CHEMSUSCHEM
FULL PAPERS
currently one of the best ILs for cellulose dissolution and the
cheapest in a bulk technical purity (BASF). The fourth genera-
tion, developed through work performed in our own laborato-
ries,[12] afforded new distillable ionic liquids (DILs), such as
1,1,3,3-tetramethylguanidinium propionate ([TMGH][CO2Et]),
that not only dissolved cellulose but were also distillable as
a means of purification. The novelty of these structures was
that these previous generations were essentially not distillable
on a large scale without the use of a high vacuum (typically
0.05 mbar). These superbase-derived ILs have been developed
further to yield RTILs that are now as low in viscosity as
[emim][OAc] and able to dissolve cellulose to high concentra-
tions.[13] Although distillation is a useful method for the purifi-
cation of ILs, it is still energy intensive and will likely only be
employed after many process cycles once impurities get to
a prohibitive level for a particular process. Therefore, new recy-
cling strategies are needed.
er, one important problem was that after several process
cycles, the separated lignin started to accumulate in the IL,
which would require additional purification techniques. Singh
et al.[26] reported recently in a conference presentation that
NaOH solutions can also have the same phase-separation
properties for [emim][OAc] as tripotassium phosphate. Howev-
er, the presence of hydroxide in enriched imidazolium-based
ILs can be problematic because of instability issues (e.g.,
through carbene formation).[27] With these kosmotrope-related
phase-separation mixtures there is still also the problem of IL
contamination through cation decomposition, anion exchange
with the kosmotropic anion, and the ubiquitous build-up of
lignocellulose-based contaminants, such as monomers, dimers,
oligomers, inorganic salts, and silicates. Typically, the salts are
recovered from the aqueous layers, which is quite energy in-
tensive. Therefore, more efficient alternatives are required, in
particular, those that minimize the use of kosmotropic re-
agents. To achieve this, it is clear that the hydrophobicity of ILs
must be increased to allow more efficient phase separation
from aqueous solutions.
As phase separation is a traditional method for the separa-
tion of components, it is logical that new generations of ILs for
cellulose processing should be phase-separable ionic liquids
(PSILs). This circumvents the need for IL distillation completely
and potentially offers considerable reduction in process energy
costs. This is best achieved by the introduction of hydrophobic
moieties into the IL to make them phase separable in the pres-
ence of water or saline solutions. Recently, we submitted
a patent application that describes this concept,[14] in which
phosphonium-based ILs in combination with cosolvents al-
lowed cellulose solubility. The phopshonium ILs in these mix-
tures are recyclable by phase separation from aqueous mix-
tures. Several reports have already described a number of po-
tential PSIL systems, which include the use of imidazolium-
and phosphonium-based salts with kosmotropic salt solu-
tions,[15–18] imdazolium-based ILs with common hydrophobic
anions, and phosphonium ILs with hydrophobic fluorinated
anions.[19–21]
Abe et al.[22] have demonstrated the use of tetrabutylphos-
phonium hydroxide (60 wt%) for the dissolution of cellulose at
room temperature. This solution is very effective to dissolve
cellulose but its application is limited as increasing the temper-
ature of these phosphonium hydroxide solutions causes irre-
versible decomposition to the phosphine oxides. Another arti-
cle by Keskar et al.[23] describes the use of long-chain phospho-
nium chlorides for the extractive fractionation of wood. How-
ever, this IL does not dissolve cellulose but does dissolve
lignin, which makes it a suitable option for the extraction of
lignin from lignocellulosic materials. Of all of these academic
reports, only one concept demonstrates the phase separation
of an IL that is also capable of dissolving cellulose. This can be
achieved if kosmotropic salt solutions, such as aqueous tripo-
Long-chain phosphonium-based ILs, in particular, have
a high degree of hydrophobicity, especially compared to
common imidazolium-based ILs. The typical long alkyl chains
on the phosphorus center contribute a high degree of hydro-
phobic character, which leads to a reduced miscibility with
water. In addition, phosphonium-based ILs have a number of
attractive advantages compared to imidazolium-based ILs,
such as increased thermal and chemical stabilities, which allow
much wider processing temperatures. As such, this may also
potentially circumvent the problem of IL reactivity with the
lignocellulosic solutes, as has been shown to occur with imida-
zolium-based ILs.[28,29]
Potential disadvantages of the use of these structures are
the increased cost of the cations and increased toxicity. Some
phosphonium salts are reported to have a relatively high toxic-
ity,[30] but the ILs in this report and the area in general are not
yet well enough studied to give a conclusion on which struc-
tures are considered to be too toxic for application in industry.
It is not established if the toxicity relates to the fact that they
are phosphonium structures or to the increasing hydrophobici-
ty as chain-lengths increase. This is a complex issue, which
could be overshadowed considerably by economic motivations
in practice. Of course, it is expected that the improved recov-
ery of the IL will offset these issues to a large degree. As such,
we would like to present our work on the characterization of
hydrophobic PSILs (Figure 1) in terms of their phase separabili-
ty and biopolymer dissolution capabilities.
tassium phosphate, are able to phase-separate cellulose-dis- Results and Discussion
solving ILs, such as [emim][OAc].[24] One study by Shill et al.[25]
Water miscibility of the pure ILs
even incorporated this phase-separation method into an IL-
+
+
based pretreatment process to enhance biofuel production
from recovered polysaccharide pulp. This proved relatively ef-
fective and even afforded improved enzymatic hydrolysis ki-
netics over the phosphate-free process. This was attributed to
the improved separation of lignin from polysaccharide. Howev-
The chloride and acetate ILs of cations [P4444
]
(C16), [P8881
]
+
+
+
(C25), [P14444
] (C26), [P8888] (C32), and [P14666] (C32) were pur-
chased or prepared (arranged in order of increasing cation
carbon content). Initial water solubility studies showed that ILs
+
from cations [P8881]+, [P8888]+, and [P14666
]
were able to phase
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 0000, 00, 1 – 14
&
2
&
ÞÞ
These are not the final page numbers!