Acylation of cellulose in a novel solvent system: Solution of dibenzyldimethylammonium fluoride in DMSO

Article abstract of DOI:10.1016/j.carbpol.2013.09.043

A novel electrolyte, dibenzyldimethylammonium fluoride has been obtained essentially anhydrous (BMAF-0.1H₂O) by a simple route. Its thermal stability, relative to tetra(1-butyl)ammonium fluoride trihydrate (TBAF3H ₂O) has been demons

Full text of DOI:10.1016/j.carbpol.2013.09.043

Carbohydrate Polymers 101 (2014) 444–450
Contents lists available at ScienceDirect
Carbohydrate Polymers
journal homepage: www.elsevier.com/locate/carbpol
Acylation of cellulose in a novel solvent system: Solution of
dibenzyldimethylammonium fluoride in DMSO
Romeu Casarano, Paulo A.R. Pires, Omar A. El Seoud∗
Institute of Chemistry, University of São Paulo, P.O.B. 26077, 05513-970 São Paulo, Brazil
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 3 July 2013
Received in revised form 2 September 2013
Accepted 14 September 2013
Available online 21 September 2013
A novel electrolyte, dibenzyldimethylammonium fluoride has been obtained essentially anhydrous
(BMAF-0.1H2O) by a simple route. Its thermal stability, relative to tetra(1-butyl)ammonium fluoride
trihydrate (TBAF 3H2O) has been demonstrated by thermogravimetric analysis. DMSO solution of
(BMAF-0.1H2O) dissolves microcrystalline- and fibrous celluloses; the dissolved biopolymers have been
acylated by ethanoic-, butanoic-, and hexanoic anhydride. The degrees of substitution of the esters are
higher than with TBAF 3H2O/DMSO. The reasons are discussed in terms of differences in electrolyte
structure and contents of water of hydration, whose presence leads to side reactions and decreases of the
basicity of (F−). This conclusion is corroborated by molecular dynamics simulations of the interactions of
glucose dodecamer/R4NF-hydrate/DMSO. These show that the interactions oligomer-F−-water is oper-
ative only for TBAF 3H2O/DMSO. The efficiency of BMAF-0.1H2O/DMSO is explained based on better
accessibility of the biopolymer due to efficient hydrogen-bonding between its hydroxyl groups and the
essentially desolvated (F−).
Keywords:
Cellulose acylation
Dibenzyldimethylammonium
fluoride-hydrate
Molecular dynamics
Tetra(1-butyl)ammonium fluoride-hydrate
© 2013 Elsevier Ltd. All rights reserved.
1. Introduction
dissolve cellulose, e.g., tetramethylammonium fluoride, or its effi-
ciency is low, benzyltrimethylammonium fluoride-H2O (Koehler
Quaternary ammonium fluoride hydrates (R4NF-hydrates) in
aprotic solvents (e.g., DMSO and THF) have been used as depro-
tecting agents and source of (F−) in organic synthesis, e.g., in
desilylation, cyclization, conjugate addition, acylation, alkylation,
condensation, and elimination (Fieser, 1967; Pliego & Pilo-Veloso,
2008; Sun & DiMagno, 2005, 2006). Tetra(1-butyl)ammonium
fluoride trihydrate (TBAF 3H2O)/DMSO has been employed for
cellulose dissolution and its subsequent derivatization, into, e.g.,
carboxylic acid esters (Casarano, Nawaz, Possidonio, da Silva, &
El Seoud, 2011; Heinze & Koehler, 2010; Leibert & Heinze, 2005),
ethers (Heinze, Wanga, Koschella, Sullob, & Foster, 2012; Heinze,
Lincke, Fenn, & Koschella, 2008), and for regioselective deacyla-
tion of carboxylic esters (Xu & Edgar, 2012; Zheng, Gandour, &
Edgar, 2013). The small volume and high charge density of (F−) play
an important role in the efficiency of this solvent system because
hydrogen-bonding between the anion and the hydroxyl groups
of the anhydroglucose units, AGU, of the biopolymer disrupts the
strong intermolecular bonding present, leading to its dissolution.
The use of (R4NF-hydrates/DMSO) is not, however, without lim-
itations, see parts (A–E) of Scheme 1. Depending on the strength of
association between the constituent ions, the electrolyte may not
& Heinze, 2007). The associated water of hydration leads to
hydrolysis of the acylating agent (Ass, Frollini, & Heinze, 2004;
Xu, Li, Tate, & Edgar, 2011), combined with a decrease of the
degree of substitution, DS, of the produced ester. This occurs
via a general base catalyzed hydrolysis (Scheme 1A), and/or a
ketene intermediate (Scheme 1B) (Casarano et al., 2011; Zheng
et al., 2013). The electrolyte itself may undergo elimination reac-
tions, including Hofmann elimination (Scheme 1C) (Albanese,
Landini, & Penso, 1998; Sharma & Fry, 1983) and/or degrada-
tion via an ylide intermediate, akin to the E1cB mechanism,
except that the species generated is a zwitterion, and not an
anion (Scheme 1D) (Chempath, Boncella, Pratt, Henson, & Pivovar,
2010; Makosza & Chesnokov, 2003). Finally, tetraallylammonium
fluoride-monohydrate (TAAF H2O), undergoes polymerization to
form linear- as well as cross-linked cyclopolymers, that do not
dissolve cellulose (Scheme 1E) (Casarano et al., 2011). Note that
the synthesis of anhydrous TBAF by the reaction of C6F6 with
tetra(1-butyl)ammonium cyanide is not a satisfactory solution to
the water-mediated side reactions, because one of the reagents is
expensive (US$ 25.6/g (C4H9)4NCN, Sigma Aldrich); the resulting
solution in DMSO (a mixture of C6(CN)6 and TBAF) is not stable; it
does not dissolve cellulose after storage for 28 h at room tempera-
ture, RT (Koehler & Heinze, 2007).
In summary, although R4NF-hydrates in DMSO are efficient sol-
vents for cellulose, their use is hampered by the above-mentioned
∗
Corresponding author. Tel.: +55 11 3091 3874; fax: +55 11 3091 3874.
E-mail address: elseoud@usp.br (O.A. El Seoud).
0144-8617/$ – see front matter © 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.carbpol.2013.09.043

R. Casarano et al. / Carbohydrate Polymers 101 (2014) 444–450
445
Scheme 1. Possible side reactions of tetraalkylammonium fluoride-hydrates. The schemes are for ester deacylation by the water of hydration, where the (F−) is acting as a
general-base (1A); ester deacylation via ketene intermediate (1B); electrolyte degradation by Hofmann elimination (1C); electrolyte degradation via ylide mechanism (1D);
electrolyte polymerization/crosslinking (1E).
problems. Therefore, the search for a thermally stable quaternary
ammonium fluoride with low water content is important, not only
for cellulose chemistry, but also for organic synthesis.
DPv = 175). Cotton sheets (Nitro Química S.A., São Paulo) were cut
into stripes, grounded, and sieved (100–200 mesh); DPv = 920.
We report here on an electrolyte, dibenzyldimethylammonium
fluoride, that is easy to obtain in practically anhydrous form
(BMAF-0.1H2O) with satisfactory thermally stability (Casarano &
El Seoud, 2013). A solution of this electrolyte in DMSO was found
to be an efficient medium for dissolution and acylation of cel-
lulose under homogeneous reaction conditions. The DS values
obtained for cellulose esters were larger than those obtained in
TBAF 3H2O/DMSO and TAAF H2O/DMSO. The results of molecu-
lar dynamics (MD) simulations for the system glucose dodecamer
(hereafter designated as “oligomer”)/R4NF-hydrate/DMSO showed
that the oligomer-F−-water interactions are operative only for
TBAF–3H2O/DMSO. This explains the relative efficiency of BMAF-
0.1H2O/DMSO as solvent for cellulose, because the essentially
desolvated (F−) binds efficiently to the hydroxyl groups of the AGU
leading, most certainly, to better accessibility, hence higher reac-
tivity of the biopolymer chains.
2.2. Equipment
The value of DPv was determined by using a shear-dilution
Cannon-Fenske viscosimeter (Schott), inserted in Schott AVS 360
automatic viscosity determination equipment. 1H- and 13C NMR
measurements were recorded with Varian Innova-300 spectrom-
eter (300 MHz for 1H). Elemental analysis was performed with
Perkin-Elmer Elemental Analyzer CHN 2400. Melting points were
determined with either Electrothermal IA 6304 mp equipment
(dibenzyldimethylammonium bromide hydrate, BMABr-hydrate)
or TA Instruments DSC Q10 Differential Scanning Calorimeter
(BMAF-hydrate). Thermogravimetric (TG) analyses were per-
formed with TA Instruments Hi-Res TG 2950 Thermogravimetric
Analyzer.
2.3. Material characterization
DPv was determined (25 ◦C) from the intrinsic viscosity of
cellulose solution in CUEN/water (1:1, v/v) according to the recom-
mended procedure (ASTM, 2007). The DS of the cellulose carboxylic
esters was determined by the back-titration method (ASTM, 2004).
2. Experimental
2.1. Materials
The reagents were purchased from Alfa Aeser or Merck and
were purified, where required, as described elsewhere (Armagero
& Chai, 2003). Microcrystalline cellulose, MCC; Avicel PH 101 was
from FMC (Philadelphia; viscosimetric degree of polymerization,
2.4. NMR and thermal analyses
DMSO-d6 was used as solvent for 1H and 13C NMR measure-
ments.

446
R. Casarano et al. / Carbohydrate Polymers 101 (2014) 444–450
2.6. Determination of the water of hydration of BMABr and BMAF
The water contents of these electrolytes were calculated from
the corresponding 1H NMR spectra, by comparing the area of the
water peak with that of known hydrogen of the electrolyte: H-8
of BMABr, Fig. SM-1a (Fig. 1a of Supplementary Material); H-8 of
BMAF, Fig. SM-1b. These calculations indicated the following com-
positions: BMABr-0.5H2O and BMAF-0.1H2O.
2.7. Dissolution and acylation of cellulose in BMAF-0.1H2O/DMSO
Scheme 2. Synthesis of BMAF-hydrate. First, BMABr-hydrate is synthesized, and
then converted into the corresponding hydroxide, neutralized with HF, followed by
methanol evaporation.
A clear, isotropic solution was obtained by stirring the elec-
trolyte (3.3 g; 13.35 mmol BMAF) in DMSO (40 mL; 0.56 mol;
always freshly distilled from CaH2) for 15–30 min at 60 ◦C. MCC
or cotton cellulose (0.50 g; 3.09 mmol) was introduced into the
above-prepared solvent system, the suspension was heated at
80 ◦C for 20–40 min, under mechanical stirring, 570 10 rpm. Cel-
lulose dissolution was followed visually by placing a lamp behind
the reaction flask; clear solutions were obtained in all cases. The
required volume of acid anhydride was added, and the mixture was
kept under the above-mentioned experimental conditions during
the required length of time, vide infra. The resulting solution was
poured into hot ethanol (60 ◦C, 400 mL); the suspension was stirred
for 30–60 min; the precipitated solid was centrifuged at 3500 × g
(IEC Centra 244 MP4R). This procedure was repeated three more
times, the product was dried under reduced pressure at 50–60 ◦C,
for 48 h, in the presence of P4O10, and its DS determined.
Thermal analyses have been carried out under the following
conditions:
– Differential scanning calorimetry (DSC): N2 atmosphere,
100 mL min−1; sample mass = 6.0 mg; heating rate, 10◦ min−1; T
range = 25–160 ◦C; The DSC cell was calibrated with In
(mp = 156.6 ◦C; ꢀHm = 28.59 J g−1
ꢀHm = 111.40 J g−1);
)
and Zn (mp = 419.6 ◦C;
– TG: Isothermal TG, N2 atmosphere, 50 mL min−1
; sample
mass = 8.1 mg; T = 80 ◦C; 18 h; TG experiments, N2 atmosphere,
50 mL min−1; sample mass = 11.5 mg; T range from 35 to 600 ◦C;
heating rate = 10 ◦C min−1
.
3. Results and discussion
2.5. Synthesis of BMAF-hydrate
3.1. Synthesis and water content of BMAF-hydrate
Scheme 2 illustrates the synthesis of BMAF-hydrate.
BMABr was synthesized by reacting neat N,N,N-
dimethylbenzylamine and benzyl bromide for ca. 2.5 h at RT,
with 93% yield. The reaction conditions/yield are superior to those
reported elsewhere, e.g., the reaction of benzyl bromide, potassium
carbonate, and DMF, 80 ◦C, 48 h, 52% yield, (Busi et al., 2004), or
the reaction of benzyl bromide and N,N,N-dimethylbenzylamine,
−10 ◦C, 3 days at RT, no yield reported (Tutaj, Gonzalez-Perez,
Czapkiewicz, Del Castillo, & Rodriguez, 2001). 1H NMR analysis
indicated that the product is BMABr-0.5H2O. Essentially anhydrous
The synthesis of BMABr-hydrate was carried out as follows: ben-
zyl bromide (25 mL; 0.21 mol) was added drop wise (ca. 15 min) to
cold (ca. −10 ◦C) N,N,N-dimethylbenzylamine (30.0 mL; 0.20 mol)
under constant, vigorous stirring and N2 atmosphere, followed by
reaction during 2.5 h, at RT. Toluene was added, the solid product
filtered off, suspended several times in the same solvent and filtered
off. The product was dried under reduced pressure, in the presence
of P4O10. Yield 93%; white powder; mp = 172–173 ◦C; literature,
mp = 176–177 ◦C (Erling & Gary, 1967; Karoly & Gyermek, 1952)
and 175 ◦C (Busi, Lahtinen, Ropponen, Valkonen, & Rissanen, 2004).
1H NMR (DMSO-d6): ı (in ppm): 7.64–7.61 (m, 4H), 7.57–7.50 (m,
6H), 4.69 (s, 4H), 2.89 (s, 6H), where ı, m, and s stand for chemical
shift, multiplet, and singlet, respectively; 13C NMR (DMSO-d6): ı (in
ppm): 133.1, 130.3, 128.9, 128.0, 66.8, and 48.0. Both spectral data
are in agreement with the literature (Busi et al., 2004). Elemental
analysis (sample dried for additional 48 h, under reduced pressure,
at 40 ◦C, in the presence of P4O10): Anal. Calcd. for C16H20NBr: C,
62.7; H, 6.6; N, 4.6%. Found: C, 62.5; H, 6.8; N, 4.6%.
BMAF-hydrate was obtained as follows: a methanolic solution
of BMABr-hydrate (0.1 mol L−1, 1 L) was passed through a column
containing macroporous resin (Purolite SGA55OOH; 1.10 eq OH−/L;
170 mL). The completeness of the (Br−/OH−) ion-exchange was
assured by treating a sample of the eluent with AgNO3 solution,
acidified with nitric acid. The pH of the methanolic hydroxide solu-
tion was adjusted to ca. 7 (expanded-scale pH-paper) by adding
methanolic/HF solution. Methanol was removed by evaporation;
the pressure was gradually reduced from 200 to 10 mmHg at 50 ◦C,
then to 2 mmHg, at 60 ◦C for ca. 4 h. The solid residue was addi-
tionally dried under reduced pressure (3 mmHg) for 48 h at 40 ◦C,
in the presence of P4O10. White powder; mp = 132–133 ◦C. 1H NMR
(DMSO-d6): ı (in ppm): 7.66–7.63 (m, 4H), 7.57–7.48 (m, 6H), 4.72
(s, 4H), 2.89 (s, 6H); 13C NMR (DMSO-d6): ı (in ppm): 134.0, 130.8,
129.5, 129.0, 67.7, and 48.6.
fluoride, BMAF-0.1H2O, was obtained after ion-exchange (Br−
→
OH−), neutralization of BMAOH with HF, methanol evaporation,
and further drying.
3.2. Discussion of the side reactions occurring during the
syntheses of R4NF-hydrates
As mentioned above, TAAF–H2O and TBAF–3H2O are susceptible
to polymerization/crosslinking and Hofmann elimination, respec-
tively. The final stage in the synthesis of the former electrolyte
involves neutralization of the methanolic TAAOH solution with HF,
followed by evaporation of the alcohol. In order to avoid poly-
merization, (time consuming) solvent evaporation was carried out
under reduced pressure at RT, or at higher temperature in the pres-
ence of a free radical scavenger, propyl gallate (Casarano et al.,
2011). Almost anhydrous TBAF can be obtained by heating com-
mercial TBAF–3H2O at 40–45 ◦C, under reduced pressure for a long
time 48–147 h; the products were found to contain from 0.1 to 0.3
H2O/TBAF (Cox, Terpinski, & Lawrynowicz, 1984; Sharma & Fry,
1983). Freshly prepared almost anhydrous TBAF is not, however,
stable at RT, even in solution, and should be used immediately.
E.g., during its storage (as a solid) under reduced pressure, the con-
centration of bifluoride ion (HF2−, see Scheme 1 C), increased from
13 mol% to ca. 80 mol% after 4 days. On standing in CD2Cl2 solution,
the concentration of (HF2−) reached 61- and 100 mol% after 3 and

R. Casarano et al. / Carbohydrate Polymers 101 (2014) 444–450
447
Scheme 3. Proposed mechanism of side reactions that may occur via intra-molecular rearrangements of BMAF ylide, formed by the abstraction of H atom by (F−). This
scheme is based on a published work where the ylide was generated by benzyllithium (Erling & Gary, 1967).
12 h, respectively. In THF, ca. 75 mol% (HF2−) was produced after 8
days. (Sharma & Fry, 1983) Although BMAF is not subject to Hof-
mann elimination (for lack of ␤-hydrogen atoms) all these R4NF are
subject, in principle, to elimination by the ylide mechanism, see 1D
of Scheme 1.
We used 1H-NMR to probe the occurrence/extent of side reac-
tions. The spectrum of BMAF-0.1H2O shows a few additional peaks,
in comparison with that of BMABr-0.5H2O (cf. Fig. SM-1). Although
a detailed characterization of these side products is beyond the
scope of the present work, their concentration can be calculated
from peak areas. Assuming that the additional peaks were produced
by intra-molecular rearrangements of BMAF ylide, see Scheme 3,
some side products can be envisaged, namely, 1 (2.1, 2.2, 4.4, and
7.0–7.4 ppm), 2 (2.2, 2.9, 3.2, 4.1, and 6.9–7.4 ppm), and 3 (2.2,
2.3, 4.1, and 7.0–7.4 ppm); chemical shifts of (1–3) were calculated
by available on-line software (Advanced Chemistry Development,
2013).
Compounds 1–3 have in common the presence of an N,N-
dimethylamino group attached to a benzylic carbon atom. These
methyl groups should have very close-, or identical chemical shifts.
If the singlet at 2.08 ppm in the spectrum of BMAF-0.1H2O is
attributed to these methyl groups (part b of Fig. SM-1; relative
area = 0.06) then the molar ratio BMAF/elimination products is
0.94/0.06, i.e., the extent of side reactions is, by this criterion, ca.
6 mol%.
3.3. Evaluation of the thermal stability of R4NF-hydrates
For TBAF–3H2O, mass decreases of 17.1 and 55.4% are expected
due to the loss of water of hydration, and other volatiles (tributy-
lamine and 1-butene), as shown in reactions 1 and 2 of Scheme 1 C.
When TBAF·3H2O is heated under reduced pressure at 77 ◦C for
15 h, the mass loss observed is 53% (Sharma & Fry, 1983), showing
that the anhydrous electrolyte undergoes Hofmann elimination on
heating, in agreement with previous discussion.
Fig. 1. (A) TG/DTG and (B) isothermal TG curves obtained for BMAF.0.1H2O.
The following can be concluded from the TG/derivative thermo-
gravimetry (DTG) of BMAF-0.1H2O, Fig. 1a: (i) A mass loss of 1.5%
occurs when the T reached 143 ◦C due, in part, to loss of water of
hydration. (ii) Subsequently, the mass loss occurs in steps, from
143 to 175 ◦C (58.4%) and from 175 to 230 ◦C (40.0%); these ranges
correspond to the DTG curve maxima at 171 and 217 ◦C, respec-
tively. We seek an explanation for the appearance of the two peaks
from the TG results of two tetraalkylammonium hydroxides. Such
comparison is valid because F− and OH− anions are isoelectronics
(Kluge & Weston, 2005). Whereas the curve of tetramethylammo-
nium hydroxide pentahydrate (no ␤-hydrogens) shows peaks from
110 to 125 ◦C, and 125 to 130 ◦C, attributed to the ylide and SN2-
type decomposition mechanisms (Macomber, Boncella, Pivovar, &
Rau, 2008), that of ethyltrimethylammonium deuteroxide hydrate

448
R. Casarano et al. / Carbohydrate Polymers 101 (2014) 444–450
(␤-hydrogens present), shows mass loss from 30 to 68 ◦C, due
essentially to Hofmann elimination. Provided that the former qua-
ternary ammonium hydroxide is a model for BMAF-0.1H2O, we
attribute the two consecutive peaks to degradation by the same
mechanisms suggested.
Finally, the dependence of mass loss on time for BMAF-0.1H2O
is shown in Fig. 1b. A mass loss of 6% occurred after 18 h at 80 ◦C.
The relevance of this result is that little electrolyte degradation
occurs, even after its exposure to extreme conditions (high T and
long time). Considering that R4NF-hydrates are more stable in solu-
tion than in the solid state (the condition for the isothermal TG
experiment) this thermal stability contrasts very favorably with the
case of almost anhydrous TBAF, vide supra (Sharma & Fry, 1983).
3.4. Dissolution and acylation of cellulose in BMAF-0.1H2O/DMSO
MCC and cotton cellulose form clear solutions in BMAF-
0.1H2O/DMSO. Table 1 shows the results of acylation of both cellu-
loses, along with literature data for other R4NF-hydrates/DMSO.
The DS values obtained for the acylation in the former system
(Entries 1 to 6) are higher than those reported for other electrolytes
(Entries 7 and 8), although the DP of the cotton employed is more
than double that acylated in TBAF–3H2O (Ass et al., 2004; Casarano
et al., 2011).
3.4.1. A rationale for the observed efficiency of BMAF-0.1H2O
Considering the results of Table 1, the following question arises:
What are the possible reasons for this efficiency? Note that for
comparison, we employed conditions similar to those used for
(TBAF–3H2O) and (TAAF–H2O), i.e., they may have not been opti-
mized. Consequently, the differences in DS may have been larger.
In order to answer this question, and to understand the
cellulose/R4NF-hydrate/DMSO system we have carried out MD
simulations of the interactions between the components presents.
To our knowledge, this is the first time that MD simulations
have been employed to probe the interactions in cellulose/R4NF-
hydate/DMSO solutions. Two systems were simulated, each
containing 1100 DMSO molecules; one oligomer molecule, and
either 24 molecules of BMAF plus 3 molecules of water (BMAF-
0.1H2O) or 24 molecules of TBAF plus 72 of water (TBAF–3H2O).
Details of these calculations are given in the Calculations section of
SM. Table 2 summarizes the main results of these calculations.
Regarding the results of Table 2, the following is relevant:
Fig. 2. Snapshots from the simulation of the system oligomer/TBAF–3H2O/DMSO.
The color codes employed are: green (F−), red (oxygen), turquoise (carbon), blue
(nitrogen), yellow (sulfur), and white (hydrogen). All three snapshots were taken
from the same (simulation) frame. Part (A) shows the oligomer and the DMSO
present in its solvation layer (0.5 nm). Part (B) shows DMSO and TBAF molecules.
Part (C) shows all system components. The arrows in parts (B and C) show the simul-
taneous binding of (F−) to two hydroxyl groups (AGU of the oligomer), and water
molecules. (For interpretation of the references to color in this figure legend, the
reader is referred to the web version of the article.)
of TBAF–3H2O. The RDF of the center of mass, COM, of these
water molecules around the oligomer surface (Fig. SM-2) shows
the presence of two peaks, with maxima at 0.182 nm and
0.298 nm, respectively. The number of water molecules asso-
ciated with these peaks can be calculated from their areas.
This calculation has indicated that the first peak is due to ca.
1.2 molecules, presumably due to water molecules that are
hydrogen-bonded to the hydroxyl groups of the AGU. The sec-
ond (broader) peak contains 18.7 water molecules. These are
most certainly hydrogen-bonded to (F−) of TBAF–3H2O that
is located within the solvation layer of the oligomer surface,
giving anion hydration number of ca. 2 (an average hydration
number of 4.7 for (F−) has been deduced from MD simulations)
(Öhrn & Karlström, 2004). As shown from NMR data of the sys-
tem TBAF–water/MCC/DMSO, hydrogen bonding involves the
hydroxyl groups of the AGU, (F−) and water (Östlund, Lundberg,
Nordstierna, Holmberg, & Nydén, 2009).
(i) Calculation of the radial distribution function, RDF, of all atoms
present around the oligomer surface has indicated that the
extension of its solvation layer can be taken as equal to 0.5 nm;
this value has been employed throughout;
(ii) Interactions of the oligomer with (F−) and water are considered
together. The extension of the oligomer solvation layer by (F−),
0.122 to 0.196 nm, is not too far from the distances between
(F−) and water in crystals of hydrated inorganic electrolytes,
calculated from X-ray diffraction (0.162–0.185 nm) (Simonov
& Bukvetsky, 1978). On the other hand, the ratios (F−/AGU) are
0.47, and 0.71, for BMAF-0.1H2O and TBAF–3H2O respectively
(E2). Likewise, the number of hydrogen-bonds of the oligomer
OH groups and (F−) are 7, and 16, respectively (E3).
These differences do not necessarily mean that the (F−/AGU)
interactions are stronger for TBAF–3H2O. The reason is that
hydration of this anion leads to dramatic decrease in its basic-
ity/nucleophilicty (Landini, Maia, & Rampoldi, 1989; Pliego &
Pilo-Veloso, 2008), hence attenuates its interactions with the
hydroxyl groups of the AGU. Thus (E6 and E7) of Table 2 show
that the surface of the oligomer remains essentially anhy-
drous in the presence of BMAF-0.1H2O. On the other hand,
the oligomer is solvated by ca. 20 water molecules in case
(iii) Our calculations have shown the following details of the inter-
actions of these components. Fig. 2A shows a snapshot of the
oligomer with DMSO (TBAF and water molecules not shown);
part 2B shows the same snapshot in the presence of TBAF and
DMSO (water molecules not shown); whereas 2C shows all sys-
tem components. The arrows in parts 2B and 2C show that (F−)

R. Casarano et al. / Carbohydrate Polymers 101 (2014) 444–450
449
Table 1
Results of acylation of MCC, and cotton cellulose, in BMAF.0.1H2O/DMSO.a
Entry
Solventb
Cellulose type
Carboxylic anhydride
Ester DS
1
2
3
4
BMAF-0.1H2O/DMSO
BMAF-0.1H2O/DMSO
BMAF-0.1H2O/DMSO
BMAF-0.1H2O/DMSO
BMAF-0.1H2O/DMSO
BMAF-0.1H2O/DMSO
TAAF–H2O/DMSO
MCC
MCC
MCC
Cotton
Cotton
Cotton
MCC
Ethanoic
Butanoic
Hexanoic
Ethanoic
Butanoic
Hexanoic
Ethanoic
Ethanoic
2.6
2.1
2.5
2.4
2.0
2.5
1.9
1.6–2.3
5
6
7c
8d
TBAF–3H2O/DMSO
Cotton
a
Experimental conditions: 3 h; 80 ◦C; carboxylic anhydride/AGU molar ratio = 6; MCC, DPV = 175; cotton cellulose, DPv = 920.
The amounts employed were: cellulose, 0.5 g (3.08 mmol); DMSO, 40 mL (0.56 mol); BMAF-0.1H2O, 3.3 g (13.35 mmol); BMAF/AGU molar ratio = 4.3.
Literature data. Conditions: TAAF = 15.33 mmol; DMSO = 0.42–0.56 mol; TAAF/AGU molar ratio = 5.0; based on 0.5 g cellulose, 30–40 mL DMSO, and 3.3 g TAAF–H2O; 3 h,
b
c
80 ◦C; ethanoic anhydride/AGU molar ratio = 6; MCC, DPV = 175 (Casarano et al., 2011).
d
Literature data. Conditions: TBAF = 20.92 mmol; DMSO = 0.84 mol; TBAF/AGU molar ratio = 3.4; based on 1 g cellulose, 60 mL DMSO, and 6.6 g TBAF–3H2O; 3 h, 60–100 ◦C;
ethanoic anhydride/AGU molar ratio of 6; cotton cellulose, DPv = 440 (Ass et al., 2004).
Table 2
Results of MD simulations.a
Entry
Results obtained
BMAF-0.1H2O
TBAF–3H2O
Fluoride ion-oligomer interactions
E1
E2
Extension of the oligomer first solvation shell by (F−
)
From 0.122 to 0.190 nm; maximum at
0.146 nm
From 0.128 to 0.196 nm; maximum at
0.148 nm
Number and distribution of (F−) at the oligomer
surface
5.64 F− (0.47 F−/AGU)
8.52 F− (0.71 F−/AGU)
0.84 F−/AGU-C2-OH
0.72 F−/AGU-C3-OH
0.24 F−/AGU-C6-OH
1.2 F−/AGU-C2-OH
1.2 F−/AGU-C3-OH
0.6 F−/AGU-C6-OH
E3
Number of hydrogen-bonds per (simulation) frame
between the oligomer OH groups and F−
7
3
16
2
Electrolyte cation-oligomer interactions
E4
Extension of the oligomer solvation shell by
electrolyte-cation
Number of cations at oligomer surface
From 0.268 to 0.594 nm, with
maximum at 0.404 nm
From 0.294 to 0.564 nm, with
maximum at 0.386 nm
E5
4.68 BMA+ (0.39 BMA+/AGU)
2.04 TBA+ (0.17 TBA+/AGU)
Water-oligomer interactions
E6
E7
Number of H2O molecules at oligomer surface
Number of hydrogen-bonds per frame between the
None
None
20.04 H2O (1.67 H2O/AGU)
1
1
oligomer OH groups and water
DMSO-oligomer interactions
E8
Number of DMSO molecules at oligomer surface
91.2 DMSO (7.6 DMSO/AGU)
None
84 DMSO (7.0 DMSO/AGU)
System component interactions, excluding the oligomer
E9
Number of H-bonds between H2O and (F−) inside the
oligomer solvation shell (0.5 nm)b
15
2
E10
E11
Number of DMSO per (F−
)
7.3
12.1
Distance between F− and the N+ of the electrolyte
cation
From 0.284 to 0.486 nm, with
maximum at 0.338 nm
From 0.262 to 0.560 nm, with
maximum at 0.462 nm
From 0.314 to 0.534 nm, with
maximum at 0.374 nm
From 0.272 to 0.618 nm, with
maximum at 0.468 nm
E12
Distance between S of DMSO and (F−
)
a
(Species) numbers and distances listed are average values, calculated for the simulation time interval 15–70 ns.
This number is given per (simulation) frame, corresponding to 0.005 ns simulation time.
b
is simultaneously associated with the hydroxyl groups of the
AGU and water. That is, this anion is acting as a “bridge”, leading
to (HOH · · · F− · · · HO-AGU).
also with DMSO (E 10), at the expense of interactions of the
latter with the other components of the system. This is clearly
shown by E8 (more DMSO molecules are present at the surface
of the oligomer-BMAF-0.1H2O), and E12 (the distance (DMSO-
F−) is larger for TBAF–3H2O. The results reported in E11 can be
explained based on the volume of the two cations, vide supra.
For example, quantum mechanics calculations (Pliego & Pilo-
Veloso, 2008) have shown that the distance (F− · · · N+) increases
as a function of increasing the cation volume.
(iv) The results of E4 and E5 may be explained on the bases of dif-
ferences in the hydration of (F−) and volumes of the cations
(0.389 and 0.286 nm3, for TBA+ and BMA+, respectively). Thus
anion hydration leads to weaker anion-cation interactions in
TBAF–3H2O relative to that of BMAF-0.1H2O. Consequently, the
oligomer solvation layer of the former electrolyte is thinner and
contains a smaller number of cations, because it contains water
instead of the voluminous TBA+, E6;
In summary, the efficiency of R4NF/DMSO as solvents for cellu-
lose is due to, in part, the simultaneous hydrogen bonding of (F−)
to C2 OH and C3 OH, as shown by the arrow in parts (B and C) of
Fig. 2. Based on our calculations, the water in TBAF–3H2O probably
affects the electrolyte performance as a solvent for cellulose, due to
the solvation of (F−). This presumably affects the physical state of
dissolved cellulose, because of the resulting weaker interactions of
(F−) with the hydroxyl groups of AGU. In this regard, BMAF-0.1H2O
offers an advantage because of its much lower water content, and
(v) Explanation of E8 to E12 rests on two known facts: (a) the free
energy of transfer of (F−) from water (or protic solvents in gen-
eral) to DMSO is positive, i.e., unfavorable (Hefter, 1991, 1997),
(b) water–DMSO interactions are stronger than water–water
counterparts (Kingston & Symons, 1973; Mizuno, Imafuji,
Ochi, Ohta, & Maeda, 2000; Shashkov, Kiselev, Tioutiounnikov,
Kiselev, & Lesieur, 1999; Symons, 1986). Accordingly, water is
expected to interact strongly not only with the fluoride ion, but

450
R. Casarano et al. / Carbohydrate Polymers 101 (2014) 444–450
smaller volume. It is also important to take into account the ther-
mal stability of these R4NF-hydrates, and the deleterious effects of
water on the reaction outcome (hydrolysis of acyl chlorides and
anhydride, hydrolysis of produced ester). A corollary to the more
extensive interaction of BMAF with cellulose is that the biopolymer
in this system is probably more accessible, hence more reactive;
this agrees with the higher DS obtained. We plan to study fur-
ther this point, i.e., the dependence of product DS on biopolymer
accessibility, hence reactivity.
Chempath, S., Boncella, J. M., Pratt, L. R., Henson, N., & Pivovar, B. S. (2010). Den-
sity functional theory study of degradation of tetraalkylammonium hydroxides.
Journal of Physical Chemistry C, 114, 11977–11983.
Cox, D. P., Terpinski, J., & Lawrynowicz, W. (1984). Anhydrous tetrabutylammonium
fluoride: A mild but highly efficient source of nucleophilic fluoride ion. Journal
of Organic Chemistry, 49, 3216–3219.
Erling, G., Jr., & Gary, W. (1967). Carbanions. IX. Evidence for a duality of mechanism
in rearrangement of carbanions. Journal of the American Chemical Society, 89,
1852–1862.
Fieser, L. F. (1967). Fieser and Fieser’s reagents for organic synthesis. New York: Wiley.
Hefter, G. T. (1991). Fluoride solvation. The case of the missing ion. Pure and Applied
Chemistry, 63, 1749–1758.
Hefter, G. T. (1997). Solvation of fluoride ions 4. Gibbs energies of transfer from water
to aqueous dimethylsulphoxide, propanone and 2-methylpropan-2-ol mixtures.
Journal of Molecular Liquids, 73-74, 267–276.
4. Conclusions
Heinze, T., & Koehler, S. (2010). Dimethyl sulfoxide and ammonium fluorides – Novel
cellulose solvents. ACS Symposium Series (Vol. 1033) (Cellulose Solvents).
Heinze, T., Lincke, T., Fenn, D., & Koschella, A. (2008). Eficient allylation of cellulose
in dimethyl sulfoxide/tetrabutylammonium fluoride trihydrate. Polymer Bulletin
(Berlin), 61, 1–9.
Heinze, T., Wanga, Y., Koschella, A., Sullob, A., & Foster, T. J. (2012). Mixed 3-mono-
O-alkylcellulose: Synthesis, structure characterization and thermal properties.
Carbohydrate Polymers, 90, 380–386.
Karoly, N., & Gyermek, G. (1952). Synthesis of compounds with ganglion-blocking
effects. I. Synthesis of some simple quaternary ammonium derivatives. Acta
Chimica Academiae Scientiarum Hungaricae, 2, 95–101.
Kingston, B., & Symons, M. C. R. (1973). Solvation spectra. Part 44.-Nuclear magnetic
resonance study of binary solvent mixtures: Water structural effects. Journal
of the Chemical Society, Faraday Transactions 2: Molecular and Chemical Physics,
2(69), 978–992.
Kluge, S., & Weston, J. (2005). Can a hydroxide ligand trigger a change in the coor-
dination number of magnesium ions in biological systems? Biochemistry, 44,
4877–4885.
Koehler, S., & Heinze, T. (2007). New solvents for cellulose: Dimethyl sulfox-
ide/ammonium fluorides. Macromolecular Bioscience, 7, 307–314.
Landini, D., Maia, A., & Rampoldi, A. (1989). Dramatic effect of the specific solvation
on the reactivity of quaternary ammonium fluorides and poly(hydrogen fluo-
rides), (HF), in media of low polarity. Journal of Organic Chemistry, 54, 328–332.
Leibert, T. F., & Heinze, T. (2005). Tailored cellulose esters: Synthesis and structure
determination. Biomacromolecules, 6, 333–340.
Macomber, C. S., Boncella, J. M., Pivovar, B. S., & Rau, J. A. (2008). Decomposition
pathways of an alkaline fuel cell membrane material component via evolved
gas analysis. Journal of Thermal Analysis and Calorimetry, 93, 225–229.
Makosza, M., & Chesnokov, A. A. (2003). Study of the mechanism of base induced
dehydrobromination of trans-␤-bromostyrene. Tetrahedron, 59, 1995–2000.
Mizuno, K., Imafuji, S., Ochi, T., Ohta, T., & Maeda, S. (2000). Hydration of the CH
groups in dimethyl sulfoxide probed by NMR and IR. Journal of Physical Chemistry
B, 104, 11001–11005.
A new quaternary ammonium fluoride-hydrate, BMAF-0.1H2O,
was synthesized by a simple procedure. The thermal stability of
this electrolyte is satisfactory; it dissolves MCC and fibrous cel-
lulose easily; the DS of the esters obtained are comparable, or
higher than those obtained with TAAF–H2O and TBAF–3H2O under
comparable experimental conditions. The reasons for the relative
efficiency of BMAF-0.1H2O have been deduced from the results
of MD simulations, available for the first time for these solvent
systems. These have shown that the (F−) of TBAF–3H2O is sol-
vated by the water present leading, presumably, to a decrease in
its basicity and hydrogen-bonding ability with the hydroxyl groups
of the AGU (lower biopolymer accessibility, hence reactivity). The
introduction of BMAF-0.1H2O is welcomed not only for cellulose
dissolution/derivatization, but also for organic synthesis in general.
Acknowledgments
We thank the FAPESP (State of São Paulo Research Founda-
tion) for financial support and a postdoctoral fellowship to R.
Casarano; the CNPq (National Council for Scientific and Techno-
logical Research) for a productivity fellowship to O. A. El Seoud;
Mariana C. Lucats and Cezar Guizzo for their help. This work has
been carried out within the frame of INCT-Catálise.
Appendix A. Supplementary data
Öhrn, A., & Karlström, G. (2004). A combined quantum chemical statistical mechan-
ical simulation of the hydration of Li+, Na+, F−, and Cl−. Journal of Physical
Chemistry B, 108, 8452.
Östlund, Å., Lundberg, D., Nordstierna, L., Holmberg, K., & Nydén, M. (2009). Disso-
lution and gelation of cellulose in TBAF/DMSO solutions: The roles of fluoride
ions and water. Biomacromolecules, 10, 2401–2407.
Supplementary data associated with this article can be
found, in the online version, at http://dx.doi.org/10.1016/
j.carbpol.2013.09.043.
Pliego, J. R., & Pilo-Veloso, D. (2008). Effects of ion-pairing and hydration on the
SNAr reaction of the F- with p-chlorobenzonitrile in aprotic solvents. Physical
Chemistry Chemical Physics, 10, 1118–1124.
References
Sharma, R. K., & Fry, J. L. (1983). Instability of anhydrous tetra-n-alkylammonium
fluorides. Journal of Organic Chemistry, 48, 2112–2114.
Shashkov, S. N., Kiselev, M. A., Tioutiounnikov, S. N., Kiselev, A. M., & Lesieur, P. (1999).
The study of DMSO/water and DPPC/DMSO/water system by means of the X-
ray, neutron small-angle scattering, calorimetry and IR spectroscopy. Physica B:
Condensed Matter (Amsterdam), 271, 184–191.
Simonov, V. I., & Bukvetsky, B. V. (1978). On the geometry of O H · · · F hydrogen
bonds in metal fluoride hydrates. Acta Crystallographica, Section B: Structural
Science, 34, 355–358.
Sun, H. R., & DiMagno, S. G. (2005). Anhydrous tetrabutylammonium fluoride. Journal
of the American Chemical Society, 127, 2050–2051.
Sun, H. R., & DiMagno, S. G. (2006). Room-temperature nucleophilic aromatic fluori-
nation: Experimental and theoretical studies. Angewandte Chemie, International
Edition in English, 45, 2720–2725.
Symons, M. C. R. (1986). Infrared and NMR studies of probes in binary solvent sys-
tems. Pure and Applied Chemistry, 58, 1121–1132.
Tutaj, B., Gonzalez-Perez, A., Czapkiewicz, J., Del Castillo, J. L., & Rodriguez, J. R.
(2001). Apparent molar volumes of benzyltrimethylammonium bromide and its
homologs in aqueous solution at 15, 25, and 35 ◦C. Journal of Solution Chemistry,
30, 1101–1109.
Xu, D., & Edgar, K. J. (2012). TBAF and cellulose esters: Unexpected deacylation with
unexpected regioselectivity. Biomacromolecules, 13, 299–303.
Advanced Chemistry Development, I. A. L. (2013). NMR Predictors V. 11.01. Scifinder
Scholar. http://scifinder.cas.org [accessed 17.11.12]
Albanese, D., Landini, D., & Penso, M. (1998). Hydrated tetrabutylammonium fluoride
as a powerful nucleophilic fluorinitaing agent. Journal of Organic Chemistry, 63,
9558–9589.
Armagero, W. L. F, & Chai, C. L. L. (2003). Purification of laboratory chemicals (5th ed.).
New York: Elsevier.
Ass, B. A. P., Frollini, E., & Heinze, T. (2004). Studies on the homogeneous acety-
lation of cellulose in the novel solvent dimethyl sulfoxide/tetrabutylammonium
fluoride trihydrate. Macromolecular Bioscience, 4, 1008–1013.
ASTM Standard D1795-96. (2007). Standard test methods for intrinsic
viscosity of cellulose. West Conshohocken, PA: ASTM International.
http://dx.doi.org/10.1520/D1795-96
ASTM Standard D871-96. (2004). Standard test methods of testing
cellulose
acetate.
West
Conshohocken,
PA:
ASTM
International.
http://dx.doi.org/10.1520/D0871-96
Busi, S., Lahtinen, M., Ropponen, J., Valkonen, J., & Rissanen, K. (2004). Synthesis,
characterization and thermal properties of new aromatic quaternary ammo-
nium bromides: Precursors for ionic liquids and complexation studies. Journal
of Solid State Chemistry, 177, 3757–3767.
Casarano, R., & El Seoud, O.A. (2013). Solvent system, a process for producing a
solvent system and its use. Patent 001097-9[pending]. BR 10.
Casarano, R., Nawaz, H., Possidonio, S., da Silva, V. C., & El Seoud, O. A. (2011). A
convenient solvent system for cellulose dissolution and derivatization: Mech-
anistic aspects of the acylation of the biopolymer in tetraallylammonium
fluoride/dimethyl. Carbohydrate Polymers, 83, 1285–1292.
Xu, D., Li, B., Tate, C., & Edgar, K. J. (2011). Studies on regioselective acylation of
cellulose with bulky acid chlorides. Cellulose, 18, 405–419.
Zheng, X., Gandour, R. D., & Edgar, K. J. (2013). Probing the mechanism of TBAF-
catalyzed deacylation of cellulose esters. Biomacromolecules, 14, 1388–1394.