Synthesis and end group structure of 2-deoxydextrans prepared via the cationic ring-opening polymerization of an anhydro sugar

Article abstract of DOI:10.1021/ma012033k

1,6-Anhydro-3,4-di-O-benzyl-2-deoxy-β-D-glucose was polymerized using 1-chloroethyl isobutyl ether or cumyl chloride as initiator and ZnI₂ as activator. The resulting polymer contained predominantly α-(1,6) linkages, and the molecular weight distributions were narrow (≤ 1.25). The kinetics of polymerization was first-order with respect to monomer concentration. There was evidence of slow chain transfer at these polymerization temperatures, although the majority of polymer chains contained the initiator fragment. Mass spectra of the polymer revealed that the initiator fragment at the polymer chain end was an isobutyl group, presumably arising from nucleophilic attack of monomer on the methylene unit of the isobutyl group in the initiating cation with loss of acetaldehyde as opposed to attack at the Sp²-methylene carbon. Cyclic oligomers were present in the polymer sample as well and could have arisen from two processes: HCl initiation or a backbiting reaction.

Full text of DOI:10.1021/ma012033k

3402
Macromolecules 2002, 35, 3402-3412
Synthesis and End Group Structure of 2-Deoxydextrans Prepared via
the Cationic Ring-Opening Polymerization of an Anhydro Sugar
Yu -zen g Lia n g, An d r ea s H. F r a n z, Ch r isa n a New bu r y, Ca r lito B. Lebr illa , a n d
Tim oth y E. P a tten *
Department of Chemistry, University of California at Davis, One Shields Avenue,
Davis, California 95616-5295
Received November 20, 2001; Revised Manuscript Received February 15, 2002
ABSTRACT: 1,6-Anhydro-3,4-di-O-benzyl-2-deoxy-â-D-glucose was polymerized using 1-chloroethyl isobu-
tyl ether or cumyl chloride as initiator and ZnI2 as activator. The resulting polymer contained
predominantly R-(1,6) linkages, and the molecular weight distributions were narrow (e1.25). The kinetics
of polymerization was first-order with respect to monomer concentration. There was evidence of slow
chain transfer at these polymerization temperatures, although the majority of polymer chains contained
the initiator fragment. Mass spectra of the polymer revealed that the initiator fragment at the polymer
chain end was an isobutyl group, presumably arising from nucleophilic attack of monomer on the
methylene unit of the isobutyl group in the initiating cation with loss of acetaldehyde as opposed to
attack at the sp2-methylene carbon. Cyclic oligomers were present in the polymer sample as well and
could have arisen from two processes: HCl initiation or a backbiting reaction.
In tr od u ction
termination processes it is as synthetically useful as a
true living polymerization.7 Previous work has shown
that molecular weight control can be achieved in the
polymerization of 6,8-dioxabicyclo[3.2.1]octane (6,8-
DBO), the structural skeleton of 1,6-anhydro sugars.8
This monomer is similar to 1,6-anhydro-3,4-di-O-benzyl-
2-deoxy-â-D-glucose, monomer 1, which was first pre-
pared and polymerized by Hatanaka et al. using PF5
initiator.9 In this report we describe the synthesis of
linear polysaccharides by the cationic ring-opening
polymerization of monomer 1 using a different polym-
erization system. We report the extent of control over
the molecular weights, end group structure, and mo-
lecular weight distributions.
The chemical synthesis of polysaccharides via the
isolation and modification of polysaccharides from natu-
ral sources, stepwise total synthesis, step-growth po-
lymerization, or chain-growth polymerization has proven
to be a valuable tool for studying relationships between
carbohydrate structure and biological function. Chain-
growth polymerizations are the most viable option for
preparing synthetic, high molecular weight polysaccha-
rides, and two classes of sugar derivatives, anhydro
sugars and monosaccharide ortho esters, have emerged
as the best monomers for chain growth polymerizations
yielding polypyranoses and their derivatives.1,2 Since
the first synthesis of a stereoregular polysaccharide by
the cationic ring-opening polymerization of 1,6-anhydro
sugars,3 various derivatives of polysaccharides have
been prepared. Methods are known for controlling the
stereochemistry of linkages during these polymeriza-
tions, such as in the polymerization of 3,6-di-O-benzyl-
R-D-glucose-1,2,4-orthopivalate4 to yield cellulose and
1,6-anhydro-2,3,4-tri-O-benzyl-â-D-glucose3 to yield dex-
tran.
Choi et al. reported a block copolymerization of
anhydro sugar derivatives that displayed many of the
elements of a controlled polymerization: a linear in-
crease in molecular weights with conversion and block
copolymer formation.5 Even in light of this report, little
is yet known about how to control the molecular weight,
molecular weight distributions, and end group structure
of the products of the range of anhydro sugar and
monosaccharide ortho ester polymerizations. The con-
trolled/“living” cationic ring-opening polymerization of
heterocyclic monomers is known.6 When the rates of side
reactions, such as chain transfer and chain termination,
are sufficiently low, polymerization conditions may be
found under which well-defined polymers can be pre-
pared. In such cases, the polymerization can be termed
“controlled” to indicate that despite slow transfer or
Exp er im en ta l Section
Ma ter ia ls. Cumyl chloride was prepared by bubbling
anhydrous HCl through a solution of R-methylstyrene in CH2-
Cl2. The sample was dried over CaH2 and distilled under
vacuum. 1-Chloroethyl isobutyl ether was prepared by adding
isobutyl vinyl ether to the excess ether solution of HCl at 0 °C
before use. The excess HCl was removed by vacuum at 0 °C.
Dry ZnI2 (Aldrich) was obtained under a nitrogen atmosphere
and transferred into a drybox. CH2Cl2 was stirred over
concentrated H2SO4 for 24 h, extracted with 10% NaHCO3 and
deionized water, dried over anhydrous MgSO4, and distilled
from P2O5 before use. Benzene, acetonitrile, etc., were dried
and purified by conventional methods. Benzene was dried by
refluxing with Na for 5 h and distilled. Acetonitrile was dried
by refluxing with P2O5 (20 g/L) for 4 h and distilled. DMF was
dried by refluxing with CaH2 (10 g/L) for 8 h and distilled
before use. Unless specified otherwise, all other reagents were
purchased from commercial sources and used as received. All
reagents for the polymerizations were handled under a nitro-
gen atmosphere using standard drybox or Schlenk techniques.
Mea su r em en ts. Number-averaged molecular weights (Mn),
weight-averaged molecular weights (Mw), and molecular weight
distribution (Mw/Mn) were determined using gel-permeation
chromatography in THF at 30 °C and a flow rate of 1.00 mL
min-1. Three Polymer Standards Services columns (100 Å,
1000 Å, and linear) were connected in series to a Thermosepa-
ration Products P-1000 isocratic pump, autosampler, column
oven, and Knauer refractive index detector. Calibration was
performed using polystyrene samples (Polymer Standard
* Corresponding author: phone (530) 754-6181; Fax (530) 752-
8995; e-mail patten@indigo.ucdavis.edu.
10.1021/ma012033k CCC: $22.00 © 2002 American Chemical Society
Published on Web 03/30/2002

Macromolecules, Vol. 35, No. 9, 2002
2-Deoxydextrans 3403
Services; Mp ) 400-1 000 000; Mw/Mn < 1.10). 1H NMR
spectra (300 MHz) and 13C {H} spectra (75 MHz) were recorded
using a Varian Inova Mecury-300 NMR spectrometer. Chemi-
cal shifts (δ, ppm) were referenced to the residual proton or
carbon signal of the solvent. Melting points were measured
using a MEL-TEMP II instrument and are uncorrected.
Polymerization conversions were determined from 1H NMR
spectra by calculating the ratio of peak integrations for the
C-2 protons of the monomer (δ ) 1.90-2.10) and polymer
(δ ) 1.56-1.80 and δ ) 2.22-2.43).
Ma ss Sp ectr om etr y. Mass spectra were recorded on an
external source HiResMALDI (IonSpec Corp., Irvine, CA)
equipped with a 4.7 T magnet. The HiResMALDI was equipped
with an LSI 337 nm nitrogen laser. Mass spectra were also
recorded on a Proflex III MALDI-TOF (Bruker-Daltonics,
Billerica, MA) with a 337 nm nitrogen laser under the same
conditions. The matrix used was 2,5-dihydroxybenzoic acid (5
mg/100 µL in ethanol). A 0.01 M solution of NaCl in methanol
was used as dopant. The solution of the polysaccharide (1 µL)
was applied to the MALDI probe followed by sodium dopant
(1 µL) and matrix solution (1 µL). The sample was dried under
a stream of hot air and subjected to mass spectrometric
analysis. For the collision-induced dissociation (CID) experi-
ment, the appropriate isolation pulses were programmed
starting at 3 s after the initial ionization and followed by
sustained off-resonance irradiation (SORI) excitation at 6 s (1
s, 5 V base to peak, +1000 Hz off-resonance). At a background
pressure of 10-10 Torr, argon gas was administered through a
pulsed valve at 6 and 6.5 s (peak pressure 5 × 10-5 Torr). Final
excitation for detection was performed 12 s after the initial
laser pulse.
4.63 (1H, m, 5-H), 4.23 (1H, dd, 6-Hendo), 3.95 (1H, m, 2-H),
3.82(1H, dd, 6-Hexo), 2.21 and 2.12 (6H, 2s, 2 × CH3CO2-).
3,4-Di-O-a cetyl-1,6-a n h yd r o-2-d eoxy-â-D-glu cose, Com -
p ou n d 4.12 A solution of compound 3 (8.60 g, 24.0 mmol) and
R,R′-azobis(isobutyronitrile) (0.640 g, 3.80 mmol) in dry ben-
zene (500 mL) was degassed by bubbling N2 through the
solution for 20 min. Tri-n-butyltin hydride (14.0 g, 48.0 mmol)
was added to the solution, and the solution was heated at
reflux for 2 h. TLC (3:2 petroleum ether-ethyl acetate) showed
complete conversion of compound 3 (Rf ) 0.4) into compound
4 (Rf ) 0.3). The mixture was concentrated, and the residue
was dissolved in acetonitrile (200 mL). The solution was
washed with petroleum ether (3 × 200 mL) and then evapo-
rated to give compound 4 (4.70 g, 85% yield). This material
was used in the next step without further purification.1H NMR
(CDCl3): δ (ppm) 5.57 (1H, s, H-1), 4.86 (1H, ddd, H-3), 4.69
(1H, s, H-4), 4.58 (1H, m, H-5), 4.19 (1H, dd, H-6endo), 3.78 (1H,
dd, H-6exo), 2.2-2.1 (1H, m, H-2ax), 2.14 and 2.08 (6H, 2s, 2 ×
CH3CO2-), 1.82 (1H, d, H-2eq).
1,6-An h yd r o-2-d eoxy-â-D-glu cose, Com p ou n d 5. A solu-
tion of compound 4 (4.60 g, 20 mmol) in 10:10:1 CH3OH-H2O-
Et3N (500 mL) was stirred for 5 h at room temperature and
then concentrated. The residue was dried by repeated distil-
lation with absolute EtOH (at least five times) and then by
placing the residue under vacuum in the presence of P2O5 to
give compound 5 (2.90 g, 98% yield). This material was used
in the next step without further purification.
1,6-An h ydr o-3,4-di-O-ben zyl-2-deoxy-â-D-glu cose, Mon o-
m er 1. A solution of compound 5 (1.46 g, 10 mmol) in 40 mL
of dry DMF was cooled to 0 °C, and benzyl bromide (2.85 mL,
24 mmol) was added. Next, NaH (60% dispersion in mineral
oil, 1.60 g, 40.0 mmol) was added in portions with stirring at
0 °C. The mixture was stirred at room temperature for 4 h
and then cooled to 0 °C. Next, CH3OH (15 mL) was added to
decompose the excess NaH. Chloroform (30 mL) was then
added to the mixture, which was washed with deionized water
(3 × 40 mL). The residue was added to 20 mL of CH3OH, and
the mineral oil was removed by separating the precipitate. The
precipitate, monomer 1, was recrystallized from CH3OH to
yield pure 6 (2.30 g, 71% yield); mp ) 56 °C. 1H NMR
(CDCl3): δ (ppm) 1.90-2.10 (1H, dd, 2-Heq), 2.00-2.10 (1H,
dd, 2-Hax), 3.41 (1H, s, 4-H), 3.63-3.73 (1H, dd, 6-Hex), 4.20
(1H, dd, 6-Hendo), 3.75 (1H, s, 3-H), 4.45 (1H, d, 5-H), 5.70 (1H,
s, 1-Heq), 4.5-4.6 (4H, m, C6H5CH2-), 7.35 (10H, m, 2 ×
C6H5-). 13C {1H} NMR (CDCl3): δ (ppm) 33.5 (C-2), 64.7
(C-6), 71.8 (C-5), 72.5 (-CH2-C6H5), 74.2 (C-4), 76.1 (C-3),
127.8 and 128.2 (-C6H5), 138.2 (C-1 of -C6H5).
NOTE: The following synthesis is an amalgamation of parts
of three procedures. The adaptations were made to eliminate
the need for column chromatography in the purification steps,
thereby allowing for a higher throughput to the synthesis.
D-Glu ca l, Com p ou n d 1.10 A solution of tri-O-acetyl-D-glucal
(10.9 g, 40.0 mmol) in 10:10:1 CH3OH-H2O-Et3N (500 mL)
was stirred for 5 h at room temperature and then concentrated.
The residue was dried by repeated distillation with absolute
EtOH (at least five times) and then by placing the residue
under vacuum in the presence of P2O5 to give compound 1.
This material was used directly in the next step without
further purification.
1,6-An h yd r o-2-d eoxy-2-iod o-â-D-glu cose, Com p ou n d 2.
This compound was synthesized largely according to the
method of Tailler et al.11 The main steps were as follows:
compound 1 (9.66 g, 39.0 mmol) was treated with bis-tri-n-
butyltin oxide (19.1 g, 32.0 mmol) and activated, powdered 3
Å molecular sieves (16 g) in refluxing dry acetonitrile (400 mL)
for 3 h. The mixture was cooled to 5 °C under nitrogen, and
iodine (15.2 g, 60.0 mmol) was added in one portion. The dark
brown mixture was stirred for 15 min at 5 °C and then for 2
h at room temperature. TLC (1:1 toluene-acetone) showed the
complete conversion of compound 1 (Rf ) 0.14) into compound
2 (Rf ) 0.45). The mixture was filtered through Celite and
concentrated. Saturated, aqueous Na2S2O3 (200 mL) and then
hexane (200 mL) were added to the residue, and the biphasic
mixture was stirred vigorously for 3 h. Next, the aqueous phase
was continuously extracted with ethyl acetate for 8 h. The
extract was concentrated to give compound 2 (8.00 g, 78%
yield), which was used in the next step without further
P olym er iza tion . The typical polymerization procedure was
as follows: in an oven-dried Schlenk flask under nitrogen
atmosphere, a solution of monomer 1 in CH2Cl2 was prepared
using the appropriate quantities of reagents. The solution was
degassed by three cycles of freezing, evacuation, and thawing.
Sequentially, solutions of initiator in CH2Cl2 (prepared by
adding stoichiometric, dry HCl in ether to isobutyl vinyl ether)
and ZnI2 in diethyl ether were added using dry syringes. The
solution was stirred at a defined temperature. At predeter-
mined intervals, a sample of the polymerization solution was
taken using a dry syringe and quenched by addition to a
solution of Et3N/CH3OH (1:1 v/v) in 2 mL of THF. The samples
were used for both the GPC and NMR measurements. The
remainder of the polymerization solution was quenched using
Et3N/CH3OH, and the polymer was isolated by precipitating
three times from THF into CH3OH and drying under vacuum
to yield a white powder. The conversions estimated by 1H NMR
1
purification. H NMR [(CD3)2SO]: δ (ppm) 5.61 (1H, s, H-1),
5.52 (1H, d, OH-3), 5.20 (1H, d, OH-4), 4.42 (1H, m, H-5), 4.01
(1H, d, H-6endo), 3.94 (1H, m, H-3), 3.83 (1H, m, H-2), 3.52 (1H,
dd, H-6exo), 3.45 (1H, m, H-4).
1
were nearly equal to the yield of polymers. H NMR (CDCl3):
3,4-Di-O-a ce t yl-1,6-a n h yd r o-2-d e oxy-2-iod o-â-D-glu -
cose, Com p ou n d 3. Crude compound 2 (8.00 g, 30.0 mmol)
was treated overnight at room temperature with pyridine (24
mL) and acetic anhydride (16 mL). The mixture was cooled to
5 °C, treated with CH3OH (40 mL), and concentrated. The
residue was dissolved in 200 mL of ethyl acetate. The solution
was washed with water (3 × 200 mL) and concentrated to give
compound 3 (9.70 g, 92% yield). This material was used in the
next step without further purification. 1H NMR (CDCl3): δ
(ppm) 5.69 (1H, s, 1-H), 5.13 (1H, m, 3-H), 4.71 (1H, m, 4-H),
δ (ppm) 0.89 (broad, end group CH3), 1.58-1.73 (broad, H-2ax),
2.15-2.42 (broad, H-2eq), 3.17-3.80 (broad multilane signal,
H-4, H-5, and H-6), 3.84-4.05 (broad singlets, H-3), 4.50-4.70
and 5.02-5.08 (broad, -CH2-C6H5), 4.90-5.01 (broad, H-1),
7.12-7.41 (broad, 2 × -C6H5). 13C {1H} NMR (CDCl3): δ (ppm)
35.5 (C-2), 66.0 (C-6), 70.8 (C-5), 75.0 (-CH2- C6H5), 77.8 (C-
4), 78.2 (C-3), 127.2 and 128.6 (-C6H5), 138.8 (C-1 of -C6H5).
The cyclization of monomer 1 was conducted using the same
procedure as the polymerization of monomer 1 except that
isobutyl vinyl ether was omitted.

3404 Liang et al.
Macromolecules, Vol. 35, No. 9, 2002
Sch em e 1. Syn th etic Sequ en ce for th e Syn th esis of a n En d -F u n ction a lized 2-Deoxyd extr a n
Resu lts a n d Discu ssion
erized previously using the standard PF5 initiation
system for 1,6-anhydro sugars.9 The cationic ring-
opening polymerization of monomer 1 was carried out
using 1-chloroethyl isobutyl ether as the initiator and
ZnI2 as the Lewis acid activator under a dry nitrogen
atmosphere in CH2Cl2 at 13 and 25 °C (Scheme 1). The
polymerization was quenched using triethylamine/
methanol (1:1 v/v), and the resulting perbenzyl polysac-
charide was isolated by precipitation into methanol.
Initial work on using the conditions for the controlled
polymerization of 6,8-DBO for 1,6-anhydro sugars cen-
tered on 1,6-anhydro-2,3,4-tri-O-benzyl-â-D-glucose. Po-
lymerizations of this monomer using 1-chloroethyl
isobutyl ether as the initiator and either ZnI2 or SnCl4
as the Lewis acid activator were very slow at room
temperature, yielding only a few monomer additions
after several hours. We hypothesized that the presence
of the inductive electron-withdrawing benzyloxy group
at the 2-position destabilized the intermediate carbo-
cation to such an extent that the two Lewis acid
activators generated an insufficient concentration of
carbocations for a reasonably fast rate of polymerization.
One remedy for this situation was to remove the
2-substitutent from the monomer.
1
Figure 1 shows the H NMR and 13C NMR spectra of
monomer 1 and poly(1) obtained from experiment 6 in
Table 1. The observed spectra were consistent with
1
previous reports.9 In the H NMR spectra, the signals
for the C-1 proton were observed at δ ) 5.70 ppm for
the monomer and 4.98 ppm for the polymer. In the 13C
{1H} NMR spectra, the signals for C-1 were observed
at δ ) 98.6 ppm for the monomer and at 104 and 99.1
ppm for the polymer. The latter signals corresponded
to the stereoisomeric â and R configurations, respec-
Monomer 1 was synthesized from tri-O-acetyl-D-glucal
with high throughput using an amalgamation of three
literature procedures.10-12 This monomer was polym-
F igu r e 1. 1H NMR spectrum (A) and 13C {1H} NMR spectrum (B) of monomer 1. 1H NMR spectrum (C) and 13C {1H} NMR
spectrum (D) of IBVE-Poly(1) (CDCl3) obtained in experiment 6.

Macromolecules, Vol. 35, No. 9, 2002
2-Deoxydextrans 3405
Ta ble 1. Ca tion ic Rin g-Op en in g P olym er iza tion of 1 in CH2Cl2 F or All P olym er iza tion s [Mon om er -1]0 ) 2.06 M
initiator
[I]0 (mM) [ZnI2]0 (mM) time (h) temp (°C) % conva Mn (NMR)b Mn (GPC)c Mw/Mn [R]D25 (deg)d
none
entry
1
2
3
4
5
6
0
68
68
68
50
80
2.8
0
4
25
25
25
13
25
25
0
0
cumyl Cl
4
cumyl Cl
cumyl Cl
IBVE + HCl
IBVE + HCl
2.8
2.8
2.8
4.0
2.5
9
2.5
2
93
87
92
96
9600
9200
8700
8600
7700
6200
5400
6500
1.25
1.22
1.21
1.24
109.7
102.7
102.4
a
b
Determined by 1H NMR spectroscopy measured in CDCl3. Calculated by the comparison of the integration of the initiator proton
signal with that of the polymer backbone protons. c Estimated from GPC eluted with THF based on polystyrene standards. CHCl3 solution
d
(3 ) 30.5 g/L, 4 ) 8.0 g/L, 5 ) 10.0 g/L).
F igu r e 2. First-order kinetic plot (left) and a plot of the dependence of the number-averaged molecular weight and the molecular
weight distributions vs monomer conversion for the cumyl chloride/ZnI2 initiated polymerization of 1. Solvent ) CH2Cl2; temperature
) 13 °C; [1]0 ) 2.06 M, [cumyl chloride]0 ) 68 mM, [ZnI2]0 ) 2.8 mM.
tively, of the repeat unit, consistent with what had been
observed previously by Hatanaka et al.9 The relative
intensities of the two signals indicated that the majority
of the polymer was comprised of R (1 f 6) linkages
although the stereoregularity was not perfect, with some
â(1 f 6) linkages present. Confirmation of this conclu-
sion was gained through optical rotation measurements
on the polymers (Table 1). The optical rotations of the
polymers ranged from 104° to 110°, consistent with
previous work9 on low-temperature PF5-initiated polym-
erizations of this monomer and with the predominance
were consistently lower than the 1H NMR molecular
weights. This difference was most likely due to differing
hydrodynamic volumes between poly(1) and the poly-
1
styrene standards, so the H NMR molecular weights
probably better reflected the true molecular weights of
the polymer samples. The molecular weight distribution
of each sample was narrow and below 1.25; however, a
definite increase in the polydispersity index was ob-
served at the end of the polymerization. We noted that
if the solutions were allowed to age without quenching
and workup, then the number-averaged molecular
weight decreased with time and the molecular weight
distribution increased with time. This observation was
a clear indication that under the polymerization condi-
tions slow chain transfer can occur. The effect of chain
transfer was also observed when the initial monomer-
to-initiator ratio was varied (experiments 5 and 6) by a
factor of 1.6, and the resulting final molecular weights
essentially did not change.
The polymerization rate decreased substantially with
decreasing temperature (experiments 3 and 4), such
that below 0 °C the polymerization rate was too slow to
observe. This rate behavior was different from what was
observed for the corresponding polymerizations of 6,8-
DBO, the isoelectronic bicyclic acetal.8 In that case, the
polymerizations proceeded down to temperatures of -40
°C. Evidently, the substituents at C-3 and C-4 in
monomer 1 either exerted a measurable inductive
electron-withdrawing effect or contributed steric inter-
actions to the transition state for propagation (or a
combination of the two effects).
1
of R (1 f 6) linkages in the polymer. In the H NMR
spectrum of poly(1) a signal was observed at δ ) 0.90
ppm that was not due to resonances from any of the
protons of the polymer backbone, and it was assigned
to the methyl protons of the isobutyl end group. This
observation and assignment were consistent with previ-
ous work on the polymerization of 6,8-DBO8 and with
the conclusion that the initiation of the polymerization
occurred via activation of the initiator by ZnI2 and
subsequent addition of the resulting carbocation to
monomer 1. Other initiators, such as cumyl chloride and
1-phenylethyl bromide, were used successfully to initiate
the polymerization of monomer 1.
Table 1 shows the data for a set of polymerizations,
and the corresponding kinetic and molecular weight
plots for experiment 6 are shown in Figure 2. For those
polymerizations that did not contain either initiator or
ZnI2, no polymerization was observed (experiments 1
and 2). The semilogarithmic kinetic plot was linear up
to ∼90% conversion, indicating that the polymerization
rate was first-order with respect to monomer concentra-
tion and that the concentration of the propagating
centers was constant during the course of the polym-
erization. The molecular weight (Mn) of the polymer
samples increased with monomer conversion in a linear
fashion. The molecular weights were measured using
To analyze the structure of the polymer chains in
more detail, poly(1) was subjected to mass spectrometric
analysis. Two samples were studied: poly(1) prepared
using 1-chloroethyl isobutyl ether initiation, IBVE-Poly-
(1), and poly(1) prepared using HCl initiation, HCl-Poly-
(1). For the latter polymer, similar polymerization
conditions were used as in the preparation of IBVE-
1
both GPC and H NMR, and the GPC molecular weights

3406 Liang et al.
Macromolecules, Vol. 35, No. 9, 2002

Macromolecules, Vol. 35, No. 9, 2002
2-Deoxydextrans 3407
F igu r e 3. (a) MALDI-TOF spectrum of compound HCl-Poly(1). (b) Details of MALDI-TOF spectrum of compound HCl-Poly-
(1). (c) FT-ICR mass spectrum of compound HCl-Poly(1).
Ta ble 2. Abu n d a n t Ion s Obser ved in F T-ICR-MS
Exp er im en ts w ith HCl-P oly(1) a n d IBVE-P oly(1); Given
fragmentation often produces different spectra. The
similarity of the TOF and FT-ICR spectra supports the
conclusion that the cyclic oligomers were formed during
the synthesis and not the mass spectrometric analysis.
A similar cyclic trimer was reported for the polymeri-
zation of the 2-deoxy sugar, 1,6-anhydro-3-O-benzyl-2-
deoxy-4-O-(2′,3′,4′,6′-tetra-O-benzyl-R-D-glucopyranosyl)-
â-D-arobino-hexopyranose.13,14
Ar e m /z Va lu es of th e Ion s
HCl-Poly(1)
IBVE-Poly(1)
[M + Na]+
[M + K]+
[M + Na]+
[M + K]+
1001.451
1075.524
1327.610
1401.687
1653.774
1727.852
1979.942
2054.015
2306.092
2380.196
2632.271
2706.347
2958.310
3032.444
1327.698
1653.946
1980.190
2306.456
2632.866
1343.729
1669.962
1996.199
2322.510
1343.577
To study the fragmentation behavior of the cyclic
tetramer, we isolated the quasi-molecular ion at m/z
1328 and performed a collision-induced dissociation
(CID) experiment. During CID the isolated ion of
interest was subjected to multiple collisions with a bath
gas (i.e., argon) in the ion cyclotron resonance cell. In
the course of collisions, energy transfer occurred, result-
ing in fragmentation. However, a common low-energy
pathway is the dissociation of the weakly coordinated
sodium cation from the quasi-molecular ion. This latter
pathway leads to a general loss of ion abundance. Low-
quality spectra resulted from isolated high-mass ions
with low abundance. Therefore, the most abundant ion
at m/z 1328 was chosen as a representative example for
the higher oligomers. The isolated quasi-molecular ion
and the CID spectrum are shown in Figure 4. Abundant
fragment ions were observed at m/z 1017, 1001, 691,
675, and 365. Figure 4a shows the isolated quasi-
molecular ion and possible fragmentations of the cyclic
tetramer at the glycosidic bonds. Fragmentation be-
tween C1/O and C5/C6 gave rise to the ion at m/z 1017.
Fragmentation between O/C6 resulted in the fragment
ion at m/z 1001. Analogous fragmentation at the next
glycosidic bond resulted in fragment ions at m/z 691 and
675 as well as 365. Accurate masses of the observed
fragmentation ions are compiled in Table 3.
2322.992
2648.208
2722.246
Poly(1) except that the IBVE was omitted. This experi-
ment provided a sample presumably with a proton end
group and served as a control.
The MALDI-TOF spectrum of the sodium-doped HCl-
Poly(1) is shown in Figure 3. Abundant ions were
observed at m/z 1328, 1654, 1980, 2306, 2632, and 2958.
The peak at m/z 1328 corresponded to the [M + Na]+
quasi-molecular ion and the peak at m/z 1344 to the
[M + K]+ quasi-molecular ion. The difference in mass
between the observed ions was ∆m ) 326 u and was
consistent with the mass of the monomeric units of
which the polymer was composed. The mass of the most
abundant ion at m/z 1328 was 18 mass units less than
would be expected for the tetrameric open-chain prod-
uct, and there was no indication of an open-chain
product at m/z 1346 (theoretical mass) in the spectrum
(Figure 3b). Thus, the signal is consistent with a cyclic
tetrameric structure. The other abundant ions that were
observed corresponded to the cyclic pentamer and the
next higher analogues. To confirm the result from the
TOF analysis and to obtain more accurate masses, we
subjected HCl-Poly(1) to MALDI-FT-ICR mass spectral
analysis. The resulting FT-ICR mass spectrum is shown
in Figure 3c. Accurate masses for the observed ions are
compiled in Table 2. Again, only the cyclic oligomers
were observed. Because the time scale for analysis
differs between TOF and FT-ICR analyzers, in source
The MALDI-TOF spectrum of IBVE-Poly(1) is shown
in Figure 5a, and the detailed view of the spectrum
(Figure 5b) revealed three series of ions. The first series
of ions was observed at m/z 1001, 1328, 1654, 1980,
2306, 2632, and 2958, and the intensity of the signals
decreased with increasing mass. These ions were con-
sistent with sodium adducts of the cyclic tetramer,
pentamer, and higher analogues as discussed above. The
second and third series exhibited the standard bell curve
of intensities normally associated with a polymer mo-
lecular weight distribution. The second series was

3408 Liang et al.
Macromolecules, Vol. 35, No. 9, 2002
F igu r e 4. (a) Isolated ion at m/z 1328 from compound HCl-Poly(1). (b) CID spectrum of m/z 1328.
Ta ble 3. F r a gm en t Ion s Obser ved in th e CID Exp er im en t
w ith Sod iu m -Dop ed HCl-P oly(1); Given Ar e m /z Va lu es of
In the FT-ICR mass spectrum of IBVE-Poly(1), two
major series were observed (Figure 5d). The first series
corresponded to the [M + Na]+ signals of the cyclic
oligomers. The second series consisted of ions represent-
ing the sodium adducts of molecules with an isobutyl
group at C6 of the initiating chain and an -OH group
at C1 of the other terminus. Accurate masses for the
observed ions are compiled in Table 2. Interestingly, the
series of open-chain oligomers with an OH group on both
chain ends was not observed in the ICR cell as opposed
to the TOF analyzer. The ions in the ICR cell were
detected several seconds after the initial ionization, and
this amount of time is sufficient for gas-phase reactions
to occur. We can conclude that the open-chain oligomers
were a product of the polymerization reaction; however,
during residence time in the gas phase those products
apparently lost water to form further cyclic oligomers.
Some low abundance signals were observed that could
be assigned. For example, small signals at m/z 2011,
2337, and 2663, corresponded to [M + Na]+ of molecules
with an OH group at the C6 of the initiating chain and
an methoxy group at C1 of the other terminus (presum-
th e F r a gm en ts
[M + Na+]
isolated ion
1327.698
1017.478
1001.485
691.305
675.312
365.140
comprised of ions at m/z 1728, 2054, 2380, 2706, 3032,
33587, 3685, 4011, 4337, 4663, 4989, and 5315 and
corresponded to the sodium adducts of open-chain
oligomers with an isobutyl group at C6 of the initiating
chain end and an -OH group at C1 of the other
terminus. The third series of signals (smaller and not
labeled in Figure 5b) showed signals at m/z 1998, 2324,
2650, 2976, 3302, 3628, 3954, 4281, 4607, and 4933.
Details of this series, which was 18 units higher in mass
than the series of cyclic oligomers, are shown in Figure
5c. These ions represent the open-chain oligomers with
an OH group at both chain ends. The assignments for
each of the observed signals are shown in Scheme 2.

Macromolecules, Vol. 35, No. 9, 2002
2-Deoxydextrans 3409

3410 Liang et al.
Macromolecules, Vol. 35, No. 9, 2002
F igu r e 5. (a) MALDI-TOF spectrum of compound IBVE-Poly(1). (b) Details of MALDI-TOF spectrum of compound IBVE-
Poly(1). (c) Details of ions recorded around m/z 3685. (d) FT-ICR mass spectrum of compound IBVE-Poly(1).
ably due to reaction with methanol upon quenching),
and small signals at m/z 2414 and 2740 corresponded
with sodium adducts of oligomers with the complete
initiator fragment at C6 and an OMe at the C1-
terminus.
Possible initiation mechanisms that account for the
products observed are shown in Schemes 3 and 4. Only
the isobutyl group of the initiator was observed on the
initiating chain end of IBVE-Poly(1). Consequently, the
mechanism of initiation must involve nucleophilic attack
of monomer on the methylene unit of the isobutyl group
in the initiating cation with loss of acetaldehyde (Scheme
3) as opposed to attack at the sp2-methylene carbon.
Only cyclic oligomers were observed in the mass spec-
trum of HCl-Poly(1). A mechanism that accounts for this
fact involves initiation by HCl to form a hydroxyl group
at the initiating end of the polymer chain followed by
end biting of the propagating cation onto this end group

Macromolecules, Vol. 35, No. 9, 2002
2-Deoxydextrans 3411
Sch em e 2. Obser ved Abu n d a n t F T-ICR a n d TOF Ma ss Sp ectr a Sign a ls for IBVE-P oly(1) a n d Th eir Assign ed
Str u ctu r es
Sch em e 3. P r op osed In itia tion Mech a n ism d u r in g th e Syn th esis of IBVE-P oly(1) In volvin g th e In itia tor
F r a gm en t
Sch em e 4. P r op osed En d -Bitin g Mech a n ism d u r in g th e Syn th esis of HCl-P oly(1)
(Scheme 4). In the mass spectrum of IBVE-Poly(1) cyclic
oligomers were also observed and could have arisen from
two possible processes. First, residual HCl present from
the in situ initiator synthesis could initiate polymeri-
zation and yield cyclic oligomers via the process de-
scribed above. The presence of the series of chains with
hydroxyl groups at the initiating chain end is consistent
with this possibility. Second, backbiting of the propagat-
ing cation onto units along the polymer backbone could
also yield cyclic oligomers. This process would be
consistent with the “chain transfer” process inferred
from the polymerization experiments described in Table
1 and above. One would expect that at lower tempera-
tures the rate of backbiting would decrease relative to
propagation and that complete molecular weight control
could be achieved in these polymerizations. Thus, the
proper combination of Lewis acid and transfer groups
(i.e., halogen) needs to be found to permit such lower
temperature polymerizations.
The polymer could be deprotected to yield end-
functionalized 2-deoxydextran via hydrogenation of the
benzyl ether groups. The polymers were stirred in 1-to-1
(v/v) THF/CH3OH solvent with 10% Pd/C and a 60 psi
pressure of hydrogen for 2 h at room temperature. After
addition of a little deionized water, the mixture was
filtered and the filtrate was concentrated. The 2-deoxy-
dextran was isolated by precipitation into acetone and
removal of volatile materials under vacuum. The poly-
1
mers were soluble in water and DMSO, and H NMR
and 13C{1H} spectra showed complete removal of the
benzyl ether groups, within the limit of detection, and
retention of the end group signal at 0.9 ppm (Figure 6).

3412 Liang et al.
Macromolecules, Vol. 35, No. 9, 2002
F igu r e 6. 1H NMR spectrum (A) and 13C {1H} NMR spectrum (B) of debenzylated IBVE-Poly(1) (d6-DMSO) obtained in experiment
6.
Con clu sion
Ack n ow led gm en t. This work was funded by the
Center for Polymeric Interfaces and Macromolecular
Assemblies (NSF-DMR 9808677) for T.E.P. and by the
National Institutes of Health (GM 49077-09) for C.B.L.
In summary, 1,6-anhydro-3,4-di-O-benzyl-2-deoxy-â-
D-glucose was polymerized using 1-chloroethyl isobutyl
ether or cumyl chloride as initiator and ZnI2 as activa-
tor. The resulting polymer contained predominantly
R-(1,6) linkages, and the molecular weight distributions
were narrow (e1.25). The kinetics of polymerization was
first-order with respect to monomer concentration.
There was evidence of slow chain transfer at these
polymerization temperatures, although the majority of
polymer chains contained the initiator fragment. Mass
spectra of the polymer revealed that the initiator
fragment at the polymer chain end was an isobutyl
group, presumably arising from nucleophilic attack of
monomer on the methylene unit of the isobutyl group
in the initiating cation with loss of acetaldehyde as
opposed to attack at the sp2-methylene carbon. Cyclic
oligomers were present in the polymer sample as well
and could have arisen from two processes: HCl initia-
tion or a backbiting reaction. The cationic ring-opening
polymerization of monomer 1 can be considered a
controlled polymerization in the extent that end-func-
tionalized polysaccharides can be prepared using this
method. While this study presents one example of the
polymerization of an anhydro sugar with some molec-
ular weight control (in addition to a previous study), one
would expect that through the proper combination of
monomer structure, protecting groups, and Lewis acid
activators other anhydro sugars may be polymerized
with control over the molecular weights and end group
structures. With the discovery of more such examples,
anhydro sugars can be added to the macromolecular
synthetic toolbox for the construction of macromolecules
containing synthetic polysaccharide segments.
Refer en ces a n d Notes
(1) Schuerch, C. Adv. Carbohydr. Chem. 1981, 39, 157-212.
(2) Schuerch, C. Synthetic Polysaccharides. In Encyclopedia of
Polymer Science; J ohn Wiley and Sons: New York, 1985; Vol.
13, pp 147-162.
(3) Ruckel, E. R.; Schuerch, C. J . Org. Chem. 1966, 31, 2233-
2236.
(4) Nakatsubo, F.; Kamitakahara, H.; Hori, M. J . Am. Chem. Soc.
1996, 118, 1677-1681.
(5) Choi, Y. S.; Uryu, T.; Yoshida, T. Macromol. Chem. Phys.
1997, 198, 2875-2888.
(6) Kubisa, P. Cationic Polymerization of Heterocycles. In Cat-
ionic Polymerizations: Mechanisms, Synthesis, and Applica-
tions; Matyjaszewski, K., Ed.; Marcel Dekker: New York,
1996; pp 437-542.
(7) Matyjaszewski, K. J . Phys. Org. Chem. 1995, 8, 197-207.
(8) Torres, L. F.; Patten, T. E. Macromolecules 1999, 32, 6958-
6962.
(9) Hatanaka, K.; Kanazawa, S.; Uryu, T.; Matsuzaki, K. J .
Polym. Sci., Part A: Polym. Chem. 1984, 22, 1987-1996.
(10) Roth, W.; Pigman, W. D-Glucal and the Glycals. In Methods
in Carbohydrate Chemistry; Whistler, R. L., Wolfram, M. L.,
Eds.; Academic Press: New York, 1963; Vol. 2, pp 405-408.
(11) Tailler, D.; J acquinet, J .-C.; Noirot, A.-M.; Beau, J .-M. J .
Chem. Soc., Perkin Trans. 1 1992, 3163-3164.
(12) Leteux, C.; Veyrieres, A.; Robert, F. Carbohydr. Res. 1993,
242, 119-130.
(13) Kasuya, M. C.; Hatanaka, K. Tetrahedron Lett. 1998, 39,
9719-9722.
(14) Kasuya, M. C.; Hatanaka, K. Macromolecules 1999, 32,
2131-2136.
MA012033K