Journal of Membrane Science 524 (2017) 449–459

Contents lists available at ScienceDirect

Journal of Membrane Science

journal homepage: www.elsevier.com/locate/memsci

Bio-based membranes for ethyl tert-butyl ether (ETBE) bio-fuel purification MARK
by pervaporation
Faten Hassan Hassan Abdellatif, Jérôme Babin, Carole Arnal-Herault, Cécile Nouvel,
⁎
Jean-Luc Six, Anne Jonquieres
Laboratoire de Chimie Physique Macromoléculaire, LCPM UMR CNRS−Université de Lorraine 7375, ENSIC, 1 rue Grandville, BP 20451, 54 001 Nancy
Cedex, France

A R T I C L E I N F O A BS T RAC T

Keywords: Ethyl tert-butyl ether (ETBE) is a fuel octane enhancer considered as a major bio-fuel in Europe. Blended with
Polymer membranes gasoline fuels, ETBE improves fuel combustion and reduces toxic hydrocarbon emissions. ETBE industrial
Cellulose acetate synthesis leads to an azeotropic mixture EtOH/ETBE (20/80 wt%), which cannot be separated by simple
Polymer modification distillation. According to former works, cellulose acetate (CA) membranes were extremely selective with a
Pervaporation
permeate EtOH content of 100% but their flux was too low for ETBE purification by the pervaporation (PV)
Bio-fuel purification
membrane process. The affinity of the bio-based polylactide (PLA) for ethanol has also been reported for
poly(vinyl pyrrolidone)/PLA blends for this application. In this work, new bio-based membranes were obtained
by grafting CA with controlled PLA amounts by a “grafting onto” strategy. The CA-g-PLA copolymers were
characterized by ATR-FTIR, 1H NMR, DSC and SAXS. Their membrane properties were investigated in terms of
structure-morphology-property relationships for the sorption and pervaporation of the targeted EtOH/ETBE
mixture. PLA grafting onto CA strongly improved the flux (×12) while the ethanol permeate content remained in
the very high range (C′ > 90 wt%) for ETBE purification by pervaporation.

1. Introduction butyl ether (ETBE) is an attractive bio-ether used as oxygenated fuel

Membrane separation processes have taken their place among the
most important industrial and research topics and they are widely used
in gas and liquid separations, waste water treatment and water
desalination. Compared to conventional processes such as liquid-liquid
extraction or distillation, membrane processes offer many advantages
like modularity, energy saving, mild operating conditions and lower
environmental impact [1–3]. In all of these processes, the membrane is
a key factor and its optimization is essential.
On the other hand, the formulation of fuels with lower environ-
mental impact and the production of new bio-fuels are two main issues
for sustainable development but their production usually requires
highly energy-intensive separation steps. Therefore, the search for
new cost-effective separation techniques is becoming critical for the
bio-fuel industry. Compared to other separation processes (i.e. azeo-
tropic distillation, liquid-liquid extraction or extractive distillation), the
pervaporation (PV) membrane process, alone or in hybrid processes,
could offer a very good alternative and greatly contribute to the future
development of bio-fuels [4–6].
Among the major bio-fuels in the European Union (EU), ethyl tert-
additive to improve fuel combustion and reduce toxic hydrocarbon
emissions. This ether has also contributed to limit the cars environ-
mental impact by replacing the carcinogenic lead derivatives formerly
used in the EU. Nevertheless, during the ETBE industrial synthesis, the
separation of the azeotropic mixture EtOH/ETBE (20/80 wt%) in-
volves a highly energy-intensive ternary distillation process [7]. This
separation process could be advantageously replaced by a PV or PV/
distillation hybrid process with an efficient organoselective membrane
for ethanol removal [4,5].
ETBE purification by PV has been relatively rarely investigated so
far and the main related works have been reviewed in recent papers [7–
9]. In the following brief literature review, the membrane fluxes are
normalized for a reference thickness of 5 µm for comparison. Most of
the related studies were based on polymeric membranes i.e. cellulosic
membranes, segmented copolymers, and poly(vinyl pyrrolidone)
(PVP)-based copolymers and blends. In particular, a cellulose acetate
(CA) membrane had nearly infinite separation factor with a permeate
ethanol content of 100% but its total normalized flux of 0.08 kg/h m2
was very low at 40 °C [10]. Several approaches have been reported for
improving CA flux while maintaining its selectivity at very high level.

⁎
Corresponding author.
E-mail address: anne.jonquieres@univ-lorraine.fr (A. Jonquieres).

http://dx.doi.org/10.1016/j.memsci.2016.11.078
Received 25 March 2016; Received in revised form 7 November 2016; Accepted 26 November 2016
Available online 28 November 2016
0376-7388/ © 2016 Elsevier B.V. All rights reserved.
F. Hassan Hassan Abdellatif et al.
Permeability and selectivity generally vary in opposite ways and the
obtained results were very encouraging [10]. Blending of cellulosic
esters improved the total flux in the range of 0.6–3 kg/h m2 with
permeate ethanol contents of 90.7–96.7 wt% at 40 °C [11]. Cellulosic
blends or semi-interpenetrated networks with poly(meth)acrylic (co)
polymers also offered improved membrane properties for the targeted
separation [1,10,12,13]. In a former work of our team, CA was
modified with short or long PEO-containing polymethacrylate grafts
by a "grafting from" strategy to assess the influence of copolymer
architecture on the membrane properties for ETBE purification. The
best membrane properties (normalized flux of 0.87 kg/h m2, C′=94 wt
% at 50 °C) were obtained for the copolymer with the highest amount
of short grafts [8]. We have also recently reported CA grafting by
different ionic liquids for this application [14]. In the last case, the
membrane flux greatly increased with the hydrogen bonding acceptor
ability of the grafted ionic liquid and the best membrane properties
(normalized flux of 0.182 kg/h m2 and permeate ethanol content of
100% at 50 °C) were obtained with an ammonium ionic liquid.
In other respect, polylactide (PLA) is a well-known bio-based
polyester derived from renewable sources such as corn starch [15].
This biodegradable and biocompatible polymer has a wide range of
medical applications in drug delivery, tissue engineering and orthope-
dic devices investigated by some of us [16–18]. Thanks to its unique
features, PLA is also widely used in packaging applications where its
gas permeation properties are strongly influenced by its crystallinity
[19]. O2 permeability through PLA/polypropylene (PP) blends and clay
nanocomposites was strongly dependent on film composition and PLA
incorporation to PP-based films enabled to control their barrier
properties with interesting prospects for packaging applications [20].
Beyond the scope of this work, PLA has also been reported for gas
permeation of small molecules of great interest like He, O2, CO2 and
CH4.
However, the separation of liquid mixtures by PLA-based mem-
branes has been very rarely reported so far [9,15,21]. Ethanol
permeation through PLA films has been investigated for bottle packa-
ging applications [15]. In the latter work, the PV of aqueous solutions
containing 0–50 wt% of ethanol showed that water permeability was
much higher than that of ethanol. The preferential water permeation
was mainly ascribed to the much higher diffusion coefficient of water
(×1000) compared to ethanol. Even more closely related to this work,
organoselective membranes made of PLA/PVP blends have also been
reported for the PV separation of two azeotropic mixtures (ethanol/
cyclohexane and ethanol/ETBE) [9,21]. The latter works clearly
showed the interesting features of PLA for removing ethanol from
liquid organic mixtures. For ETBE purification, the best membrane
properties corresponded to a very high flux of ca. 2.7 kg/h m2 and to an
ethanol permeate content of 80 wt% at 30 °C for a PLA(97 wt
%)/PVP(3 wt%) blend [9].
In this work, we are considering new bio-based membrane materi-
als combining the unique features of CA and PLA in grafted copoly-
mers. Apart from PLA grafting of cellulose by polycondensation of
lactic acid affording very small grafts [22], PLA grafting of cellulose and
cellulose derivatives was achieved by ring-opening polymerization
(ROP) of lactide from the polysaccharide hydroxide side groups in
homogeneous conditions in aprotic dipolar solvents (DMAc/LiCl or
DMSO) or in ionic liquids (1-allyl-3-methylimidazolium chloride
AmimCl and 1-butyl-3-methylimidazolium chloride) [22–30]. A re-
lated strategy was also described for obtaining CA-g-PLA copolymers
by ‘‘one pot’’ cellulose acetylation followed by lactide ROP in AmimCl
with 4-dimethylaminopridine (DMAP) catalyst [28]. Cellulose nano-
crystals (CNC) and nanowhiskers (CNW) were also grafted by PLA
surface-initiated ROP or melt polycondensation of lactic acid for
producing PLA-based nanocomposite films with improved mechanical
properties [31–34]. Biodegradable cellulose-based fibers with im-
proved surface properties were obtained by in situ graft polymerization
of L-lactide onto cellulose by reactive extrusion [30]. Another reactive
450
Journal of Membrane Science 524 (2017) 449–459
processing method was successfully developed for preparing CA-g-PLA
for drug delivery [29]. To the best of our knowledge, the most efficient
methods reported so far for PLA grafting of cellulose, cellulose
derivatives and nanocelluloses were all based on "grafting from"
strategies using ring opening polymerization (ROP).
For the first time, this paper reports CA modification by a simple
“grafting onto” strategy of low molecular weight PLA oligomers by
"click" chemistry based on the copper(I) catalyzed 1,3-dipolar cycload-
dition of azides and alkynes (CuAAC). The low molecular weight for the
PLA grafts was chosen to ensure the best membrane properties
according to the results obtained in our former work on CA grafted
with PEO-based polymethacrylates [8]. The new grafting strategy
enabled to precisely control the number and the length of the PLA
grafts in the membrane materials. The chemical structure and mor-
phology of the grafted cellulosic copolymers were investigated by ATR-
FTIR, 1H NMR, DSC and SAXS. The sorption and permeability
properties of the new bio-based membranes were then assessed for
the separation of azeotropic mixture EtOH/ETBE (20/80 wt%) by the
PV sustainable process. The influence of the PLA content on the
membrane properties was analyzed on the basis of structure-properties
relationships revealing the key features of the new bio-based mem-
branes for ETBE purification by PV.
2. Material and methods
2.1. Materials
CA (CA, acetyl 39.7 wt%, MW=50,000 g/mol), 6-bromohexanoic
acid (97%), stannous octoate (SnOct2, 92.5–100.0%), copper bromide
(CuBr, 99.99%), propargyl alcohol (99%), 4-dimethylaminopyridine
(DMAP, ≥99%), sodium azide (NaN3 ≥99.99%) and N,N,N’,N’’,N’’’-
pentamethyldiethylenetetramine (PMDETA, 99%) were purchased
from Sigma-Aldrich. 1H NMR analysis confirmed the acetyl content
of CA, corresponding to 2.46 of acetyl groups and 0.54 of hydroxyl
groups per glycosidic ring. The degrees of substitution (DS) of the side
groups were commonly defined as the number of side groups per
glycosidic ring. 1-(3-Dimethylaminopropyl)-3-ethylcarbodiimide hy-
drochloride (EDC.HCl, > 98%) and dimethyl sulfoxide (DMSO,
99.7%) were purchased from TCI Company and Fisher Scientific,
respectively. These reagents were used as received without further
purification. D,L-Lactide was purchased from Lancaster, recrystallized
in dry toluene twice and dried under vacuum before using. Toluene was
refluxed, distilled over calcium hydride (CaH2) and then added in the
reaction media via cannula under nitrogen atmosphere. The solvents
dichloromethane (DCM, ≥99.9%) and tetrahydrofuran (THF, 99.7%)
were dried over activated molecular sieves before using. To avoid
contamination by atmospheric moisture, all reagents and solvents were
stored under nitrogen atmosphere.
2.2. CA functionalization and grafting
2.2.1. Modification of CA with azide side groups
In three necks round bottom reactor, 15 g of virgin CA (correspond-
ing to 56.54 mmol of glycosidic rings and 30.53 of hydroxyl groups)
were dissolved in 300 mL of dry THF under argon flow and vigorous
stirring at room temperature. A solution of 4 g (25.48 mmol) of 6-
azidohexanoic acid (see Supporting information) and 0.373 g
(2.29 mmol) of DMAP in 150 mL of dry DCM were added to the CA
solution. After cooling the mixture at 5 °C, 4.84 g (25.24 mmol) of
EDC.HCl were added to the reaction medium and stirred for 48 h at
room temperature. The crude polymer was precipitated from 2 L of
ethanol (96%), washed twice in ethanol, filtered and dried under
vacuum at 60 °C overnight. The polymer was obtained as white solid
fibers with a yield of 88%. The degree of substitution of CA by azide
groups DSazido=0.39 was estimated by 1H NMR in CDCl3 using
equation (1):
F. Hassan Hassan Abdellatif et al.
2DSAzido A1.4
=
3DSAcetyl +2DSAzido A1.8−2.2 + A2.35 (1)
where A1.4 was the area of the peak at 1.4 ppm corresponding to 2
methylene protons Hd in the anchor, A1.8–2.2 was the area of the
peaks in the range of 1.8–2.2 ppm corresponding to 3 protons Ha of the
acetyl side groups, and A2.35 was the area of the peak at 2.35 ppm
corresponding to 2 methylene protons Hb in the anchor (see main text).
2.2.2. Grafting of the azido CA with α-alkyne PLA1640 by CuAAC
"click" chemistry
As way of example, the following procedure describes the synthesis
of CA grafted with 40.5 wt% of PLA. In a Schlenk tube, 1 g of azido-
functionalized CA (corresponding to 3.13 mmol of glycosidic rings and
1.22 mmol of azide groups) and 0.788 g (0.4807 mmol) of a functional
polylactide oligomer containing a terminal alkyne group (α-Alkyne
PLA1640, see Supporting information) were dissolved in 30 mL of
DMSO under argon atmosphere. Consequently, a theoretical degree
of substitution DStheo
PLA = 0.156 was targeted for the PLA grafts. After
complete dissolving of the two polymers, the solution was degassed
with argon for several times. 68.9 mg (0.4807 mmol) of copper
bromide and 2 mg of copper powder dendritic “Cu°” were added under
argon flow to the reaction mixture. The tube was sealed under argon
atmosphere and stirred for 24 h at room temperature. After reaction
completion, the grafted CA was obtained by precipitation from cold
ethanol in presence of 10 eq. of PMDETA as copper ligand. This
precipitation step ensured the removal of any unreacted PLA oligomer,
which was well soluble in cold ethanol because of its particularly low
molecular weight. The precipitated fibers were washed twice in ethanol,
filtered and dried at 60 °C under vacuum overnight. All grafted
polymers were characterized by ATR-FTIR and 1H NMR (see main
text). Experimental degree of substitution for the PLA grafts (DSexpPLA
= 0.135) was calculated from equation (2).
exp A7.6
DSPLA =
A4.8 (2)
where A7.6 was the area of the peak at 7.6 ppm corresponding to the
proton Hg of the triazole ring formed by CuAAC "click" chemistry and
A4.8 was the area of the peak at 4.8 ppm corresponding to the proton
H2 belonging to the glycosidic rings.
2.2.3. Physicochemical characterization
FT-IR spectra were recorded for thin polymeric films on a Bruker
Tensor 127 FTIR spectrometer equipped with liquid nitrogen cooled
MCT detector. The standard sample cell in the FT-IR spectrometer was
a Pike Miracle single-bounce attenuated total reflectance (ATR) cell
equipped with a Ge crystal.
1
H NMR spectra were recorded on a Bruker Avance 300 spectro-
meter operating at 300 MHz. All 1H NMR experiments used polymer
solutions in CDCl3. The chemical shifts were referenced to TMS and
were calculated using the residual isotopic impurities of the deuterated
solvent. The peak integrations were used to calculate the degrees of
substitution of the different CA side groups or polymer grafts.
The number and weight average molecular weights were estimated
by size exclusion chromatography (SEC) using multi-angle laser light
scattering detector (MALLS; Mini Dawn Treos, Wyatt), differential
refractometer detector (OPTILab Rex, Wyatt), HPLC pump (Waters
515), degazer AF (Waters In-Line), and three PLgel 5 µl (105, 103, and
100 Å) columns (300×7.5 mm2). THF was used as eluent at 40 °C with
an elution rate of 0.7 mL per min. Polymer solutions (30 mg/mL) were
prepared using HPLC-grade THF and were filtered through PTFE
membranes (Alltech, average pore diameter: 0.2 µm) prior injection. A
refractive index increment (dn/dC) of 0.054 was used for PLA [35].
Thermal analysis was performed by standard Differential Scanning
Calorimetry (DSC) using a TA Instruments DSC Q2000. Polymer
samples of ca. 10 mg were used for analysis under a continuous flow
451
Journal of Membrane Science 524 (2017) 449–459
of nitrogen. Two cycles of measurements from 0 to 210 °C were
performed with heating and cooling rates of 10 °C/min. The thermal
transitions corresponding to the first and second heating scans differed
significantly and showed the strong influence of thermal history. The
data reported in this work corresponded to the glass transition
temperatures for the second heating scan for better comparison.
The nano-scale morphology of virgin CA and CA grafted with
different PLA contents was determined by means of small X-ray
scattering using polymer membranes with thicknesses of ca. 200 µm,
which were obtained in the same way as for sorption experiments.
Small angle X-ray scattering (SAXS) experiments were carried out at
room temperature, without any additional thermal treatment, using
synchrotron radiation at the European Synchrotron Radiation Facility
(ESRF, Grenoble, France). The SAXS experiments were performed on
the BM2-D2AM beamline at an incident energy of 16 keV. A 2D X-ray
camera was used (Roper Scientific), and the data were corrected from
the dark image, normalization with flat field and tapper camera
distortion. Finally, radial average around the incident beam center
were calculated. Silver behenate was used for the channel-q calibration.
The background (empty cell) was subtracted in the radial average
profiles, after the measurement of the attenuation coefficients.
2.3. Membrane preparation for pervaporation and sorption
experiments
The different polymer materials were dissolved in DMF (pure for
synthesis) to obtain a polymer concentration of 2.5% w/v. The
solutions were then cast on a PTFE mold. They were carefully removed
from the mold after DMF evaporation at 45 °C. Most of the DMF
solvent was evaporated at 45 °C during the first stage of membrane
preparation. In the final stage, the membranes were dried under
vacuum overnight at 60 °C to remove any residual DMF. DSC and 1H
NMR experiments confirmed that there was no DMF in the membranes
after their vacuum drying at 60 °C and that the drying procedure was
suitable for these relatively thin materials. The pervaporation mem-
brane thicknesses ranged from 60 to 70 µm and the difference in
thickness between two membrane points was less than 6 µm.
The procedure for sorption membrane casting was similar as that
for pervaporation membranes except for polymer concentration, which
was increased to 5% w/v in order to obtain membrane thicknesses of
ca. 200 µm. The sorption membranes were then carefully dried at 60 °C
under vacuum for 24 h before sorption experiments.
2.4. Sorption experiments
For sorption experiments, the dried polymer membranes were
weighted and then immersed in the azeotropic mixture EtOH/ETBE
(20/80 wt%) in hermetically closed bottles. These bottles were kept at
30 °C in a thermostated oven. After one week, the membranes were
taken out the bottles and quickly wiped with paper tissue and weighted
in a closed tare bottle. The membranes were regularly weighted until
they reached constant swelling weight. At the end of each sorption
experiment, the membranes were dried under vacuum and weighted
again. The results obtained with the different sorption membranes
showed that no polymer dissolution had occurred during the sorption
experiments. The total swelling, S (wt%), was calculated from equation
(3):
wS − wD
S= × 100%
wD (3)
where wS and wD were the membrane weights after and before
sorption experiment. The average experimental error for the total
swelling S was ± 0.2 wt%.
For determining the composition of the absorbed mixture, the
swollen membranes at equilibrium were soaked for 1 night at ambient
temperature into 10 mL of diethylether, which is a non solvent for the
F. Hassan Hassan Abdellatif et al. Journal of Membrane Science 524 (2017) 449–459

H OR H H
H O H OR
O RO
RO O O
H OR H O
n
H H H N3
OR + OH
O

R=H Or 6-Azido-hexanoic acid

acetate

THF, DCM,
DMAP, EDCl, 48 hours, R T

H OR H H
H O H OR
O RO
RO O
H OR H O
n
H H H
OR

O O

N3
R= H Or
Or
acetate azide anchor

0
C uB r, C u , D M S O O OH
O
24h, RT m
O
P LA

H OR H H
H O H OR
O RO
RO O
H OR H O
n
H H H
OR

O O

N3
R = H Or Or
Or
acetate azide anchor

N O
O N
O OH
N O

cellulose acetate-g-P LA O m

Fig. 1. General reaction scheme for the PLA grafting onto CA by "click" chemistry.

membrane polymers and a good solvent for both ethanol and ETBE. thermal conductivity detector and hydrogen as carrier gas. The ethanol
The desorption solution was then analyzed by gas chromatography weight fraction CSEtOH in the desorbed mixture was calculated from the
using a Shimadzu GC-8A chromatograph with a Porapak Q column, a area of the EtOH and ETBE peaks on the basis of a former calibration.

452
F. Hassan Hassan Abdellatif et al.
Fig. 2. ATR-FTIR spectra for a) the α-alkyne PLA, b) azido-CA and c) the grafted
copolymer CA-g-40.5 PLA.
The average experimental error for the ethanol weight fraction CSEtOH
was ± 0.005.
Sorption separation factors were then calculated from Eq. (4) where
CSEtOH and C were the ethanol weight fractions in the sorption mixture
and the azeotropic mixture EtOH/ETBE, respectively. The symbol for
the sorption separation factor was that recently proposed by Baker
et al.[3].
S known pre-requisite to introduce azide or alkyne side groups for
CEtOH C
βS = S
/ CuAAC "click" chemistry. This functionalization is usually achieved
1 − CEtOH 1−C
2.5. Pervaporation experiments
Pervaporation experiment was performed for the azeotropic mix-
ture ETOH/ETBE at 50 °C using a pervaporation set-up described
elsewhere [8]. A temperature of 50 °C has already been reported
several times for ETBE purification by pervaporation by us [7,8] and
others [13,36]. This temperature ensures high productivity and re-
mains fully compatible with the temperature range involved in the
ETBE industrial synthesis process [37]. The cell downstream pressure
was maintained at less than 0.04 KPa using an Alcatel vacuum pump.
The permeate was collected continuously after condensation by liquid
nitrogen using two parallel traps. The permeate flux was calculated
using equation (5):
wp
Permeate flux =
Δt × A
where wp was the permeate sample weight collected during a permea-
tion time Δt and A was the active membrane area.
To make a comparison between different membranes with close but
non-equal thicknesses, normalized fluxes, Jnormalized, were reported
for a reference thickness of 5 µm (Eq. (6)). This reference thickness was
chosen because it is easily reached for dense polymer layers on top of
asymmetric membranes and it allowed a straightforward comparison
with the separation properties for other membranes reported for ETBE
purification.
Membrane thickness
Jnormalized = Permeate flux
5 (6)
The permeate ethanol weight fraction, C', was determined by gas
chromatography in the same conditions as for sorption experiments.
Journal of Membrane Science 524 (2017) 449–459
The pervaporation separation factor, βPV, was calculated by analogy
with the sorption separation factor according to Eq. (7). The experi-
mental errors were less than 5% for the total fluxes and ± 0.005 for the
ethanol weight fractions.
′
CEtOH C
βPV = ′
/
1 − CEtOH 1−C (7)
Ethanol and ETBE permeability data were calculated by a fairly
complex procedure allowing to estimate the pervaporation driving
force for the separation of the azeotropic mixture EtOH/ETBE from
thermodynamic models [38].
3. Results and discussion
3.1. Synthesis and characterization of CA grafted with PLA
Over the last few years, CuAAC "click" chemistry has been widely
used as a new way of modifying cellulose and cellulosic derivatives to
produce bio-based materials for a wide range of applications [38] e.g. i)
cross-linked networks for structural materials and hydrogels [39–41],
ii) block and graft copolymers [42–45], and iii) dendronised celluloses
[46–49]. In this work, the CA-g-PLA membrane materials were
obtained in two steps by a "grafting onto” strategy based on CuAAC
"click" chemistry (Fig. 1). In a first step, CA was chemically modified to
introduce azide side groups. The second step consisted of simple
grafting with a functional PLA containing a complementary alkyne
terminal group by CuAAC "click" chemistry.
3.1.1. Synthesis and characterization of azido CA
Pre-click modification of cellulose and its derivatives is a well
(4)
after activating the hydroxyl side groups e.g. by tosylation [39,50]. In
the present work, CA with azide side groups was obtained in a single
step through an activated esterification of the hydroxyl side groups
(DSOH= 0.54) with 6-azido hexanoic acid by adapting the mild
conditions reported by Neises and Steglich (Fig. 1) [51]. EDC.HCl
and DMAP were used as coupling agent and catalyst, respectively, and
a theoretical degree of substitution of the azido groups equal to 0.45
was targeted for azido CA.
ATR-FTIR characterization of azido CA showed a band at νC≡C
= 2100 cm−1 (weak) corresponding to the new azide side groups
(Fig. 2). 1H NMR analysis also confirmed the chemical structure of
azido CA by the appearing of new characteristic peaks assigned to the
azido hexanoate group in the range of 1.3–1.7 ppm (Hc, Hd, He), at
2.4 ppm (Hb) and at 3.3 ppm (Hf) (Fig. 3). The experimental degree of
substitution DSazido= 0.39 calculated from 1H NMR analysis was in
good agreement with the targeted one, given the 1H NMR error for
polymer analysis, and the corresponding substitution rate for the
hydroxyl groups (87%) was typical for high hydroxyl substitution of
(5) cellulosic derivatives [38].
3.1.2. Synthesis and characterization of CA grafted with PLA
CA-g-poly(D,L-lactide) (CA-g-PLA) glycopolymers were obtained
by a "grafting onto" method involving an α-alkyne PLA oligomer
(Fig. 1). This functional PLA oligomer (Mn NMR = 1640 g/mol) was
synthesized in a single step from D,L-lactide by controlled ROP
according to Laville et al. method [52] using SnOct2 catalyst and
propargyl alcohol as an alkyne-functionalized initiator (see Supporting
information). Azido CA (N3-CA, DSazido = 0.39) was grafted with
different amounts of α-alkyne PLA by CuAAC "click" chemistry. In this
way, the PLA content could be easily varied over a broad range with a
good control of the copolymer architecture. The glycopolymers were
called CA-g-wt%PLA, where wt% was the weight fraction of grafted
PLA.
453
F. Hassan Hassan Abdellatif et al. Journal of Membrane Science 524 (2017) 449–459

Fig. 3. 1H NMR characterization (CDCl3) for the α-alkyne PLA, azido CA and the grafted copolymer CA-g-40.5 PLA.

ATR-FTIR spectrometry was first used to characterize the grafted Table 1

bio-based copolymers by comparison of the stacked spectra of α-alkyne Results of the 1H NMR characterization of the grafted copolymers CA-g-PLA: theoretical
PLA, N3-CA and CA-g-40.5 PLA corresponding to the maximum and experimental degrees of substitution for the PLA grafts, grafting rates and PLA
weight contents in the grafted copolymers.
targeted degree of substitution for the PLA grafts (Fig. 2). Due to the
absence of a common invariable band in the three spectra, their Copolymer DS theo DSexp Grafting rate % PLA wt%
PLA PLA
normalization was not possible and the ATR-FTIR analysis remained
qualitative. Nevertheless, after performing the "click" reaction, the CA-g-23 PLA 0.065 0.058 89% 23
appearance of a new signal of medium intensity at νC=N= 1670 cm−1 CA-g-31.5 PLA 0.1 0.087 87% 31.5
CA-g-40.5 PLA 0.15 0.135 90% 40.5
confirmed the formation of the triazole ring proving covalent grafting.
At 2100 cm−1, a low intensity band characteristic for residual azido
groups was still observed, as expected from the maximum targeted [14]. Therefore, a good control of the copolymers architecture (i.e.
theoretical degree of substitution for the PLA grafts (DStheo
PLA = 0.156 < molecular weight and degree of substitution for the PLA grafts) was
DSazido =0.39). easily obtained by the "grafting onto" strategy. The PLA content was
Similarly as the ATR-FTIR analysis, Fig. 3 shows a comparison of varied between 23 and 40.5 wt%. The focus was made on the PLA
1
H NMR spectra before and after CA modifications. The 1H NMR content range providing the best improvement in membrane properties
spectrum of CA-g-40.5 PLA showed typical signals of both parts, compared to virgin cellulose acetate and lower PLA contents were not
especially the methyl protons peaks at 1.5 ppm assigned to the PLA considered in this work. Furthermore, the maximum PLA grafting rate
grafts and the glucosidic protons peaks (3–5 ppm) from the CA was limited to less than 50 wt% to maintain CA as the principal
backbone. The covalent link of PLA onto N3-CA was confirmed by component in the grafted copolymers. This limitation ensured a good
the appearing of the Hg triazole proton at 7.6 ppm and the correspond- mechanical withstanding of the different membranes in the PV
ing decrease in the Hf signal at 3.3 ppm adjacent to the azido group. operating conditions.
The weight fraction of PLA in the grafted copolymers was calculated The cellulosic membrane materials were investigated by DSC and
from the corresponding experimental degree of substitution of CA with synchrotron SAXS measurements to assess the influence of PLA
the PLA grafts (DSexp
PLA ) (see Experimental part) using equation (8). content on membrane Tg and morphology in relation to the PV
DSexp
PLA × MwtPLA performances obtained for ETBE purification. The thermograms of
PLA wt% = the new grafted copolymer membranes revealed a single glass transi-
Mwt N3CA + DSexp
PLA × MwtPLA (8)
tion temperature and no endothermic melting peaks irrespective of
For each glycopolymer, the data of DStheo
PLA , DSexp
PLA ,
grafting rate and their PLA content (Fig. 4). Fig. 5 showed that the grafting of CA (Tg =
weight fraction in PLA are reported in Table 1. The CuAAC "click" 184 °C) with 23 wt% of PLA (Tg = 13 °C) was responsible for a strong
chemistry allowed to access to grafting rates in the range of 87–90%, Tg decrease (Tg = 118 °C). The short PLA grafts used in this work had a
which corresponded to high modification rates for cellulose chemistry

454
F. Hassan Hassan Abdellatif et al. Journal of Membrane Science 524 (2017) 449–459

0.3

0.2

0.1
Heat Flow (W/g)

CA-g-40.5 PLA

0
CA-g-31.5 PLA
-0.1
CA-g-23 PLA

-0.2

-0.3

Fig. 6. Synchrotron SAXS patterns for virgin CA, azido-CA and the grafted copolymers
-0.4 CA-g-PLA.
0 50 100 150 200

Temperature (° C) Rg close to 75 Å, that could be related to the presence of CA crystallites.

Nevertheless, in a former work [57], we had performed synchrotron
Fig. 4. DSC thermograms obtained during the second scan for the grafted copolymers WAXS on virgin cellulose acetate and shown that the WAXS pattern
CA-g-PLA.
was characteristic of a very weakly crystalline polymer. The PLA
grafting onto CA changed the SAXS pattern significantly (Fig. 6). The
200
scattering laws I(q)~q−3 in the scattering vector q range from 10−2 to
4×10−2 Å−1 showed the disappearance of the CA heterogeneities after
PLA grafting, associated with a more homogeneous structure at the
nanoscale but with possible larger scale fluctuations at the micron-
150 range. The absence of any correlation peak for the PLA grafts showed
that they were finely dispersed in the new bio-based copolymers and
that the new membrane materials were homogeneous at the nanoscale.
T (°C)

100
3.2. Sorption properties of CA grafted with PLA for ETBE purification
g

The sorption-diffusion model [58] describes permeation through a

50
0 brane material. Determining the sorption properties was thus consid-
0 10 20 30
PLA content (wt%)
Fig. 5. Influence of PLA weight content on the glass transition temperature (Tg) of the
grafted copolymers CA-g-PLA.
very low Tg compared to high molecular weight PLA (Tg = 54 °C[53])
and they induced very good plasticizing of the grafted copolymer. By
increasing the PLA content to 31.4 wt%, the plasticizing effect was even
better (Tg = 69 °C)) but it tended to level off for higher PLA content in
the percolation threshold range. In any case, the DSC results showed
that the PLA grafting greatly increased the segmental mobility and
plasticization of the rigid CA chains, which are both expected to
improve the membrane flux.
Further insights into the morphology of the grafted cellulosic
materials were obtained by synchrotron SAXS measurements on fairly
thick membranes (ca. 200 µm) of virgin CA, N3-CA and the grafted
copolymers CA-g-PLA with increasing PLA contents from 23 to 40.5 wt
% (Fig. 6). The weak and complex scattering pattern obtained for virgin
dense polymer membrane by two sequential steps. The first step
corresponds to the sorption of a part of the liquid mixture at the
membrane upstream side. This step is usually selective and contributes
to the feed and permeate enrichments. The second step corresponds to
the diffusion of the absorbed molecules through the polymeric mem-
40 50 ered as a first step towards the understanding of the overall membrane
separation features.
The sorption properties were investigated for virgin CA and CA with
increasing contents of grafted PLA (Table 2). The experiments were
carried out for the azeotropic mixture EtOH/ETBE (20/80 wt%) at
30 °C for safety reasons. As expected, the presence of hydroxyl side
groups ((DSOH = 0.54) in virgin CA led to a very high sorption
separation factor βs= 129 due to their very high affinity for ethanol
while the polymer chains rigidity strongly limited membrane swelling
(S = 4.2 wt%). The cellulosic membranes with increasing PLA contents
from 23 to 40.5 wt% showed a weak almost linear increase in
membrane swelling S from 7.5 to 11 wt% (Fig. 7). At the same time,
the increase in the PLA content induced a decrease in the membrane
Table 2
Sorption results for the azeotropic mixture EtOH/ETBE (20/80 wt%) for virgin CA and
the grafted copolymers CA-g-PLA at 30 °C.
Membrane WPLA (wt%) S (wt%) CsEtOH (wt%) βs

CA was characteristic for large electronic density fluctuations with Virgin CA 0 4.2 97 129
characteristic sizes larger than 50 nm. Previous works have shown the CA-g-23 PLA 23 7.5 65 7.4
presence of intermolecular aggregation and heterogeneities in CA CA-g-31.5 PLA 31.5 9 61 6.25
solutions, which subsisted in the dry membranes after solvent evapora- CA-g-40.5 PLA 40.5 11 53 4.5

tion [54–56]. The CA scattering pattern could be analyzed through the

WPLA: PLA weight fraction in the grafted CA; S: total swelling at sorption equilibrium;
Guinier's law I(q)=I0.exp(−Rg2q2/3) and we assessed a gyration radius CsEtOH: ethanol content in the absorbed mixture; βs: sorption separation factor.

455
F. Hassan Hassan Abdellatif et al. Journal of Membrane Science 524 (2017) 449–459

12 100

Total swelling S (wt%)

10
80
Total swelling S (wt%)

C
60

s
ETOH
6

(wt %)
s
C (wt %)
ETOH
40
4

20
2

0 0
0 10 20 30 40
PLA content (wt%)
Fig. 7. Influence of the PLA content on the total swelling and ethanol content in the
absorbed mixture for the sorption of the azeotropic mixture EtOH/ETBE (20/80 wt%)
for virgin CA and the grafted copolymers CA-g-PLA at 30 °C.
affinity for ethanol and the sorption separation factor decreased from
129 for virgin CA to 4.5 for the membrane with the highest PLA content
(40.5 wt%). Nonetheless, all grafted copolymer membranes were still
selective towards ethanol during the sorption step, as shown by the
ethanol content in the absorbed mixture (≥53 wt%) always greatly
exceeding that of the feed azeotropic mixture EtOH/ETBE (20 wt%).
Therefore, the sorption step systematically ensured a first enrichment
of ethanol in the absorbed mixture, which was an advantage for ethanol
removal from the feed mixture.
3.3. Pervaporation properties of CA grafted with PLA for ETBE
purification
The PV experiments were carried out at 50 °C (rather than 30 °C for
the sorption measurements) to improve the fluxes and facilitate a
comparison between the different membranes (Table 3). We have
already shown that this change in temperature has a negligible
influence on the sorption properties for this particular feed mixture,
due to the fact that the ethanol and ETBE activity coefficients are
almost constant between 30 °C and 50 °C [8].
For virgin CA, the permeate contained ethanol only (C′ = 100 wt%),
which corresponded to an outstanding infinite pervaporation separa-
tion factor βPV, but the normalized flux was low (Jnormalized = 0.023 kg/
h m2). These results were consistent with the pervaporation properties
reported by Nguyen et al. for this separation (Jnormalized = 0.08 kg/h m2
and C′ = 100 wt% at 40 °C) [10]. Despite the higher temperature
(50 °C), the slightly decreased flux obtained in this work may be
ascribed to the much higher molecular weight of the CA used in this
work and to different membrane casting procedures.
50
Fig. 8. Influence of the PLA content on the total normalized flux and permeate ethanol
content for the PV of the azeotropic mixture EtOH/ETBE (20/80 wt%) for virgin CA and
the grafted copolymers CA-g-PLA at 50 °C.
Fig. 8 showed that the normalized PV flux increased linearly with
the membrane PLA content up to almost twelve times that of virgin CA.
This linear increase was in good agreement with the homogeneous
morphology of the membrane materials. In absence of PLA phase
separation, a linear variation of the normalized flux was expected as a
function of the PLA content. Furthermore, the influence of PLA
grafting was almost four times stronger on flux compared to that on
membrane swelling. According to the sorption-diffusion model, this
difference was characteristic for a major contribution of the diffusion
step on the PV mass transfer. The results reported in Table 3 showed
that PLA grafting had a more complex effect on PV selectivity.
After PLA grafting, the ethanol permeate content C′ remained in the
extremely high range (C′ ≥ 99 wt%) for the PLA content of 23 wt%.
This grafted copolymer corresponded to the most rigid grafted
copolymer of the whole series (as shown by the highest Tg of 118 °C
obtained for this material). For this grafted copolymer, the soft PLA
grafts were strongly constrained by the rigid cellulose acetate main
chains, which accounted for 77 wt% of the membrane material. In that
case, the diffusion of the bulky ETBE molecules was extremely
hindered and the permeate contained a very low amount of ETBE
(1 wt%). When the PLA content was further increased up to 40.5 wt%,
the ethanol permeate content gradually decreased to 93 wt% but still
remained in the very high range for this application. According to the
percolation theory of multi-component polymer materials, the percola-
tion threshold usually occurs at a minor component content of ca.
33 wt% and corresponds to the formation of a continuous path for the
minor component inside the polymer material. Therefore, for a PLA
content of 40.5 wt% in the grafted copolymer i.e. above the percolation
threshold, the soft PLA grafts formed a soft percolating network, which
facilitated the diffusion of the ETBE molecules through the copolymer

Table 3
Pervaporation results for the azeotropic mixture EtOH/ETBE (20/80 wt%) for virgin CA and the grafted copolymers CA-g-PLA at 50 °C.

Membrane Thickness (μm) J (g/h m2) Jnormalized (kg/h m2) C′ (wt%) βPV βD PEtOH×1000 (kg µm/h m2 kPa) PETBE×1000 (kg µm/h m2 kPa)

Virgin CA 58 1.98 0.023 100 ∞ – 2.22 0

CA-g-23 PLA 57 14.47 0.165 99 396 53.3 15.8 0.17
CA-g-31.5 PLA 59 18.64 0.22 97 129 20.67 20.6 0.66
CA-g-40.5 PLA 62 21.77 0.27 93 46 10.19 24.3 1.9

WPLA: PLA weight fraction in the grafted CA; J: total flux, Jnormalized: total flux normalized for a reference thickness of 5 µm for comparison; C′: ethanol content in permeate; βpv:
pervaporation separation factor; βD: diffusion separation factor; Pi: partial permeability of species i.

456
F. Hassan Hassan Abdellatif et al.
25
20
x 1000 (kg/h m kPa)
2
15
10
EtOH
P
5 5
0 0
0 10 20 30 40 50
PLA content (wt%)
Fig. 9. Influence of the PLA content on the partial permeability of ethanol (■) and ETBE
(▲)for virgin CA and the grafted copolymers CA-g-PLA during the PV of the azeotropic
mixture EtOH/ETBE (20/80 wt%) at 50 °C.
membrane. Nevertheless, the permeate ethanol content (93 wt%) still
remained very high for this separation owing to the strong cellulose
acetate constraints on the percolating PLA grafts.
Therefore, in any case, the different membranes CA-g-PLA were
highly selective to ethanol (C′ ≥93 wt%) and corresponded to high PV
separation factors βPV as shown in Table 3 entry 4.
For a further understanding of the membrane performances, partial
permeability data were then considered as intrinsic membrane proper-
ties, which only reflected the influence of polymer material and
permeating species irrespective of the operating conditions [59]
(Table 3). These partial permeability data were calculated as the ratio
of partial flux over the corresponding driving force for mass transfer by
analogy with the common practice in gas permeation [60]. For the
virgin CA membrane, the partial permeability of ethanol was
2.22×10−3 kg µm /h m2 kPa with infinite PV separation factor βPV
and nil ETBE permeability. The grafting of PLA onto CA increased the
partial permeability of both ethanol and ETBE (Fig. 9). Nevertheless,
the permeability increase was much stronger for ethanol (×11) than for
ETBE (×2) for the maximum PLA content of 40.5 wt%.
A comparison between the sorption and pervaporation results
showed that the ethanol content in the sorption mixture was always
much less than that of the permeate. As an example, for the best
membrane reported in this work (CA-g-40.5 PLA), the liquid absorbed
by the membrane during sorption contained 53 wt% of ethanol only.
During pervaporation with the same grafted copolymer, the permeate
ethanol content was 93 wt%, meaning that the diffusion step ensured a
further very strong enrichment in ethanol during the PV mass transfer
compared to the sorption step. The sorption and diffusion separation
factors βS and βD were thus compared for virgin CA and the grafted
copolymers membranes for assessing the respective roles of the
sorption and diffusion steps on membrane selectivity (Fig. 10). The
sorption and PV separation factors βS and βPV were first calculated
from the sorption and PV experimental data with Eqs. (4) and (7),
respectively. The corresponding diffusion separation factors βD were
then deduced from Eq. (9) according to the assumptions of the
sorption-diffusion model [58].
βPV = βS × βD
The sorption and pervaporation experiments were carried out at
different temperatures of 30 and 50 °C, respectively. As explained, this
change in temperature had a negligible influence on the sorption
Journal of Membrane Science 524 (2017) 449–459
25 400 β
PV
350
20
P
300
ETBE
Separation factors
x 1000 (kg/h m kPa)
250
15
200
10 150
2
100
βD
50
β
S
0
20 25 30 35 40 45
PLA content (wt%)
Fig. 10. Influence of the PLA content on the separation factors related to sorption,
diffusion and pervaporation for the grafted copolymers CA-g-PLA.
properties and the sorption separation factors at 30 °C were assumed
quasi-identical at 50 °C for calculating the diffusion separation factors
βD. The highest sorption separation factor βs (129) was found for virgin
CA. In addition, the pervaporation separation factor βPV of virgin CA
reached the infinite, leading to an infinite diffusion separation factor
βD. For this reason, the corresponding separation factor data were not
plotted in Fig. 10, which was limited to the data obtained for the
grafted copolymers. For all of them, the diffusion separation factor was
systematically higher than the sorption separation factor. Therefore,
the diffusion step was mainly contributing to the PV separation factor.
When the PLA content increased in Fig. 10, the sorption and diffusion
steps were less selective due to a simultaneous decrease in both H-
bonding ability and rigidity of the membrane materials, which favored
the permeation of the bulky ETBE molecules. Nevertheless, the grafted
copolymers membranes still remained highly selective for ETBE
purification as shown by the very high ethanol permeate contents
(C′≥93 wt%).
4. Conclusion
A new strategy for PLA grafting onto CA was developed for high
performance bio-based membranes for ETBE purification by PV. A
short α-alkyne PLA oligomer was first synthesized by controlled ROP.
This functional PLA oligomer was then grafted onto an azido CA by
CuAAC "click" chemistry. The grafting rate was easily controlled and
the PLA content was varied from 0 to 40.5 wt%. In this study, the
maximum PLA content was limited to less than 50 wt% to ensure the
mechanical withstanding of the different membranes in the PV
operating conditions.
During the sorption of the azeotropic mixture EtOH/ETBE, the new
membranes absorbed ethanol selectively and ensured a first enrich-
ment of the targeted mixture. During PV, the flux greatly increased
with the PLA content and the maximum flux was nearly 12 times that
of virgin CA. At the same time, the ethanol permeate content decreased
slightly with the PLA content but still remained in the very high range
(C′≥93 wt%) for this application and the membrane properties ob-
tained with the best grafted copolymer (CA-g-40.5 PLA) are among the
best ones reported so far for ETBE purification by PV. An analysis
(9)
based on the sorption-diffusion model showed that the diffusion step
mainly contributed to the very high membrane selectivity obtained
with the new bio-based membranes.
The "grafting onto" strategy developed in this work could be
457
F. Hassan Hassan Abdellatif et al.
extended to other polysaccharides and bio-based polymer grafts to
provide a new generation of entirely bio-based membranes for challen-
ging separations in the bio-fuels and petroleum industries.
Acknowledgments
The authors would like to thank the ELEMENT Erasmus Mundus
Programme for the PhD scholarship and the corresponding extension
offered to Mrs Faten HASSAN HASSAN ABDELLATIF. They would like
also to thank Pr Laurent David of the Laboratoire IMP@Lyon1,
Université Claude Bernard Lyon 1, Univ Lyon, CNRS UMR 5223, for
the synchrotron SAXS analyses at the European Synchrotron Radiation
Facility (ESRF) in Grenoble.
Appendix A. Supporting information
Supplementary data associated with this article can be found in the
online version at doi:10.1016/j.memsci.2016.11.078.
References
[1] E.M.V. Hoek, V.V. Tarabara (Eds.) Encyclopedia of Membrane Science and
Technology, Hoboken, New Jersey, John Wiley & Sons, 2013.
[2] E. Drioli, L. Giorno (Eds.) Comprehensive Membrane Science and Engineering,
Kidlington (UK), Elsevier Science, 2010.
[3] R.W. Baker, Membrane Technology and Applications, 2nd ed., John Wiley & Sons,
Chichester, 2004.
[4] A. Jonquieres, C. Arnal-Herault, J. Babin, Pervaporation, in: E.M.V. Hoek, V.V.
Tarabara (Eds.) Encyclopedia of Membrane Science and Technology, Hoboken,
New Jersey, John Wiley & Sons, vol. 3, 2013, pp. 1533–1559.
[5] H. Noureddini, Ethyl tert-butyl ether and methyl tert-butyl ether: status, review,
and alternative use. Exploring the environmental issues of mobile, recalcitrant
compounds in gasoline, ACS Symp. Ser. 799 (2002) 107–124.
[6] F. Lipnizki, Membrane process opportunities and challenges in the bioethanol
industry, Desalination 250 (2010) 1067–1069.
[7] M. Wang, C. Arnal-Herault, C. Rousseau, A. Palenzuela, J. Babin, L. David,
A. Jonquieres, Grafting of multi-block copolymers: A new strategy for improving
membrane separation performance for ethyl tert-butyl (ETBE) bio-fuel purification
by pervaporation, J. Membr. Sci. 469 (2014) 31–42.
[8] M. Billy, A. Ranzani Da Costa, P. Lochon, R. Clement, M. Dresch, A. Jonquieres,
Cellulose acetate graft copolymers with nano-structured architectures: Application
to the purification of bio-fuels by pervaporation, J. Membr. Sci. 348 (2010)
389–396.
[9] S. Zereshki, A. Figoli, S.S. Madaeni, F. Galiano, E. Drioli, Pervaporation separation
of ethanol/ETBE mixture using poly(lactic acid)/poly(vinyl pyrrolidone) blend
membranes, J. Membr. Sci. 373 (2011) 29–35.
[10] Q.-T. Nguyen, C. Léger, P. Billard, P. Lochon, Novel membranes made from a semi-
interpenetrating polymer network for ethanol–ETBE separation by pervaporation,
Polym. Adv. Technol. 8 (1997) 487–495.
[11] G.S. Luo, M. Niang, P. Schaetzel, Pervaporation separation of ethyl tert-butyl ether
and ethanol mixtures with a blended membrane, J. Membr. Sci. 125 (1997)
237–244.
[12] Q.-T. Nguyen, R. Clément, I. Noezar, P. Lochon, Performances of poly(vinylpyr-
rolidone-co-vinyl acetate)-cellulose acetate blend membranes in the pervaporation
of ethanol diethyl tert-butyl ether mixtures: Simplified model for flux prediction,
Sep. Purif. Technol. 13 (1998) 237–245.
[13] G.S. Luo, M. Niang, P. Schaetzel, A high performance membrane for sorption and
pervaporation separation of ethyl tert-butyl ether and ethanol mixtures, Sep. Sci.
Technol. 34 (1999) 391–401.
[14] F. Hassan Hassan Abdellatif, J. Babin, C. Arnal-Herault, A. Jonquieres, Grafting of
cellulose and cellulose derivatives by CuAAC click chemistry Chapter 25, in:
V.K. Thakur (Ed.)Cellulose-based Graft Copolymers: Structure and Chemistry, CRC
Press Taylor & Francis Publisher, Boca Raton, Florida, 2015, pp. 563–591.
[15] Y. Shinkawa, Y. Hayashi, S. Sato, K. Nagai, Permeability of ethanol solution
through poly(lactic acid) film, J. Appl. Polym. Sci. 132 (2015). http://dx.doi.org/
10.1002/APP.4203.
[16] J.A. Camargo, A. Sapin, C. Nouvel, D. Daloz, M. Leonard, F. Bonneaux, J.-L. Six,
P. Maincent, Injectable PLA-based in situ forming implants for controlled release of
Ivermectin a BCS Class II drug: solvent selection based on physico-chemical
characterization, Drug Dev. Ind. Pharm. 39 (2013) 146–155.
[17] C. Duval, C. Nouvel, J.-L. Six, Is Bismuth subsalicylate an effective non-toxic
catalyst for PLGA synthesis?, J. Polym. Sci. Polym. Chem. 52 (2014) 1130–1138.
[18] C. Gavory, A. Durand, J.-L. Six, C. Nouvel, E. Marie, M. Léonard, Polysaccharide-
covered nanoparticles prepared by nanoprecipitation, Carbohyd. Polym. 84 (2011)
133–140.
[19] K. Hamad, M. Kaseem, H.W. Yang, F. Deri, Y.G. Ko, Properties and medical
applications of polylactic acid: a review, Express Polym. Lett. 9 (2014) 435–455.
[20] H. Ebadi-Dehaghani, M. Barikani, H.A. Khonakdar, S.H. Jafari, U. Wagenknecht,
G. Heinrich, On O2 gas permeability of PP/PLA/clay nanocomposites: A molecular
458
Journal of Membrane Science 524 (2017) 449–459
dynamic simulation approach, Polym. Test. 45 (2015) 139–151.
[21] S. Zereshki, A. Figoli, S.S. Madaeni, S. Simone, J.C. Jansen, M. Esmailinezhad,
E. Drioli, Poly(lactic acid)/poly(vinyl pyrrolidone) blend membranes: Effect of
membrane composition on pervaporation separation of ethanol/cyclohexane
mixture, J. Membr. Sci. 362 (2010) 105–112.
[22] Y. Teramoto, Y. Nishio, Cellulose diacetate-graft-poly(lactic acid)s: synthesis of
wide-ranging compositions and their thermal and mechanical properties, Polymer
44 (2003) 2701–2709.
[23] D. Chen, B. Sun, New tissue engineering materials copolymers of derivatives of
cellulose and lactide: Their synthesis and characterization, Mater. Sci. Eng. C 11
(2000) 57–60.
[24] A. Mayumi, T. Kitaoka, H. Wariishi, Partial substitution of cellulose by ring-
opening esterification of cyclic esters in homogeneous system, J. Appl. Polym. Sci.
102 (2006) 4358–4364.
[25] H. Dong, Q. Xu, Y. Li, S. Mo, S. Cai, L. Liu, The synthesis of biodegradable graft
copolymer cellulose-graft-poly(L-lactide) and the study of its controlled drug
release, Coll. Surf. B: Biointerface 66 (2008) 26–33.
[26] C. Yan, J. Zhang, Y. Lv, J. Yu, J. Wu, J. Zhang, J. He, Thermoplastic cellulose-graft-
poly(L-lactide) copolymers homogeneously synthesized in an ionic liquid with 4-
dimethylaminopyridine catalyst, Biomacromolecules 10 (2009) 2013–2018.
[27] Y. Guo, X. Wang, X. Shu, Z. Shen, R.-C. Sun, Self-Assembly and Paclitaxel Loading
Capacity of Cellulose-graft-poly(lactide) Nanomicelles, J. Agric. Food Chem. 60
(2012) 3900–3908.
[28] Y. Luan, J. Wu, M. Zhan, J. Zhang, J. Zhang, J. He, "One pot" homogeneous
synthesis of thermoplastic cellulose acetate-graft-poly(L-lactide) copolymers from
unmodified cellulose, Cellulose 20 (2013) 327–337.
[29] R. Quintana, O. Persenaire, Y. Lemmouchi, L. Bonnaud, P. Dubois, Grafted D/L-
lactide to cellulose acetate by reactive melt processing: Its role as CA/PLA blend
compatibilizer, Eur. Polym. J. 57 (2014) 30–36.
[30] Y. Zhang, X. Li, Y. Yang, A. Lan, X. He, M. Yu, In situ graft copolymerization of L-
lactide onto cellulose and the direct melt spinning, RSC Adv. 4 (2014)
34584–34590.
[31] S. Salmieri, F. Islam, R. Khan, F. Hossain, H.M. Ibrahim, C. Miao, W. Hamad,
M. Lacroix, Antimicrobial nanocomposite films made of poly(lactic acid)-cellulose
nanocrystals (PLA-CNC) in food applications: Part A Effect of nisin release on the
inactivation of Listeria monocytogenes in ham, Cellulose 21 (2014) 1837–1850.
[32] P. Ma, L. Jiang, P. Xu, W. Dong, M. Chen, P.J. Lemstra, Rapid stereocomplexation
between enatiomeric comb-shaped cellulose-g-poly(L-lactide) nanohybrids and
poly(D-lactide) from the melt, Biomacromolecules 16 (2015) 3723–3729.
[33] E. Lizundia, E. Fortunati, F. Dominici, J.L. Vilas, L.M. Leon, I. Armentano,
L. Torre, J.M. Kenny, PLLA-grafted cellulose nanocrystals: Role of the CNC content
and grafting on the PLA bionanocomposites film properties, Carbohydr. Polym. 142
(2016) 105–113.
[34] J. Ambrosio-Martin, M.J. Fabra, A. Lopez-Rubio, J.M. Lagaron, Melt polycon-
densation to improve the dispersion of bacterial cellulose into polylactide via melt
compounding: enhanced barrier and mechanical properties, Cellulose 22 (2015)
1201–1226.
[35] C. Nouvel, P. Dubois, E. Dellacherie, J.-L. Six, Controlled synthesis of amphiphilic
biodegradable polylactide-grafted dextran copolymers, J. Polym. Sci. Part A: Polym.
Chem. 42 (2004) 2577–2588.
[36] I. Ortiz, P. Alonso, A. Urtiaga, Pervaporation of azeotropic mixtures ethanol/ethyl
tert-butyl ether: influence of membrane conditioning and operation variables on
pervaporation flux, Desalination 149 (2002) 67–72.
[37] K.F. Yee, A.R. Mohamed, S.H. Tan, A review on the evolution of ethyl tert-butyl
ether (ETBE) and its future prospects, Renew. Sustain. Energy Rev. 22 (2013)
604–620.
[38] F. Hassan Hassan Abdellatif, J. Babin, C. Arnal-Herault, L. David, A. Jonquieres,
Grafting of cellulose acetate with ionic liquids for biofuel purification by a
membrane process: Influence of the cation, Carbohydr. Polym. 147 (2016)
313–322.
[39] P.-H. Elchinger, P.-A. Faugeras, B. Boens, F. Brouillette, D. Montplaisir,
R. Zerrouki, R. Lucas, Polysaccharides: The "Click" Chemistry Impact, Polymers 3
(2011) 1607–1651.
[40] T. Liebert, C. Hänsch, T. Heinze, Click chemistry with polysaccharides, Macromol,
Rapid Commun. 27 (2006) 208–213.
[41] A. Uliniuc, M. Popa, T. Hamaide, M. Dobromir, New approaches in hydrogel
synthesis – click chemistry: a review, Cellulose, Chem. Technol. 46 (2012) 1–11.
[42] Q. Li, H. Kang, R. Liu, Block and hetero ethyl cellulose graft copolymers
synthesized via sequent and one-pot ATRP and "click" reactions, Chin. J. Chem. 30
(2012) 2169–2175.
[43] W.Z. Xu, J.F. Kadla, Honeycomb films of cellulose azide: molecular structure and
formation of porous films, Langmuir 29 (2013) 727–733.
[44] W. Xu, G. Gao, J. Kadla, Synthesis of antibacterial cellulose materials using a
clickable quaternary ammonium compound, Cellulose 20 (2013) 1187–1199.
[45] K. Negishi, Y. Mashiko, E. Yamashita, A. Otsuka, T. Hasegawa, Cellulose chemistry
meets click chemistry: syntheses and properties of cellulose-based glycoclusters
with high structural homogeneity, Polymers 3 (2011) 489–508.
[46] M. Pohl, J. Schaller, F. Meister, T. Heinze, Selectively dendronized cellulose:
synthesis and characterization, Macromol. Rapid Commun. 29 (2008) 142–148.
[47] M. Pohl, N. Michaelis, F. Meister, T. Heinze, Biofunctional surfaces based on
dendronized cellulose, Biomacromolecules 10 (2009) 382–389.
[48] M. Pohl, G.A. Morris, S.E. Harding, T. Heinze, Studies on the molecular flexibility
of novel dendronized carboxymethyl cellulose derivatives, Eur. Polym. J. 45 (2009)
1098–1110.
[49] M.I. Montanez, Y. Hed, S. Utsel, J. Ropponen, E. Malmstrom, L. Wagberg, A. Hult,
M. Malkoch, Bifunctional dendronized cellulose surfaces as biosensors,
F. Hassan Hassan Abdellatif et al.
Biomacromolecules 12 (2011) 2114–2125.
[50] P.-A. Faugeras, F. Brouillette, R. Zerrouki, Crosslinked cellulose developed by
CuAAC, a route to new materials, Carbohydr. Res. 356 (2012) 247–251.
[51] B. Neises, W. Steglich, Simple method for the esterification of carboxylic acids,
Angew. Chem. Int. Ed. 17 (1978) 522–524.
[52] M. Laville, J. Babin, I. Londono, M. Legros, C. Nouvel, A. Durand, R. Vanderesse,
M. Leonard, J.-L. Six, Polysaccharide-covered nanoparticles with improved shell
stability using click-chemistry strategies, Carbohydr. Polym. 93 (2013) 537–546.
[53] S. Jacobsen, H.G. Fritz, Filling of poly(lactic acid) with native starch, Polym. Eng.
Sci. 36 (1996) 2799–2804.
[54] E. Fleury, J. Dubois, C. Léonard, J.P. Joseleau, H. Chanzy, Microgels and ionic
associations in solutions of cellulose diacetate, Cellulose 1 (1994) 131–144.
[55] K.D. Goebel, G.C. Berry, D.W. Tanner, Properties of cellulose acetate. III. Light
scattering from concentrated solutions and films, Tensile creep and desalination
459
Journal of Membrane Science 524 (2017) 449–459
studies on films, J. Polym. Sci.: Polym. Phys. Ed. 17 (1979) 917–937.
[56] H. Suzuki, Y. Muraoka, M. Satio, K. Kamide, Light-scattering study on cellulose
diacetate in 2-butanone, Eur. Polym. J. 18 (1982) 831–837.
[57] M. Billy, A. Ranzani Da Costa, P. Lochon, M. Dresch, R. Clément, S. Etienne,
J.M. Hiver, L. David, A. Jonquieres, Cellulose acetate-g-poly(methyl diethylene
glycol methacrylate) copolymers with nano-structured architectures: Synthesis and
characterization, Eur. Polym. J. 46 (2010) 944–957.
[58] J.G. Wijmans, R.W. Baker, The solution-diffusion model: a review, J. Membr. Sci.
107 (1995) 1–21.
[59] R.W. Baker, J.G. Wijmans, Y. Huang, Permeablity, permeance and selectivity: A
preferred way of reporting pervaporation performance data, J. Membr. Sci. 348
(2010) 346–352.
[60] J.G. Wijmans, Process performance=membrane properties+operating conditions,
J. Membr. Sci. 220 (2003) 1–3.