]+, [P8,8,8,8
]
and compared with that of [C2mim]+cellulose (52%). The residual
+
[P6,6,6,14]+cellulose were markedly softened at about 140, 140 weights of [P6,6,6,6]+, [P8,8,8,8
]
and [P6,6,6,14]+xylan were 47, 57
+
and 120 1C, respectively (although [P6,6,6,6]+cellulose did not and 59%, respectively, suggesting also a higher flame retar-
form a film with hot pressing). The increase of the thermal dancy compared with that of [C2mim]+xylan (48%). The residual
fluidity of the samples may be due to the long-alkyl-chain weights of [P6,6,6,6]+, [P8,8,8,8]+ and [P6,6,6,14]+lignin were 61, 71
phosphonium cations of the bulky structures, which have large and 71%, respectively, suggesting lower or similar flame retar-
free volumes (0.311 for [C2mim][(MeO)(H)PO2] vs. 0.363–0.369 dancy compared with that of [C2mim]+lignin (74%). Based on
for phosphonium[(MeO)(H)PO2]). Therefore, the phosphoinum- these results, the flame retardancy of the derivatives with long-
type ionic liquids consequently increased the free volume of the alkyl-chain phosphonium cations was confirmed by TGA.
derivatised polymers. The softening points of [P6,6,6,6]+cellulose
The samples were then burned (Fig. S9 (ESI†) shows the
and [P8,8,8,8]+cellulose were the same, because there was no photos of the samples after burning). When the films of
substantial difference between the fractional free volumes of [P8,8,8,8]+, [P6,6,6,14]+cellulose, [P6,6,6,14]+xylan, [P6,6,6,6]+, and
[P6,6,6,6][(MeO)(H)PO2] and [P8,8,8,8][(MeO)(H)PO2] (0.363 vs. 0.369). [P8,8,8,8]+lignin made contact with the flame of the alcohol
Further investigation is required to explain the lower softening lamp, combustion started but a char layer was immediately
point of [P6,6,6,14]+cellulose despite the similar fractional free formed and then the fire was self-extinguished. The actual
volume ([P6,6,6,14][(MeO)(H)PO2]: 0.367).
flame retardancy of the derivatives with long-alkyl-chain phos-
In the case of xylan, while [C2mim]+xylan was markedly phonium cations was confirmed.
softened at about 190 1C, [P6,6,6,6]+, [P8,8,8,8]+ and [P6,6,6,14]+xylan
The formed char layers were smooth, not foamed structure
were markedly softened at about 130, 130 and 120 1C, respec- (see Fig. S9 (ESI†), and a SEM image of [P8,8,8,8]+cellulose is
tively; that is, they showed the same tendency as the cellulose shown in Fig. S10, ESI†). Therefore, the phosphonium-type ILs
derivatives. In the case of lignin, [P6,6,6,6]+ and [P8,8,8,8]+lignin did not work as intumescent flame retardants. It is due to the
were markedly softened at about 190 and 140 1C, whereas absence of a N atom in the phosphonium cations unlike
[C2mim]+lignin did not show a macroscopic softening beha- imidazolium cations. On the other hand, phosphine is also
viour, as mentioned above. On the other hand, the softening of known as a phosphorus-type flame retardant and phosphonium
[P6,6,6,14]+lignin was not confirmed, although it formed a thin cations may play the role of such flame retardants. SEM-EDX
film upon hot pressing. Thus, totally different behaviours were analysis indicates that the char layer contains polyphosphate
observed for the lignin derivatives compared to the poly- (Fig. S11, ESI†).
saccharide derivatives, which cannot be explained by the effects
Solubility of the polymers in water was investigated to
of free volume. The differences were probably because lignin determine the controllability of properties of the polymers.
has a three-dimensional structure, whereas polysaccharides [C2mim]+cellulose was miscible in water. On the other hand,
basically have straight-chain structures (although xylan has the cellulose derivatives produced by exchanging the cation for
+
very short branches). It might be necessary to confirm the phosphonium cations ([P6,6,6,6]+, [P8,8,8,8
]
and [P6,6,6,14]) were
increase in the free volumes of the polymers substituted by immiscible in water. Notably, the film of [P6,6,6,14]+cellulose was
the ionic liquids, in order to elucidate these behaviours. How- broken when immersed in water, which may indicate a relatively
ever, such investigation is too complex to attempt at this early weak entanglement of the polymer chains in [P6,6,6,14]+cellulose.
stage because of the complex structure of lignin. As a summary [P6,6,6,6]+, [P8,8,8,8]+ and [P6,6,6,14]+xylan were immiscible in water,
of this experiment, it was possible to lower the softening point whereas [C2mim]+xylan was miscible. These films were also
by introducing long-chain phosphonium cations.
broken in water, similar to [P6,6,6,14]+cellulose. In the case of
+
TGA was conducted under air to investigate the thermal decom- lignin, [P6,6,6,6
]
and [P8,8,8,8]+lignin were immiscible in water,
position temperature (see Table 1 and Fig. S4, ESI†). [P6,6,6,6]+, whereas [C2mim]+lignin was miscible. On the other hand,
+
[P8,8,8,8
]
and [P6,6,6,14]+cellulose showed decomposition tempera- [P6,6,6,14]+lignin partially dissolved in water. The hydrophobicity
tures of 234, 248 and 243 1C, respectively, which were sufficiently of ions generally becomes stronger by increasing the alkyl chain
higher than their softening points (while [P6,6,6,6]+cellulose did length, but these results showed a different tendency. Therefore,
+
not form a film with hot pressing). [P6,6,6,6]+, [P8,8,8,8
]
and we considered the possibility that the cation was not introduced
[P6,6,6,14]+xylan showed decomposition temperatures of 228, 232 in a 1 : 1 ratio relative to the anion. However, the ratio of anion to
and 223 8C, respectively, which were also higher than their soft- cation was measured using 31P NMR spectroscopy and was
ening points (while [P6,6,6,6]+ and [P8,8,8,8]+xylan did not form films found to be 1 : 1. Further investigation is necessary to clarify
with hot pressing). [P6,6,6,6]+ and [P8,8,8,8]+lignin showed decomposi- why [P6,6,6,14]+lignin was partially soluble in water. In summary,
tion temperatures of 287 and 250 1C, which were higher than their it was possible to add hydrophobicity to the cellulose, xylan and
softening points. [P6,6,6,14]+lignin showed a decomposition tempera- lignin derivatives by substitution with long-alkyl-chain phos-
ture of 225 1C, while its softening point could not be determined phonium cations, suggesting that the properties of the polymers
(although it formed a thin film with hot pressing).
can be controlled by introduction of different cation species.
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2019
New J. Chem.

View Article Online
Paper
NJC
There is much room to design the properties of the biopolymers Sigma-Aldrich Co., LLC, and used as received. Triethyl phosphate,
because there are various cations such as ammonium, pyridinium trihexylphosphine, 1-bromohexane, tetra-n-octylphosphonium
and many inorganic cations.
bromide and dimethyl phosphite were purchased from Tokyo
The results above show that introduction of a long-chain Chemical Industry Co., Ltd, and used as received. Dialysis
phosphonium cation enables adduction of hydrophobicity and membrane (MWCO 1000) was purchased from Funakoshi Co.,
control of the softening temperature. Although some of the Ltd, and used as received. 1H and 31P NMR spectra were
samples are not amenable to forming films, at least by hot recorded on a JEOL ECA-600 spectrometer. FT-IR spectra were
pressing (we have not attempted solution casting), they can be recorded on a Thermo Scientific NICOLETis10 spectrometer.
used as composites. For instance, such polymers can be used as Thermogravimetric analysis was carried out on a Shimazu DTG-
high-molecular-weight flame retardants. Low-molecular-weight 60AH analyser. Softening points were measured using an
flame retardants have problems, e.g. bleeding out and ATM-01, AS ONE.
damaging the human body via volatilisation. By contrast,
flame-retardant polymers do not readily bleed out owing to
Synthesis of [C2mim]+cellulose, xylan, and lignin
entanglement with polymer matrices and have low toxicity
owing to their non-volatility.
Cellulose, hemicellulose and lignin (0.5 g), respectively, were
mixed with [C2mim][(MeO)(H)PO2] (5 g) and then heated in an
oil bath at 160 1C for different periods of time (cellulose and
xylan: 1 h, lignin: 3 h). After the reaction, the resulting mixture
Conclusions
was cooled to room temperature and diluted with 20 mL of
Thermoplasticity and flame retardancy can be simultaneously water (for cellulose and xylan) or methanol (for lignin). Dialysis
imparted to each of the three components of plant cell walls, was performed to remove the unreacted [C2mim][(MeO)(H)PO2]
while maintaining more than 66% of the hydroxyl groups in 500 mL water or methanol three times. The resulting
of the polymers, by substitution with [C2mim][(MeO)(H)PO2]. dialysed solution was dried under reduced pressure. After
[C2mim]+cellulose, [C2mim]+xylan and [C2mim]+lignin formed thin drying [C2mim]+cellulose, [C2mim]+xylan, and [C2mim]+lignin
films and their residual weights at 400 1C were 52, 48 and 74%, were obtained.
higher than those of the un-derivatised cellulose, xylan and lignin
The substitution ratios of the hydroxyl groups of the bio-
(10, 39 and 67%, respectively), showing a high flame retar- polymers with [C2mim][(MeO)(H)PO2] were measured by
dancy. In the actual combustion test, they extinguished the fire 31P NMR spectroscopy. Triethyl phosphate was added as a
before being completely burned. Long-alkyl-chain phospho- standard for quantification. The substitution ratios of [C2mim]+-
nium cations were introduced into each polymer instead of cellulose and [C2mim]+xylan were obtained as follows. First, we
the [C2mim] cation to control the softening point and other calculated DS by the following equation.
Triethyl phosphate ðgÞ ꢁ peak ratio
½C2mimꢂþsample added ðgÞ ꢁ Degree of substitution ðDSÞ
¼
Molecular weightðMWÞ of triethyl phosphate MW of one unit of cellulose or xylan þ ðMW of ½C2mimꢂ½ðMeOÞðHÞPO2ꢂꢀ32Þ ꢁ DS
properties such as hydrophobicity. It was confirmed that the where the peak ratio represents the integrated value of the
softening point of long-alkyl-chain phosphonium cation polymers phosphorus atoms in the anionic moiety of the sample when
can be controlled while maintaining the flame retardancy equi- the integrated value of triethyl phosphate is 1.0. Then, the
valent to that of the [C2mim]+ samples. In addition, almost all of substitution ratio can be expressed by the following equation.
the long-alkyl-chain phosphonium cation-based polymers became
DS
hydrophobic, clearly indicating that the design of the cation
species results in control of their properties.
Substitution ð%Þ ¼
ꢁ 100
Maximum value of the DS ð3 or 2Þ
where the maximum values of DS is 3 for cellulose and 2
for xylan.
Experimental
Materials and equipment
The substitution ratio of [C2mim]+lignin was obtained by the
following equation.
[C2mim][(MeO)(H)PO2], methanol, hexane, nitric acid, 0.1 mol Lꢀ1
silver nitrate solution and dichloromethane were purchased from
Kanto Chemical Co., Inc. and used as received. Filter paper
(No. 5C) derived from cellulose was purchased from Advantec
Co., Ltd and used as received. Xylan from beech wood was
purchased from Apollo Scientific Co., Ltd and used as received.
Alkali lignin (low-sulfonate content), trihexyltetradecyl-
Triethyl phosphate ðgÞ ꢁ peak ratio
Substitution ð%Þ ¼
MW of triethyl phosphate ðmol gꢀ1
Þ
1
ꢁ
ꢁ 100
OH content of untreated lignin ðmol gꢀ1
Þ
phosphonium bromide, Amberlite IRN 77A hydrogen form The OH content of untreated lignin we used was 4.26 ꢁ
and Amberlite IRN 78A hydroxide form were purchased from 10ꢀ3 (mol gꢀ1), measured as reported.32,33
New J. Chem.
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2019

View Article Online
NJC
Paper
Synthesis of tetrahexylphosphonium bromide ([P6,6,6,6]Br)
0.920 g cmꢀ3 for [P8,8,8,8][(MeO)(H)PO2], 0.924 g cmꢀ3 for
[P6,6,6,14][(MeO)(H)PO2].
Trihexylphosphine (6.29 g, 0.022 mol) and 1-bromohexane
(4.17 g, 0.025 mol) were added to hexane at room temperature
under an argon atmosphere. The resulting mixture was stirred
for 24 h at 160 1C. The resulting liquid was washed repeatedly
with hexane. The liquid was dried under reduced pressure at
room temperature for 24 h to yield [P6,6,6,6]Br as a solid.
1H NMR (600 MHz, DMSO-d6; Me4Si) d = 2.01–2.25 (m, 8H,
PCH2(CH2)4CH3), 1.38–1.45 (m, 8H PCH2CH2(CH2)3CH3), 1.31–
1.36 (m, 8H, P(CH2)2CH2(CH2)2CH3), 1.11–1.28 (m, 16H;
P(CH2)3(CH2)2CH3), 0.83 (t, J = 5.7 Hz, 12H, P(CH2)5CH3).
Synthesis of [phosphonium][(MeO)(H)PO2]
Here we describe the example of [P8,8,8,8][(MeO)(H)PO2]. The
bromide anion of [P8,8,8,8]Br (10 g) was converted into hydroxide
by mixing a [P8,8,8,8]Br/water/methanol solution with the anion
exchange resin (Amberlite IRN 78). An excess amount of
dimethyl phosphite relative to [P8,8,8,8][hydroxide] (5 eq) was
added. The resulting liquid was dried in vacuo and dissolved in
dichloromethane. The solution was washed repeatedly with
water. The dichloromethane layer was dried in vacuo at room
temperature to yield [P8,8,8,8][(MeO)(H)PO2]. [P6,6,6,6][(MeO)(H)PO2]
and [P6,6,6,14][(MeO)(H)PO2] were synthesised using the same
method.
Synthesis of [phosphonium]+cellulose, xylan, and lignin
Here we describe the example of [P8,8,8,8]+cellulose. The cation
of [C2mim]+cellulose (0.1 g) was converted into a proton by
mixing the aqueous solution of the imidazolium salt and the
cation exchange resin (Amberlite IRN 77). The bromide anion
of [P8,8,8,8]Br was converted into hydroxide by mixing the water/
methanol solution of the bromide salt with the anion exchange
resin (Amberlite IRN 78). An equimolar amount of [H]+cellulose
relative to [P8,8,8,8][hydroxide] was added. Dialysis was per-
formed to remove unreacted [P8,8,8,8][hydroxide] in 500 mL
methanol three times. The resulting liquid was dried under
reduced pressure at room temperature for 24 h to yield
[P8,8,8,8]+cellulose.
1H NMR (600 MHz; DMSO-d6; Me4Si), d = 6.88 and 5.95
(s, 1H, PH), 3.16 (d, J = 11.7 Hz 3H, POCH3), 2.10–2.15 (m, 8H,
PCH2(CH2)6CH3), 1.39–1.45 (m, 8H, PCH2CH2(CH2)5CH3),
1.31–1.36 (m, 8H, P(CH2)2CH2(CH2)4CH3), 1.22–1.26 (m, 32H,
P(CH2)3(CH2)4CH3), 0.83 (t, J = 7.0 Hz, 12H, P(CH2)7CH3).
Synthesis was confirmed in the same way as for the other
samples.
Conflicts of interest
1H NMR (600 MHz; DMSO-d6; Me4Si), d = 6.97 and 6.02 (1H,
P-H), 2.10–2.15 (m, 8H, PCH2(CH2)6CH3), 1.39–1.45 (m, 8H,
PCH2CH2(CH2)5CH3), 1.31–1.35 (m, 8H, P(CH2)2CH2(CH2)4-
CH3), 1.22–1.26 (m, 32H, P(CH2)3(CH2)4CH3), 0.83 (t, J = 7.0 Hz,
12H, P(CH2)7CH3), and many cellulose backbone signals
(Fig. S12, ESI†).
These are no conflicts to declare.
Acknowledgements
This research was supported in part by the COI program ‘‘Con-
struction of next-generation infrastructure using innovative
materials–Realization of a safe and secure society that can
coexist with the Earth for centuries’’, which is supported by
Ministry of Education, Culture, Sports, Science and Technology-
Japan (MEXT) and Japan Science and Technology (JST), the
Advanced Low Carbon Technology Research and Development
Program (ALCA, No. JPMJAL1104) and the Cross-ministerial
Strategic Innovation Promotion Program (SIP) of JST. This study
was also partly supported by KAKENHI (18K14281 from the
Japan Society for the Promotion of Science) and Leading Initia-
tive for Excellent Young Researchers (from MEXT).
The introduction of [P8,8,8,8]+cellulose was also confirmed by
the FT-IR spectrum (Fig. S2, ESI†). Peaks derived from the C–H
bond of phosphonium cation at 2854 and 2924 cmꢀ1 were
detected in the spectrum of [P8,8,8,8]+cellulose, in addition to
the P–O–C bond at 821 cmꢀ1 and the PQO bond at 1211 cmꢀ1
All the other samples were synthesised using the same
method and the synthesis was confirmed in the same way for
the other samples (Fig. S2, S13 and S14, ESI†).
.
Calculation of free volumes of ionic liquids
To calculate the free volumes of the ionic liquids, density
functional theory (DFT) calculations were used, as reported.34
We write the method briefly here. DFT calculations for the ionic
liquids were performed by using the Gaussian 09 program
package.35 Full geometry optimization (gas phase) was con-
ducted by using 6-311+G(d,p) basis sets based on Becke’s three-
parameter hybrid method36 with LYP correlation (B3LYP).37,38
Ionic volumes of the ionic liquids were estimated39 and can be
considered as the van der Waals volume of the ionic liquids.40
We used the density of the [phosphonium][(MeO)(H)PO2] for
the calculation. The synthesis of the [phosphonium][(MeO)-
(H)PO2] is described below.
References
1 K. J. Edgar, C. M. Buchanan, J. S. Debenham, P. A. Rundquist,
B. D. Seiler, M. C. Shelton and D. Tindall, Prog. Polym. Sci.,
2001, 26, 1605–1688.
2 D. Klemm, B. Heublein, H. P. Fink and A. Bohn, Angew.
Chem., Int. Ed., 2005, 44, 3358–3393.
3 T. Mekonnen, P. Mussone, H. Khalil and D. Bressler,
J. Mater. Chem. A, 2013, 1, 13379–13398.
4 P. M. Kosaka, Y. Kawano, H. M. Petri, M. C. A. Fantini and
D. F. S. Petri, J. Appl. Polym. Sci., 2007, 103, 402–411.
5 B. K. Kandola, A. R. Horrocks, D. Price and G. V. Coleman,
J. Macromol. Sci., Polym. Rev., 1996, 36, 721–794.
The used densities are described here: 1.197 g cmꢀ3 for
[C2mim][(MeO)(H)PO2], 0.955 g cmꢀ3 for [P6,6,6,6][(MeO)(H)PO2],
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2019
New J. Chem.

View Article Online
Paper
NJC
6 S. L. Levan and J. E. Winandy, Wood Fiber Sci., 1990, 22, 24 J. Song, H. Fan, J. Ma and B. Han, Green Chem., 2013, 15,
113–131. 2619–2635.
7 K. Xie, A. GaO and Y. Zhang, Carbohydr. Polym., 2013, 98, 25 K. Ohira, Y. Abe, M. Kawatsura, K. Suzuki, M. Mizuno,
706–710. Y. Amano and T. Itoh, ChemSusChem, 2012, 5, 388–391.
8 R. Stevens, S. Daan, E. Van, R. Bezemer and A. Kranenbarg, 26 H. T. Vo, Y. J. Kim, E. H. Jeon, C. S. Kim, H. S. Kim and
Polym. Degrad. Stab., 2006, 91, 832–841.
9 B. Schartel, Materials, 2010, 3, 4710–4745.
10 G.-L. Wei, D.-Q. Li, M.-N. Zhuo, Y.-S. Liao, Z.-Y. Xie,
H. Lee, Chem. – Eur. J., 2012, 18, 9019–9023.
27 A. Pinkert, K. N. Marsh, S. S. Pang and M. P. Staiger, Chem.
Rev., 2009, 109, 6712–6728.
T.-L. Guo, J.-J. Li, S.-Y. Zhang and Z.-Q. Liang, Environ. 28 D. M. Suflet, G. C. Chitanu and V. I. Popa, React. Funct.
Pollut., 2015, 196, 29–46.
Polym., 2006, 66, 1240–1249.
11 D. Aoki and Y. Nishio, Cellulose, 2010, 17, 963–976.
12 S. Marceneiro, Q. Hu, A. M. A. Dias, I. Lobo, I. Dias,
29 J. L. E. Campbell, K. E. Johnson and J. R. Torkelson, Inorg.
Chem., 1994, 33, 3340–3345.
ˇ
´
E. Pinho, M. G. Rasteiro and H. C. Sousa, Ind. Eng. Chem. 30 I. Simkovic and K. Csomorova, Polym. Degrad. Stab., 2003,
Res., 2014, 53, 16061–16071. 80, 519–523.
13 J. Wu, j. Bai, Z. Xue, Y. Liao, X. Zhou and X. Xie, Cellulose, 31 W. D. Ellis, R. M. Rowell, S. L. Levan and R. A. Susott, Wood
2015, 22, 89–99. Fiber Sci., 1987, 19, 439–445.
14 X. Yang, N. Ge, L. Hu, H. Gui, Z. Wang and Y. Ding, Polym. 32 S. Suzuki, Y. Shibata, D. Hirose, T. Endo, K. Ninomiya,
Adv. Technol., 2013, 24, 568–575.
R. Kakuchi and K. Takahashi, RSC Adv., 2018, 8, 21768–
15 R. Sonnier, L. Dumazert, S. Livi, T. K. L. Nguyen, J. Duchet-
21776.
Rumeau, H. Vahabi and P. Laheurte, Polym. Degrad. Stab., 33 A. Granata and D. S. Argyropoulos, J. Agric. Food Chem.,
2016, 134, 186–193. 1995, 43, 1538–1544.
16 H. Miyafuji and Y. Fujiwara, Holzforschung, 2013, 67, 787–793. 34 T. Endo, S. Nemugaki, Y. Matsushita, Y. Sakai, H. Ozaki,
17 S. Han, J. Li, S. Zhu, R. Chen, Y. Wu, X. Zhang and Z. Yu,
Y. Hiejima, Y. Kimura and K. Takahashi, Chem. Phys., 2016,
472, 128–134.
BioResources, 2009, 4, 825–834.
18 R. P. Swatloski, S. K. Spear, J. D. Holbrey and R. D. Rogers, 35 M. J. Frisch, G. W. Trucks and H. B. Schlegel, et al., Gaussian
J. Am. Chem. Soc., 2002, 124, 4974–4975. 09, Revision D.01, Gaussian, Inc., Wallingford CT, 2013.
19 H. Wang, G. Gurau and R. D. Rogers, Chem. Soc. Rev., 2012, 36 A. D. J. Becke, Chem. Phys., 1993, 98, 5648–5652.
41, 1519–1537.
37 C. Lee, W. Yang and R. G. Parr, Phys. Rev. B: Condens. Matter
Mater. Phys., 1988, 37, 785–789.
38 B. Miehlich, A. Savin, H. Stoll and H. Preuss, Chem. Phys.
Lett., 1989, 157, 200–206.
20 A. Brandt, J. Grasvik, J. P. Hallett and T. Welton, Green
Chem., 2013, 15, 550–583.
21 H. Ohno and Y. Fukaya, Chem. Lett., 2009, 38, 2–7.
22 M. Armand, F. Endres, D. R. MacFarlane, H. Ohno and 39 D. F. Parsons and B. W. Ninham, J. Phys. Chem. A, 2009, 113,
B. Scrosati, Nat. Mater., 2009, 8, 621–629. 1141–1150.
23 I. Kilpelainen, H. Xie, A. King, M. Granstrom, S. Heikkinen 40 T. Endo, H. Murata, M. Imanari, N. Mizushima, H. Seki,
¨
and D. S. Argyropoulos, J. Agric. Food Chem., 2007, 55,
S. Sen and K. Nishikawa, J. Phys. Chem. B, 2013, 117,
9142–9148.
326–332.
New J. Chem.
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2019