Honeycomb membranes prepared from 3-O-amino acid functionalized cellulose derivatives

Article abstract of DOI:10.1016/j.carbpol.2012.12.076

The development of value-added wood-derived polymer products is of significant importance. Of particular interest is the synthesis of advanced bioactive cellulosic materials. In the present research, novel cellulosic honeycomb films are reported. Cellulos

Full text of DOI:10.1016/j.carbpol.2012.12.076

Carbohydrate Polymers 100 (2014) 126–134
Contents lists available at ScienceDirect
Carbohydrate Polymers
journal homepage: www.elsevier.com/locate/carbpol
Honeycomb membranes prepared from 3-O-amino acid functionalized cellulose
derivatives
∗
William Z. Xu, Batia Ben-Aroya Bar-Nir, John F. Kadla
Advanced Biomaterials Chemistry Laboratory, University of British Columbia, Vancouver, BC V6T 1Z4, Canada
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 13 June 2012
Received in revised form
The development of value-added wood-derived polymer products is of significant importance. Of par-
ticular interest is the synthesis of advanced bioactive cellulosic materials. In the present research,
novel cellulosic honeycomb films are reported. Cellulose was reacted with dimethylthexylsilyl chlo-
ride to form regioselective 2,6-di-O-thexyldimethylsilyl cellulose followed by substitution of the C3
with functionalized poly(ethylene glycol) (PEG). The free end of the PEG side chains of the regiose-
lective 3-O-poly(ethylene glycol)–2,6-di-O-thexyldimethylsilyl cellulose served as an attachment point
for bioactive molecules. As an example, Fmoc–Gly–OH was linked to the free end of PEG to produce
2
1 November 2012
Accepted 27 December 2012
Available online 18 January 2013
Keywords:
Cellulose
3
-O-Fmoc–Gly–poly(ethylene glycol)–2,6-di-O-thexyldimethylsilyl cellulose. Honeycomb films were
produced through film casting under a humid airflow. AFM analysis revealed the directed self-assembly
of the 3-O-Fmoc–Gly–poly(ethylene glycol)–2,6-di-O-thexyldimethylsilyl cellulose wherein the pendent
3-O-Fmoc–Gly–poly(ethylene glycol) groups allocated preferentially around the edges of the honeycomb
pores.
Regioselective
Poly(ethylene glycol)
3
-O-Poly(ethylene
glycol)–2,6-di-O-thexyldimethylsilyl
cellulose
Honeycomb
©
2013 Elsevier Ltd. All rights reserved.
1
. Introduction
tert-butyldimethylsilyl (Heinze, Pfeifer, & Petzold, 2008), and
thexyldimethylsilyl (Koschella, Heinze, & Klemm, 2001; Petzold,
Einfeldt, Gunther, Stein, & Klemm, 2001) groups are often employed
in regioselectively modifying cellulose.
Over the last several years the significance of sustainable
development has become widely understood and accepted. As
a result a growing trend in the research and application of
renewable resources has occurred. Cellulose is the most abun-
dant polymer on earth and as such has been extensively studied
for the development of advanced materials such as nanocom-
posites (Goetz, Foston, Mathew, Oksman, & Ragauskas, 2010;
Peresin, Habibi, Zoppe, Pawlak, & Rojas, 2010), blood compat-
ible (Liu et al., 2009) and thermally responsive (Azzam, Heux,
Putaux, & Jean, 2010; Ifuku & Kadla, 2008) materials. Most recently,
particular interest has been in the development of regioselec-
tively substituted cellulose derivatives due to their designable
properties, such as solubility and crystallinity (Kondo, Koschella,
Heublein, Klemm, & Heinze, 2008). In polysaccharide chemistry,
regioselectivity means a pre-set functionalization of one or two
selected hydroxyl groups within the repeating unit (Heinze, 2009).
Mesoporous and macroporous membranes are receiving con-
siderable research interest, due to their wide range of applications
such as separation and dialysis (Risbud & Bhonde, 2001), drug
delivery (Ma & McHugh, 2007), electronics (Imada et al., 1999),
photonics (Wijnhoven & Vos, 1998), and surfaces for cell growth
(Beattie et al., 2006; Stenzel, Barner-Kowollik, & Davis, 2006).
Lithography and “stamping” have been applied to deposit biomem-
branes onto cellulose to form interesting materials with ordered
porous structure (Tanaka, Wong, Rehfeldt, Tutus, & Kaufmann,
2004), but they are time-consuming and require special equip-
ment. Simply by placing a thin layer of water in oil (W/O) emulsion
in a closed box containing water, honeycomb patterned cellu-
lose acetate films with size of 1–100 ␮m were formed (Kasai &
Kondo, 2004). By means of saponification of the honeycomb films
of cellulose triacetate, uniform honeycomb films of cellulose were
fabricated by Nemoto et al. (2005) using a three-step transcription
method, by which the films could be easily reproduced. Ultrathin
(<20 nm) films of poly(methyl methacrylate) (PMMA) with cavities
2
,2,6,6-Tetramethylpiperidine-1-oxyl radical (TEMPO)-mediated
oxidation (Isogai, Saito, & Fukuzumi, 2011) has been extensively
applied in functionalizing the primary hydroxyl group at C6 posi-
tion. Bulky protecting groups such as trityl (Yue & Cowie, 2002),
(diameter: 10–160 nm, depth: 1–10 nm) containing cellulose deco-
ration were also reported by using binary polymer blends (Kontturi,
Johansson, & Laine, 2009). But the most convenient method
to prepare porous films with regular hexagonal arrays, termed
∗ Corresponding author. Tel.: +1 604 827 5254; fax: +1 604 822 0661.
E-mail address: john.kadla@ubc.ca (J.F. Kadla).
0
144-8617/$ – see front matter © 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.carbpol.2012.12.076

W.Z. Xu et al. / Carbohydrate Polymers 100 (2014) 126–134
127
honeycomb structured films and membranes, is to cast films by
2.2. Characterization
1H and 13C NMR spectra were measured using a Bruker AVANCE-
controlling solvent evaporation under a humid atmosphere, known
as “breath figures” (Stenzel, 2002). Specifically during the evapo-
ration of solvent from a polymer solution there is condensation of
water onto the cooling surface. By controlling the system the water
droplets create an ordered hexagonal template in which the poly-
mer precipitates at the water–solvent interface, resulting in the
encapsulation of water droplets and preventing their coalescence
◦
◦
300 spectrometer at 25 C for small molecules and 40 C for
polymers. Chemical shifts were referenced to tetramethyl silane
(TMS; 0.0 ppm). Infrared spectra were obtained with a Perkin-
Elmer Spectrum One FT-IR spectrometer as potassium bromide
pellets in a 3:1 sample to salt ratio (insoluble samples) or as
neat films cast from CHCl3 (soluble samples) onto ZnSe win-
(Widawski, Rawiso, & Francois, 1994). Amphiphilicity of polymers
−
1
is often needed for the formation of honeycomb membranes since
the porous micro-structure is coupled with a micro-phase mor-
phology that leads to a hydrophobic surface while the pores are
enriched with hydrophilic groups (Wong et al., 2006). Although
honeycomb membranes can be formed with either regioselectively
or randomly functionalized cellulose derivatives, regioselectively
functionalized cellulose derivatives have a more regular molecu-
lar structure, so are more suitable for the study of the effect of
molecular structure such as the degree of substitution (DS) and
the chain length (DP) of the hydrophilic groups on the formation of
honeycomb membranes.
It is hypothesized that the poly(ethylene glycol) segments will
preferentially interact with the condensing water droplets and act
to direct the self-assembly of the regioselective cellulosic honey-
comb films. As a result, the attachment of biomolecules to the
terminal end of the poly(ethylene glycol) chains would be pref-
erentially allocated to the walls of the honeycomb pores and be
accessible to interact with other biomolecules in solution when
exposed to water. Not only are some Fmoc-amino acids bioac-
tive with anti-inflammatory activities (Burch et al., 1991), but
they are also widely used in the synthesis of peptides when pro-
tection of the amino groups is needed (Wu & Xu, 2001). In the
present research, Fmoc–Gly–OH was selected as a model amino
acid compound to link to the free end of PEG to produce 3-
O-Fmoc–Gly–poly(ethylene glycol)–2,6-di-O-thexyldimethylsilyl
cellulose. Honeycomb films of 3-O-Fmoc–Gly–poly(ethylene
glycol)–2,6-di-O-thexyldimethylsilyl cellulose were cast by con-
trolled solvent evaporation under a humid atmosphere and
analyzed using atomic force microscopy (AFM) and a scanning
electron microscope (SEM). This work is believed to provide a plat-
form for preparation of bioactive membranes by attaching bioactive
molecules for applications such as detection and deactivation of
bacteria.
dows. Spectra were recorded at a resolution of 4 cm
using
a total of 32 scans. Elemental analysis was measured using
a Perkin-Elmer Series II CHNS/O analyzer. Polymer molecular
weight was determined by using a GPC (Agilent 1100, UV and
RI detectors) connected to a multi-angular light scattering detec-
tor (DAWN Heleos-II, Wyatt Technologies) and calibrated with
polystyrene standards. Columns: Stryagel HR-1, HR-3, and HR-4
◦
(Waters Corp., Milford, USA); temperature: 35 C, eluting sol-
vent: THF; flowrate: 0.5 mL/min; sample concentration: 2 mg/mL;
injection volume: 100 ␮L. Modulus of honeycomb films was mea-
sured by using PeakForce QNM (Quantitative NanoMechanics)
mode with AFM (Veeco Multimode coupled with Nanoscope V
Controller). The spring constant and deflection sensitivity of a
Scanasyst-Air silicon tip on nitride lever were calibrated with a
Sapphire-15M standard (Veeco). The surface morphology of the
honeycomb films was observed with AFM and a SEM (Hitachi S-
2600N scanning electron microscope). Pore size was measured
on SEM images using Image-J software (NIH, USA) and the sta-
tistical analysis of pore size distribution was performed based on
more than 3000 pores for each sample by means of MATLAB soft-
ware.
2.3. Synthesis
Poly(ethylene glycol) monotosylate (1a–c) was synthesized from
the corresponding poly(ethylene glycol) [Mw = 200 (EG ) for 1a,
4
Mw = 600 (EG13) for 1b, and Mw = 1000 (EG22) for 1c]. Poly(ethylene
glycol) (0.006 mol) was dissolved in 10/40 mL H O/THF. Potas-
2
sium hydroxide (0.34 g, 0.006 mol, 1 equiv.) was added and the
solution was cooled in ice water bath. p-Tosyl chloride (0.86 g,
0.0045 mol, 0.75 equiv.) solution in 50 mL THF was added drop-
wise. The reaction mixture was warmed to room temperature and
stirred for additional 18 h. The THF was evaporated. Brine solu-
tion was added and washed 5 times with chloroform. The organic
phases were combined, washed with water, dried over magne-
2
. Experimental
sium sulfate and the solvent was removed by evaporation in vacuo.
2.1. Materials
1
Yield: 1a 66%; 1b 77%; 1c 87%. H NMR (CDCl , 300 MHz) (Canaria,
3
Smith, Yu, Fraser, & Lansford, 2005): ı (ppm) 7.69 (d, J = 7.89 Hz,
Polyethylene glycol Mw = 200 (EG ), polyethylene gly-
4
2H), 7.25 (d, J = 7.89 Hz, 2H), 4.06 (t, J = 4.82 Hz, 2H), 3.7–3.5 (mul-
col Mw = 600 (EG13), polyethylene glycol Mw = 1000 (EG22),
dimethylthexylsilyl chloride (TDMSCl), imidazole, anhydrous
N,N-dimethylacetamide (DMA), hydroxybenzotriazole (HOBt),
N-(9-fluorenylmethoxycarbonyl)-glycine (Fmoc–Gly–OH), p-tosyl
13
tiplet, 14.5/50.9/87.3H), 2.35 (s, 3H). C NMR (CDCl , 75.4 MHz):
3
(
ppm) 144.3, 133.0, 129.4, 127.5 (C H ), 72.2–68.3 (CH EG), 61.3
6 4 2
(
CH2 OH), 21.2 (CH3 C H ).
6
4
Poly(ethylene glycol) monoiodide (2a–c) was synthesized from
the corresponding poly(ethylene glycol) monotosylate (1a–c).
Poly(ethylene glycol) monotosylate (0.006 mol) was dissolved in
ꢀ
chloride, sodium hydride (60% dispersion in mineral oil), N,N -
dicyclohexylcarbodiimide (DCC), 4-(dimethylamino)pyridine
DMAP), sodium sulfite, and tetra-n-butylammonium iodide
TBAI) were purchased from Sigma–Aldrich and used as received.
Potassium hydroxide, sodium hydroxide, anhydrous magnesium
sulfate, potassium iodide, acetone (ACS certified), chloroform (ACS
certified), sodium chloride (ACS certified), sodium bicarbonate
(
(
5
0 mL of acetone and carefully added to a refluxing potassium
iodide (2.9 g, 0.18 mol, 3 equiv.) in 50 mL of acetone solution. The
mixture was refluxed for 18 h and cooled to room temperature.
Water was added to receive a clear solution (∼50 mL). The ace-
tone was removed under reduced pressure. Saturated sodium
sulfite solution (50 mL) was added gaining a colorless solution
and the aqueous phase was washed 5 timed with chloroform.
The organic phases were combined, dried over magnesium sul-
fate and the solvent was removed by evaporation in vacuo. Yield:
(
ACS certified), and dichloromethane (DCM) were purchased from
Fisher Scientific and used as received. Tetrahydrofuran (THF) and
anhydrous lithium chloride (Sigma–Aldrich) were dried before
use. Cellulose was obtained from the deacetylation of cellulose
acetate (average Mn ∼30 kDa, Aldrich) using sodium hydroxide in
methanol and thoroughly dried prior to use.
a 57%; 2b 65%; 2c 98%. 1H NMR (CDCl , 300 MHz): ı (ppm)
.7–3.5 (multiplet, 14.5/50.9/87.3H), 3.15 (t, 2H). C NMR (CDCl3,
2
3
3
13

1
28
W.Z. Xu et al. / Carbohydrate Polymers 100 (2014) 126–134
7
(
5.4 MHz): ı (ppm) 72.1–69.8 (CH2 EG), 61.4 (CH2 OH), 2.6
CH2 I).
,6-di-O-thexyldimethylsilyl cellulose (3) (2,6-TDMS cellulose)
was synthesized according to the procedure of Koschella et al.
Kadla, Asfour, & Bar-Nir, 2007; Koschella et al., 2001), as reported
in the previously published paper (Bar-Nir & Kadla, 2009). Yield:
9%. Elemental analysis: C59.03%, H10.58%. DS: 1.96 (calculated
3. Results and discussion
2
3.1. Synthesis and characterization of poly(ethylene glycol)
monoiodide
(
In order to synthesize 3-AA-PEG–2,6-TDMS cellulose,
poly(ethylene glycol) monoiodide (2a–c) was first synthesized.
The reaction scheme for the preparation of poly(ethylene glycol)
monoiodide is illustrated in Scheme 1. Poly(ethylene glycol) mono-
tosylate (1a–c) was synthesized by reacting poly(ethylene glycol)
with p-tosyl chloride at room temperature for 18 h, followed by
further conversion to poly(ethylene glycol) monoiodide (2a–c) by
reacting 1a–c with potassium iodide.
8
with the data from elemental analysis). Molecular weight: Mw
7
1
8 kDa, PDI 1.4. H NMR (C D , 313 K, 300 MHz): ı (ppm) 4.77–3.52
6
6
13
(
cellulose), 1.86–0.27 (TDMS). C NMR (C D , 313 K, 75.4 MHz):
6 6
ı (ppm) 102.5 (C1), 76.9, 75.6, 72.1, 71.9, 60.6 (C6), 35.1 to −2.8
TDMS).
-O-hydroxypoly(ethylene glycol)–2,6-di-O-thexyldimethylsilyl
(
3
1
cellulose (4a–c) (3-OHPEG–2,6-TDMS cellulose). General pro-
cedure: sodium hydride (0.01 mol, 5 equiv.) was washed three
times with freshly distilled anhydrous THF. 2,6-TDMS cellulose
The synthesized products were characterized with NMR. H and
C NMR spectra of the synthesized poly(ethylene glycol) mono-
13
tosylate 1b and poly(ethylene glycol) monoiodide 2b, as well as
the purchased poly(ethylene glycol) Mw = 600 (EG13) are com-
pared in Figs. 1 and 2, respectively. As seen in Fig. 1, in contrast
to the spectrum of the starting material poly(ethylene glycol),
additional peaks at 7.69, 7.25, 4.06, and 2.35 ppm of the prod-
uct 1b indicate successful introduction of tosyl groups at the end
of poly(ethylene glycol). Although the reaction stoichiometry was
selected to functionalize only one of the hydroxyl groups, a small
amount (ca. 5%, see supporting information) of poly(ethylene gly-
col) di-tosylate was formed and some unreacted poly(ethylene
glycol) remained after the reaction. The unreacted poly(ethylene
glycol) was removed through washing. Complete conversion of 1b
to 2b was confirmed by the disappearance of the above mentioned
tosyl-related peaks and the appearance of the new peak of 2b at
3.15 ppm. According to Fig. 2, the introduction of tosyl group to
the end of poly(ethylene glycol) was also evidenced by the new
appearance of tosyl peaks (1b) at 144.3, 133.0, 129.4, 127.5, and
21.2 ppm compared to the spectrum of poly(ethylene glycol). After
the conversion of 1b to 2b, these tosyl-related peaks disappeared
and a new peak appeared at 2.6 ppm due to replacement of the
tosyl group with an iodo group. The unreacted hydroxyl group at
the other end of 1b and 2b was confirmed by the peak at 61.4 ppm.
(
(
3) (1 g, 0.002 mol) and tetra-n-butylammonium iodide (TBAI)
75 mg, 0.02 mmol, 0.01 equiv.) were added and stirred in 24 mL
anhydrous THF, at room temperature for 1 h. Poly(ethylene glycol)
monoiodide (2a–c) (3.3/7.3/11.3 g, 0.01 mol, 5 equiv.) was added.
The mixture was stirred for 72 h at room temperature, and then
slowly precipitated into methanol while destroying the excess
of sodium hydride. Further purification was done by redissolving
the crude polymer in chloroform and precipitating in methanol.
Yield: 4a 0.70 g; 4b 0.85 g; 4c 0.60 g. Elemental analysis: 4a
C58.30%, H10.12%; 4c C58.46%, H10.18%. Molecular weight: 4a
Mw 110 kDa, PDI 1.8; 4c Mw 160 kDa, PDI 2.8. FT-IR: 3509 (OH),
2
957, 2870 (CH ), 1465, 1378 (CH , CH ), 1252 (Si C), 1152,
3 2 3
1
117 (Si O C), 1078, 1036 (C O CAGU), 834, 778 (Si C/Si O C).
1
H NMR (C D , 313 K, 300 MHz): ı (ppm) 4.76–3.52 (cellulose),
6
6
13
3
.51 (EG), 1.85–0.28 (TDMS). C NMR (C D , 313 K, 75.4 MHz):
6 6
ı (ppm) 71.5 (EG), 62.4 (EG CH2 OH), 35.1 to −2.8 (methyl of
TDMS).
3-O-Fmoc–Gly–poly(ethylene glycol)–2,6-di-O-thexyldimethyl-
silyl cellulose (5a–c) (3-AA-PEG–2,6-TDMS cellulose). General
procedure: 3-O-hydroxypoly(ethylene glycol)–2,6-di-O-thexyldi-
methylsilyl cellulose (4a–c) (0.0017 mol) was dissolved in 10 mL
methylene chloride (DCM). Fmoc–Gly–OH (1.3 g, 0.0043 mol,
2
.5 equiv.) and HOBt (0.574 g, 0.0043 mol, 2.5 equiv.) were dis-
solved in 10 mL DCM and added to the cellulose solution. DMAP
0.020 g, 0.1 equiv.) and DCC (0.7 mL, 0.0043 mol, 2.5 equiv.)
3.2. Synthesis and characterization of 3-AA-PEG–2,6-TDMS
cellulose
(
were added and the reaction was stirred at room tempera-
ture for 8 h. The solution was filtered and concentrated, then
precipitated into methanol. Further purification was done by
redissolving the crude polymer in chloroform and precipitating in
methanol. Yield: 5a 0.70 g; 5b 0.84 g; 5c 0.62 g. Elemental anal-
ysis: 5a C59.57%, H9.80%, N0.42%; 5b C59.28%, H9.87%, N0.36%;
By using the synthesized poly(ethylene glycol) monoiodide
(2a–c),
3-O-Fmoc–Gly–poly(ethylene
glycol)–2,6-di-O-
thexyldimethylsilyl cellulose was then synthesized. Scheme 2
displays the synthetic scheme of 3-O-Fmoc–Gly–poly(ethylene
glycol)–2,6-di-O-thexyldimethylsilyl cellulose. The regener-
ated cellulose (from cellulose acetate, average Mn 30 kDa) was
employed in this research because it has known and relatively low
molecular weight. Cellulose derivatives of lower molecular weight
should have better solubility for chemical reaction and film casting.
The regenerated cellulose is not soluble in common solvents but
can be dissolved in DMA/LiCl. After the silylation of cellulose, the
formed product can be easily dissolved in many solvents including
THF, chloroform, toluene, benzene, etc. By reacting cellulose
5
c C58.99%, H9.86%, N0.33%. Degree of substitution calculated
with the nitrogen content from the elemental analysis: 5a 0.16;
b 0.15; 5c 0.15. Molecular weight: 5a Mw 95 kDa, PDI 1.4; 5b
Mw 100 kDa, PDI 1.6; 5c Mw 97 kDa, PDI 2.2. FT-IR: 3503 (OH),
957, 2870 (CH ), 1760 ( CH2 COO ), 1732 ( NH COO ), 1515
5
2
3
(
amide II band) (Socrates, 2001), 1465, 1378 (CH , CH ), 1252
2 3
(
Si C), 1152, 1118 (Si O C), 1079, 1036 (C O CAGU), 834, 778
1
3
(
Si C/Si O C). C NMR (C D , 313 K, 75.4 MHz): ı = 145.0, 142.2,
6 6
◦
1
25.9, 120.6, 76.2, 71.4, 69.4, 67.4, 64.7, 48.1, 43.4, 35.1 to −2.8
with TDMSCl in DMA/LiCl in the presence of imidazole at 100 C,
(
TDMS).
silylation of cellulose took place preferentially at the C6 and C2
positions (Koschella et al., 2001). The bulky TDMS groups bonded
to the C6 and C2 positions prevented the less reactive hydroxyl
group at the C3 position from being further silylated, resulting
in regioselective protection of the hydroxyl groups of cellulose.
Poly(ethylene glycol) monoiodide (2a–c) was then attached to the
C3 position at the regioselective 2,6-TDMS cellulose (3) to generate
-OHPEG–2,6-TDMS cellulose (4a–c), with which Fmoc–Gly–OH
was subsequently reacted to afford 3-O-Fmoc–Gly–poly(ethylene
glycol)–2,6-di-O-thexyldimethylsilyl cellulose (5a–c).
2.4. Film preparation
Honeycomb-patterned films were prepared by applying 10 ␮L
of neat solution (1% in toluene) onto a glass slide in a humid envi-
ronment (flow rate: 700 mL/min; relative humidity: 70–80%; room
temperature).
3

W.Z. Xu et al. / Carbohydrate Polymers 100 (2014) 126–134
129
O
H
O
n
i
H
O
n
ii
H
I
n
O
H
O
S
O
O
1
a-c
2a-c
Scheme 1. Synthetic scheme to poly(ethylene glycol) monoiodide. Note: n = 4 for a, n = 13 for b, and n = 22 for c. (i) p-Tosyl chloride, KOH, THF, H2O, room temperature, 18 h;
ii) KI, acetone, refluxing, 18 h.
(
2
,3
4
2
5
6
1
O
O
4
n-1
HO
I
3
5
2
b
1
O
S
7
7
8
8
9
9
HO
O
7
8
O
6
n-1
1
b
Poly(ethylene glycol) MW=600
7.5
7.0
6.5
6.0
5.5
5.0
4.5
4.0
3.5
3.0
2.5
Chemical Shift (ppm)
Fig. 1. A comparison of 1H NMR spectra of poly(ethylene glycol) Mw = 600, and the synthesized poly(ethylene glycol) monotosylate (1b) and poly(ethylene glycol) monoiodide
2b).
(
The synthesized cellulose derivatives were characterized with
should be noted that the small amount of impurity, poly(ethylene
glycol) di-tosylate, could act as a crosslink agent to bond differ-
ent cellulose chains together. However, the effect of crosslink is
very limited as the formed insoluble crosslinked products during
the synthesis of 4a–c would be filtered off during the purifica-
tion process after the reaction. In the previous work (Bar-Nir &
Kadla, 2009), iodide of poly(ethylene glycol) monomethyl ether
was attached to the C3 position of 2,6-TDMS cellulose, resulting
FTIR, 1H and C NMR, and elemental analysis. The characteriza-
13
tion results of product 3 have been previously reported (Bar-Nir
1
&
Kadla, 2009). Fig. 3 displays a comparison of H NMR spectra of
1
the synthesized products 3, 4b, and 5b. In contrast to the H NMR
spectrum of 2,6-TDMS cellulose, the additional peak in the spec-
trum of 4b appearing at 3.51 ppm suggests successful attachment
of poly(ethylene glycol) to the C3 position of 2,6-TDMS cellulose. It
3-5
1
4
5
8
O
O
HO
1
O
I
CDCl3
2
6
3
7
2
n-3
8
6
7
2
b
O
13 14
13 14
11
12
15
O
O
S
16
HO
O
O
9
10
13
n-3
O
1
4
9
11
16
15
12
1b
Poly(ethylene glycol) MW=600
1
40
130
120
110
100
90
80
70
60
50
40
30
20
10
Chemical Shift (ppm)
Fig. 2. A comparison of 13C NMR spectra of poly(ethylene glycol) Mw = 600, and the synthesized poly(ethylene glycol) monotosylate (1b) and poly(ethylene glycol) monoiodide
2b).
(

1
30
W.Z. Xu et al. / Carbohydrate Polymers 100 (2014) 126–134
Si
O
Si
O
OH
O
(
O
(
(
i
O
ii
H
O
O
)
HO
O
HO
O
)
)
O
O
O
OH
n
Si
Si
3
4a-c
7
'
8'
9'
11'
9'
10'
Si
O
O
7'
8'
2
4
O
6
2
3
4
iii
(
5
2
25
2
2
20
17
14
13
O
O 3
O
1
)
2
1
19
16
O
NH
15
O
O
1
2
7
8
8
9
9
1
8
n-1
2
6
2
7
10
7
2
2
1
1
O
Si
27
2
4
3
2
6
2
5
a-c
2
5
Scheme 2. Synthetic scheme to 3-O-Fmoc–Gly–poly(ethylene glycol)–2,6-di-O-thexyldimethylsilyl cellulose. Note: n = 4 for a, n = 13 for b, and n = 22 for c. (i) TDMSCl,
imidazole, DMA, LiCl; (ii) poly(ethylene glycol) monoiodide 2a–c, NaH, TBAI, THF; (iii) Fmoc–Gly–OH, HOBt, DMAP, DCC, DCM, room temperature, 8 h.
in relatively high DS 3-O-methoxypoly(ethylene glycol)–2,6-di-O-
thexyldimethylsilyl cellulose. In order to link Fmoc–Gly–OH to
the free end of PEG at the C3 position of 2,6-TDMS cellulose,
poly(ethylene glycol) monoiodide (2a–c) was employed in the syn-
thesis of 4a–c. The relatively low DS of 4a–c and thus 5a–c may be
due to the side reactions of intramolecular and intermolecular elim-
ination of 2a–c during the synthesis of 4a–c. In addition, it could
not be excluded that 2a–c reacted with itself to increase the side
chain length during the synthesis of 4a–c, but according to the inte-
( CH2 CH2 OH) and other methylene groups in the PEG side
chain, respectively. After the reaction of 4b with Fmoc–Gly–OH, the
peak at 62.4 ppm disappears and many new 13C NMR peaks appear
at 145.0, 142.2, 125.9, 120.6, 69.4, 67.4, 64.7, 48.1, and 43.4 ppm
in the spectrum of the formed product 5b (Fig. 4). The assignment
of these peaks is shown in Fig. 4. The peaks of aromatic carbons
at positions 24 and 25 (Scheme 2) are overlapped with the solvent
(C D ) peaks; hence they could not be assigned. The disappearance
6
6
of the peak at 62.4 ppm indicates the reaction of the end hydroxyl
group at the free end of PEG side chain. In addition, FTIR results sug-
gest successful conversion of 4b to 5b by the additional carbonyl
gration analysis of the 1H NMR spectrum of product 5b in CD Cl2
2
(
see supporting information), this reaction should be insignificant.
The additional peaks at 7.58, 7.43, and 7.21 ppm in the spectrum of
b are attributed to the aromatic group, indicating successful link-
age of Fmoc–Gly–OH to the free end of PEG at the C3 position of
−
1
−1
peaks at 1760 and 1732 cm and the peak at 1515 cm of sample
5
5b.
2
,6-TDMS cellulose.
3.3. Formation of honeycomb membrane from
3-AA-PEG–2,6-TDMS cellulose
The successful synthesis of 5a–c was also confirmed by 13C
NMR and FTIR. As seen in Fig. 4, by attaching poly(ethylene gly-
col) to 2,6-TDMS cellulose (3), additional 13C peaks of product
In contrast to the top-down techniques such as the lithographic
method to form micro-sized patterns, bottom-up techniques are
4
b appear at 62.4 and 71.4 ppm, attributed to the end methylene
8,9,8',9'
7,7'
12,13
2-6,14,15,17,20,21
11,11'
1
23-26
5b
4b
3
7
.5
7.0
6.5
6.0
5.5
5.0
4.5
4.0
3.5
3.0
2.5
2.0
1.5
1.0
0.5
0
Chemical Shift (ppm)
Fig. 3. A comparison of 1H NMR spectra of the synthesized products of 2,6-TDMS-cellulose (3), 3-OHPEG–2,6-TDMS cellulose (4b), and 3-AA-PEG–2,6-TDMS cellulose (5b)
see Scheme 2 for the labeling).
(

W.Z. Xu et al. / Carbohydrate Polymers 100 (2014) 126–134
131
12,13
1
0,10'
8
,8'
9,9'
11,11'
7,7'
2
7
26
14 15
23
21
2
2
5b
17
2
0
4b
3
1
40
130
120
110
100
90
80
70
60
50
40
30
20
10
0
Chemical Shift (ppm)
Fig. 4. A comparison of 13C NMR spectra of the synthesized products of 2,6-TDMS-cellulose (3), 3-OHPEG–2,6-TDMS cellulose (4b), and 3-AA-PEG–2,6-TDMS cellulose (5b)
see Scheme 2 for the labeling).
(
often applied to form honeycomb structured porous materials.
Wong et al. (2006) investigated the four different casting variations
of the water droplet templating method, i.e., airflow, cold stage,
casting on water, and emulsion, and found that the airflow casting
technique was a versatile method for many polymer materials. In
addition to many physical parameters such as humidity and airflow
rate, chemical aspect is a major factor to affect the formation of reg-
ular honeycomb-patterned films (Wong, Davis, Barner-Kowollik, &
Stenzel, 2007). Our synthesized amphiphilic 3-AA-PEG–2,6-TDMS
celluloses were cast under a humid airflow to form honeycomb
films. A schematic diagram of the formation of honeycomb films
with the “breath figures” method is shown in Fig. 5. When a thin
layer of polymer solution was placed under a humid air flow, the
decreasing temperature of the polymer solution due to evaporation
of the solvent resulted in condensation of water droplets onto the
surface of polymer solution. With the solvent further evaporating,
the water droplets were rearranged to form an ordered hexagonal
template by capillary force. Honeycomb films were formed with
precipitation of the polymer at the water–solvent interface and
complete evaporation of the solvent and water. Fig. 6 displays the
surface profiles with height (Fig. 6a) and modulus (Fig. 6b) skins for
the honeycomb films from the product 5b, respectively. Quantita-
tive nanomechanical analysis of the surfaces (Fig. 6b) reveals a clear
difference in the modulus along the edge of the pores (red region) as
compared to that along the main skeleton between the pores (yel-
low region). As seen in Fig. 6c, the moduli vary from ca. 100 MPa
at the edges of the pores to ca. 400 MPa at the regions between
the pores. The larger modulus on the main skeleton is in good
agreement with that found for TDMS cellulose films (see suppor-
ting information). Because the PeakForce QNM mode of AFM is not
applicable for liquids or waxes, the modulus of the PEG oligomers
(EG , EG13, and EG22) could not be measured for comparison. These
4
findings indicate that when honeycomb membranes are formed via
the “breath figures” method, the condensed water droplets seem
to interact with the hydrophilic groups in the macromolecule, i.e.,
PEG side chain and results in a preferential allocation of the softer
hydrophilic PEG chains around the pores.
Further evidence of the preferential allocation of the PEG groups
around the edges of the honeycomb pores was apparent from AFM
analysis of the honeycomb films in water. Fig. 7 shows AFM images
of the honeycomb membrane in air (Fig. 7A) and after the addition
of water (Fig. 7B). The image obtained using the liquid clearly shows
Fig. 5. A schematic diagram of the formation of honeycomb films: (a) exposure of a thin layer of polymer solution under humid air flow; (b) condensation of water droplets
onto the cooling surface of the polymer solution due to evaporation of the solvent; (c) rearrangement of the water droplets to form an ordered hexagonal template resulting
from the capillary force; (d) precipitation of the polymer at the water–solvent interface to form honeycomb films after complete evaporation of the solvent and water.

1
32
W.Z. Xu et al. / Carbohydrate Polymers 100 (2014) 126–134
Fig. 6. AFM images of honeycomb membrane from the synthesized product 5b. (a) 3D height skin; (b) 3D modulus skin; (c) modulus profile along the black line in (b). (For
interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
that the morphology of the membrane surface changes upon expo-
sure to water. As the polymer is not soluble in water, this observed
phenomenon is due to mobilization of the hydrophilic PEG side
chains rather than the hydrophobic skeleton.
As observed by Kontturi et al. (2011), molecular rearrangements
occurred during exposure of the regenerated cellulose film to water
due to its hydrophilicity. When water penetrates the amphiphilic
polymer chains, the hydrophilic part, i.e., PEG side chains, interact
with water to move to energetically more favorable positions and
conformations and the resulting film morphology should be differ-
ent from the initial state. During the casting of the membrane, the
hydrophilic PEG side chains were limited to interact with micro-
sized water droplets. When the formed honeycomb membrane was
exposed to water, the PEG side chains were no longer limited to be
around the pores; hence the PEG side chains extend out from the
regular pores. This “activation” or “mobilization” of the PEG side
chains upon exposure to water could be of significant importance
in the development and utilization of such materials as bioactive
films and membranes.
In addition to AFM, the formed honeycomb films from products
5a, 5b, and 5c were also examined with SEM. The images and pore
size distribution histograms are displayed in Fig. 8. It was found that
Fig. 7. AFM images of the honeycomb membrane from product 5b in air (a) and in water (b).

W.Z. Xu et al. / Carbohydrate Polymers 100 (2014) 126–134
133
Fig. 8. SEM images and pore size distribution histograms of honeycomb films of samples 5a, 5b and 5c cast at a concentration of 1% in toluene.
the films from products 5b and 5c had more narrow distribution
standard deviation: 0.10 and 0.09 ␮m, respectively) and smaller
honeycomb-patterned films. Directed self-assembly and prefer-
(
ential allocation of the hydrophilic PEG chains by interaction
with water was confirmed by AFM. Exposure of these films to
water revealed that the PEG side chains became “activated” or
“mobilized”. The length of the hydrophilic PEG chains showed a
significant effect on the pore size and the pore size distribution of
the formed honeycomb films. These materials illustrate the poten-
tial for the development of bioactive cellulosic materials wherein
immobilized biomolecules can be preferentially allocated to the
pore region of the honeycomb-patterned films and can be “acti-
vated” to interact with molecules in solution.
pore size (average pore diameter: 0.67 and 0.81 ␮m, respectively)
than that from 5a (standard deviation: 0.20 ␮m, average pore diam-
eter: 1.11 ␮m). Work is underway to investigate the effect of PEG
side chain length, degree of substitution, and casting conditions on
honeycomb film formation and properties.
4
. Conclusions
Poly(ethylene glycol) monoiodide of three different lengths
was synthesized and attached to the C3 position of the regios-
elective 2,6-TDMS cellulose to afford an amphiphilic polymer.
The free hydroxyl end of the PEG side chains was successfully
linked an amino acid, Fmoc–Gly–OH and the formed 3-
O-Fmoc–Gly–poly(ethylene glycol)–2,6-di-O-thexyldimethylsilyl
cellulose was cast using the “breath figures” method to form
Acknowledgment
The authors appreciate financial support by the SENTINEL bioac-
tive paper network.

1
34
W.Z. Xu et al. / Carbohydrate Polymers 100 (2014) 126–134
Appendix A. Supplementary data
Kontturi, E., Johansson, L.-S., & Laine, J. (2009). Cellulose decorated cavities on ultra-
thin films of PMMA. Soft Matter, 5(9), 1786–1788.
Kontturi, E., Suchy, M., Penttila, P., Jean, B., Pirkkalainen, K., Torkkeli, M., et al. (2011).
Amorphous characteristics of an ultrathin cellulose film. Biomacromolecules,
12(3), 770–777.
Koschella, A., Heinze, T., & Klemm, D. (2001). First synthesis of 3-O-functionalized
cellulose ethers via 2,6-di-O-protected silyl cellulose. Macromolecular Bioscience,
Supplementary data associated with this article can be
found, in the online version, at http://dx.doi.org/10.1016/j.carbpol.
2
012.12.076.
1
(1), 49–54.
References
Liu, P.-S., Chen, Q., Liu, X., Yuan, B., Wu, S.-S., Shen, J., et al. (2009). Grafting of zwit-
terion from cellulose membranes via ATRP for improving blood compatibility.
Biomacromolecules, 10(10), 2809–2816.
Ma, D., & McHugh, A. J. (2007). The interplay of phase inversion and membrane for-
mation in the drug release characteristics of a membrane-based delivery system.
Journal of Membrane Science, 298(1–2), 156–168.
Azzam, F., Heux, L., Putaux, J.-L., & Jean, B. (2010). Preparation by grafting onto,
characterization, and properties of thermally responsive polymer-decorated
cellulose nanocrystals. Biomacromolecules, 11(12), 3652–3659.
Bar-Nir, B. B.-A., & Kadla, J. F. (2009). Synthesis and structural characterization of 3-
O-ethylene glycol functionalized cellulose derivatives. Carbohydrate Polymers,
Nemoto, J., Uraki, Y., Kishimoto, T., Sano, Y., Funada, R., Obata, N., et al. (2005). Produc-
tion of mesoscopically patterned cellulose film. Bioresource Technology, 96(17),
76(1), 60–67.
1
955–1958.
Beattie, D., Wong, K. H., Williams, C., Poole-Warren, L. A., Davis, T. P.,
Barner-Kowollik, C., et al. (2006). Honeycomb-structured porous films from
polypyrrole-containing block copolymers prepared via RAFT polymerization as
a scaffold for cell growth. Biomacromolecules, 7(4), 1072–1082.
Peresin, M. S., Habibi, Y., Zoppe, J. O., Pawlak, J. J., & Rojas, O. J. (2010). Nanofiber
composites of polyvinyl alcohol and cellulose nanocrystals: Manufacture and
characterization. Biomacromolecules, 11(3), 674–681.
Petzold, K., Einfeldt, L., Gunther, W., Stein, A., & Klemm, D. (2001). Regioselective
Burch, R. M., Weitzberg, M., Blok, N., Muhlhauser, R., Martin, D., Farmer, S. G., et al.
1
functionalization of starch: Synthesis and H NMR characterization of 6-O-silyl
ethers. Biomacromolecules, 2(3), 965–969.
(
1991). N-(Fluorenyl-9-methoxycarbonyl) amino acids, a class of antiinflam-
matory agents with a different mechanism of action. Proceedings of the National
Academy of Sciences of the United States of America, 88(2), 355–359.
Risbud, M. V., & Bhonde, R. R. (2001). Suitability of cellulose molecular dialysis
membrane for bioartificial pancreas: In vitro biocompatibility studies. Journal
of Biomedical Materials Research, 54(3), 436–444.
Socrates, G. (2001). Infrared and Raman characteristic group frequencies: Tables and
charts. Chichester, England: John Wiley & Sons Ltd.
Stenzel, M. H. (2002). Formation of regular honeycomb-patterned porous film by
self-organization. Australian Journal of Chemistry, 55(4), 239–243.
Stenzel, M. H., Barner-Kowollik, C., & Davis, T. P. (2006). Formation of honeycomb-
structured, porous films via breath figures with different polymer architectures.
Journal of Polymer Science Part A: Polymer Chemistry, 44(8), 2363–2375.
Tanaka, M., Wong, A. P., Rehfeldt, F., Tutus, M., & Kaufmann, S. (2004). Selective
deposition of native cell membranes on biocompatible micropatterns. Journal of
the American Chemical Society, 126(10), 3257–3260.
Widawski, G., Rawiso, M., & Francois, B. (1994). Self-organized honeycomb mor-
phology of star-polymer polystyrene films. Nature, 369(6479), 387–389.
Wijnhoven, J. E. G. J., & Vos, W. L. (1998). Preparation of photonic crystals made of
air spheres in titania. Science, 281(5378), 802.
Wong, K., Hernández-Guerrero, M., Granville, A., Davis, T., Barner-Kowollik, C., &
Stenzel, M. (2006). Water-assisted formation of honeycomb structured porous
films. Journal of Porous Materials, 13(3), 213–223.
Wong, K. H., Davis, T. P., Barner-Kowollik, C., & Stenzel, M. H. (2007). Honeycomb
structured porous films from amphiphilic block copolymers prepared via RAFT
polymerization. Polymer, 48(17), 4950–4965.
Wu, Y., & Xu, J.-C. (2001). Synthesis of chiral peptide nucleic acids using Fmoc chem-
istry. Tetrahedron, 57(38), 8107–8113.
Yue, Z., & Cowie, J. M. G. (2002). Preparation and chiroptical properties of a regios-
electively substituted cellulose ether with PEO side chains. Macromolecules,
Canaria, C. A., Smith, J. O., Yu, C. J., Fraser, S. E., & Lansford, R. (2005). New syntheses for
1
1-(mercaptoundecyl)triethylene glycol and mercaptododecyltriethyleneoxy
biotin amide. Tetrahedron Letters, 46(28), 4813–4816.
Goetz, L., Foston, M., Mathew, A. P., Oksman, K., & Ragauskas, A. J. (2010). Poly(methyl
vinyl ether-co-maleic acid)–polyethylene glycol nanocomposites cross-linked
in situ with cellulose nanowhiskers. Biomacromolecules, 11(10), 2660–2666.
Heinze, T. (2009). Hot topics in polysaccharide chemistry – selected examples.
Macromolecular Symposia, 280(1), 15–27.
Heinze, T., Pfeifer, A., & Petzold, K. (2008). Functionalization pattern of tert-
butyldimethylsilyl cellulose evaluated by NMR spectroscopy. Bioresources, 3(1),
7
9–90.
Ifuku, S., & Kadla, J. F. (2008). Preparation of a thermosensitive highly regioselec-
tive cellulose/N-isopropylacrylamide copolymer through atom transfer radical
polymerization. Biomacromolecules, 9(11), 3308–3313.
Imada, M., Noda, S., Chutinan, A., Tokuda, T., Murata, M., & Sasaki, G. (1999). Coherent
two-dimensional lasing action in surface-emitting laser with triangular-lattice
photonic crystal structure. Applied Physics Letters, 75(3), 316–318.
Isogai, A., Saito, T., & Fukuzumi, H. (2011). TEMPO-oxidized cellulose nanofibers.
Nanoscale, 3(1), 71–85.
Kadla, J. F., Asfour, F. H., & Bar-Nir, B. (2007). Micropatterned thin film honey-
comb materials from regiospecifically modified cellulose. Biomacromolecules,
8
(1), 161–165.
Kasai, W., & Kondo, T. (2004). Fabrication of honeycomb-patterned cellulose films.
Macromolecular Bioscience, 4(1), 17–21.
Kondo, T., Koschella, A., Heublein, B., Klemm, D., & Heinze, T. (2008). Hydrogen
bond formation in regioselectively functionalized 3-mono-O-methyl cellulose.
Carbohydrate Research, 343(15), 2600–2604.
3
5(17), 6572–6577.