Cyclization in hyperbranched polymer syntheses: Characterization by MALDI-TOF mass spectrometry

Article abstract of DOI:10.1021/ja9739547

Matrix-assisted laser desorption/ionization (MALDI)-time-of-flight (TOF) mass spectrometry allowed direct determination of the extent of macrocycle formation that occurred during the polymerization and copolymerization of A₂B and A₄B

Full text of DOI:10.1021/ja9739547

10180
J. Am. Chem. Soc. 1998, 120, 10180-10186
Cyclization in Hyperbranched Polymer Syntheses: Characterization
by MALDI-TOF Mass Spectrometry
Jonathon K. Gooden, Michael L. Gross,* Anja Mueller, Andrei D. Stefanescu, and
Karen L. Wooley*
Contribution from the Department of Chemistry, Washington UniVersity, One Brookings DriVe,
St. Louis, Missouri 63130-4899
ReceiVed NoVember 18, 1997
Abstract: Matrix-assisted laser desorption/ionization (MALDI)-time-of-flight (TOF) mass spectrometry allowed
direct determination of the extent of macrocycle formation that occurred during the polymerization and
copolymerization of A2B and A4B monomers. Cyclization in hyperbranched pentafluorophenyl-terminated
poly(benzyl ether)s was indicated by the presence of ions 20 Da less than the masses of acyclic species, owing
to the loss of the HF chain ends during polymerization. This loss occurs by intramolecular nucleophilic
aromatic substitution of the benzyl oxide focal point functionality upon a pentafluorophenyl chain end.
Homopolymerizations and copolymerization of A2B and A4B monomers gave cyclic products in all cases, and
the extent of cyclization depended on counterion, reaction time, and reaction temperature. In the
copolymerization, product distributions revealed that larger proportions of A2B repeat units in the product led
to increased amounts of cyclic products.
Introduction
although cyclization has often been considered to be negligible,
it was the focus of a recent theoretical study.3 The extent of
cyclization may depend on the synthetic techniques employed
and should affect the structure/property relationships of the end
products.4 Comparison of the hyperbranched polymer molecular
weight calculated from end group analysis with the absolute
molecular weight gives an indication of the extent of cycliza-
tion,2c,d where increased amounts of cyclized products cor-
respond to decreased concentrations of the B focal point group
and increased apparent molecular weight values by end group
analysis. In one case,2c unique 1H NMR resonances for
macrocyclic and acyclic end groups allowed the overall extent
of cyclization in the polyetherification of flexible A2B monomer
units containing phenolic and alkyl bromide functional groups
to be determined. Isolation of the cyclized monomer was
possible through a combination of precipitation and chroma-
tography. However, the extent of cyclization that occurs at each
degree of polymerization during the preparation of hyper-
branched polymers has not been determined quantitatively, and
thus, it remains as an important but experimentally unanswered
question.5 Herein, we demonstrate the matrix-assisted laser
desorption/ionization (MALDI)-time-of-flight (TOF) mass
spectrometric evaluation of cyclization of hyperbranched poly-
mers, including the direct detection of acyclic and cyclic species
Highly branched polymers have attracted considerable atten-
tion as materials that possess unusual properties in comparison
to linear polymers.1 Dendritic macromolecules are prepared by
stepwise methodologies, involving polymerization of AxB
monomers with controlled growth occurring selectively at either
the A chain ends (divergent growth strategy) or at the B focal
point site (convergent growth strategy). In contrast, the
preparation of the less highly branched and polydisperse
hyperbranched polymers is achieved by the one-step polymer-
ization of AxB monomers, involving reaction of the A and B
functionalities with each other. As with the condensation
polymerization of AB monomers, reaction may occur intermo-
lecularly to give polymer growth, or it may occur intramolecu-
larly to give a macrocyclic species. However, unlike linear
polymerizations, where cyclization gives termination of growth
by removal of the two reactive chain ends, the hyperbranched
macrocyclic species bear a large number of remaining A chain
end functionalities (#A ) (x - 1)DPn) that are available for
further polymer growth. The formation of a macrocycle during
hyperbranched polymerizations consumes the one unique B focal
point functional group, but only when all B functionalities are
consumed will polymerization no longer proceed. Therefore,
macrocyclic species should continue to grow throughout the
polymerization, and hyperbranched polymers containing a
macrocyclic unit should exist at each degree of polymerization.
Detection of these macrocyclic units by non-mass-spectrometric
methods is not trivial, because the chemical linkage constituting
the macrocycle is the same as those units forming the polymer
backbone linkages, which result from intermolecular reactions,
and the molecular weight is not affected to a large extent by
ring formation.
(2) (a) Flory, P. J. Principles of Polymer Chemistry; Cornell University
Press: Ithaca, NY, 1953; pp 368, 370. (b) Miller, T. M.; Neenan, T. X.;
Kwock, E. W.; Stein, S. M. J. Am. Chem. Soc. 1993, 115, 356-7. (c) Percec,
V.; Chu, P.; Kawasumi, M. Macromolecules 1994, 27, 4441-53. (d) Chu,
F.; Hawker, C. J.; Pomery, P. J.; Hill, D. J. T. J. Polym. Sci. A: Polym.
Chem. 1997, 35, 1627-33. (e) Matyjaszewski, K.; Gaynor, S. G.; Kulfan,
A.; Podwika, M. Macromolecules 1997, 30, 5192-4. (f) Ho¨lter, D.; Burgath,
A.; Frey, H. Acta Polym. 1997, 48, 30-5. (g) Beginn, U.; Drohmann, C.;
Mo¨ller, M. Macromolecules 1997, 30, 4112-6.
(3) Cameron, C.; Fawcett, A. H.; Hetherington, C. R.; Mee, R. A. W.;
McBride, F. C. ACS Polym. Prepr. (Am. Chem. Soc., DiV. Polym. Chem.)
1997, 38 (1), 56-7.
(4) For example, an acyclic hexamer composed of A2B repeat units and
containing branching sites at 50% of the repeat units when cyclized gives
a calculated decrease to 33% degree of branching.
The possibility for macrocycle formation has been discussed
in several reports2 of hyperbranching polymerizations, and
(1) Hawker, C. J.; Malmstro¨m, E. E.; Frank, C. W.; Kampf, J. P. J. Am.
Chem. Soc. 1997, 119, 9903-4.
S0002-7863(97)03954-1 CCC: $15.00 © 1998 American Chemical Society
Published on Web 09/17/1998

Cyclization in Hyperbranched Polymer Syntheses
J. Am. Chem. Soc., Vol. 120, No. 39, 1998 10181
Scheme 1a
a Reaction of the A2B monomer 1 via intermolecular reactions gives the acyclic tetramer 2, containing five A groups and one B group. Reaction
of 1 via an intramolecular reaction consumes the B functionality and gives the cyclic tetramer 3, containing only four A groups.
produced during hyperbranching polymerizations of A2B and
A4B monomers. Additionally, we report the use of MALDI
mass spectrometry for the determination of copolymer product
distributions in hyperbranched copolymerizations.
tion gives termination of growth by removal of the two reactive
chain ends, the hyperbranched macrocyclic species bear a large
number of pentafluorophenyl (A) functionalities (#A ) DPn)
that remain available for further polymer growth; only when
all benzyl oxide (B) functionalities are consumed will poly-
merization no longer proceed. It is important to note that
because the benzyl-ether bond in the macrocycle is similar to
those formed upon intermolecular polymerization, it is indis-
tinguishable by spectroscopic techniques, such as IR and NMR.
Results and Discussion
To investigate the formation of macrocycles during hyper-
branched polymerizations, pentafluorophenyl-terminated hy-
perbranched poly(benzyl ether)s6 were prepared. The reaction
of the A2B monomer, 3,5-dipentafluorobenzyloxybenzyl alcohol
(1) proceeds via intermolecular nucleophilic aromatic substitu-
tion reactions upon the para positions of the pentafluorophenyl
groups to give a hyperbranched polymer (Scheme 1, shown only
as tetramer 2). The hyperbranched polymer has increasing
numbers of pentafluorophenyl chain ends (#A ) DPn + 1) as
the polymerization degree (DP) increases but maintains a single
benzylic alcohol at the focal point. Competing with polymer-
ization is intramolecular attack of the benzyl oxide upon the
para position of a pentafluorophenyl chain end group to yield
the hyperbranched cyclophane polymer (shown as the tetramer
3 in Scheme 1). Unlike linear polymerizations, where cycliza-
An alternate structural tool is MALDI-TOF mass spectrom-
etry, which has been developed7 for the direct analysis of
polymer molecular weights,8 molecular weight distributions,9
and end group compositions.10 One issue with MALDI of
(7) (a) Schriemer, D. C.; Li, L. Anal. Chem. 1997, 69, 4169-75. (b)
Schriemer, D. C.; Li, L. Anal. Chem. 1997, 69, 4176-83.
(8) (a) Tanaka, K.; Waki, H.; Ido, Y.; Akita, S.; Yoshida, Y.; Yoshida,
T. Rapid Commun. Mass Spectrom. 1988, 2 (8), 151-3. (b) Belu, A. M.;
DeSimone, J. M.; Linton, R. W.; Lange, G. W.; Friedman, R. M. J. Am.
Soc. Mass Spectrom. 1996, 7, 11-24.
(9) (a) Bahr, U.; Deppe, A.; Karas, M.; Hillenkamp, F.; Giessmann, U.
Anal. Chem. 1992, 64, 2866-9. (b) Ra¨der, H. J.; Spickermann, J.;
Kreyenschmidt, M.; Mu¨llen, K. Macromol. Chem. Phys. 1996, 197, 3285-
96. (c) Vitalini, D.; Mineo, P.; Scamporrino, E. Macromolecules 1997, 30,
5285-9.
(10) (a) Danis, P. O.; Karr, D. E.; Westmoreland, D. G.; Piton, M. C.;
Christie, D. I.; Clay, P. A.; Kable, S. H.; Gilbert, R. G. Macromolecules
1993, 26, 6684-5. (b) Montaudo, G.; Garozzo, D.; Montaudo, M. S.;
Puglisi, C.; Samperi, F. Macromolecules 1995, 28, 7983-9. (c) Lee, S.;
Winnik, M. A.; Whittal, R. M.; Li, L. Macromolecules 1996, 29, 3060-
72. (d) Weidner, St.; Ku¨hn, G. Rapid Commun. Mass Spectrom. 1996, 10,
942-6. (e) Pasch, H. Polym. Prepr. (Am. Chem. Soc., DiV. Polym. Chem.)
1996, 37 (1), 319-20. (f) Zammit, M. D.; Davis, T. P.; Haddleton, D. M.;
Suddaby, K. G. Macromolecules 1997, 30, 1915-20.
(5) Cyclization of the chain ends of poly(propyleneimine) dendrimers
has recently been reported by an elegant electrospray mass spectrometry
study. This cyclization was by a side reaction of nitrile chain ends during
hydrogenation and not a condensation reaction between focal point and
chain end during hyperbranched polymer growth: Hummelen, J. C.; van
Dongen, J. L. J.; Meijer, E. W. Chem. Eur. J. 1997, 3 (9), 1489-93.
(6) (a) Mueller, A.; Wooley, K. L. Polym. Prepr. (Am. Chem. Soc., DiV.
Polym. Chem.) 1997, 38 (1), 58-9. (b) Mueller, A.; Kowalewski, T.; Wooley,
K. L. Polym. Mater. Sci. Eng. 1997, 77, 89-90. (c) Mueller, A.;
Kowalewski, T.; Wooley, K. L. Macromolecules 1998, 31, 776-86.

10182 J. Am. Chem. Soc., Vol. 120, No. 39, 1998
Gooden et al.
Figure 2. MALDI-TOF mass spectrum of 4, cationized with silver in
trans-3-indoleacrylic acid matrix. Degrees of polymerization from 5
to 13 are included, and the cyclized products are observed as resolved
ion peaks at -20 m/z units from their acyclic analogues (see inset).
Figure 1. MALDI-TOF mass spectrum of 4, in a dithranol matrix.
The cationization is from adventitious sodium ions present in the matrix,
on the sample plate, or in the sample. The cyclized products were
cleanly resolved from the oligomers (see inset).
and a single experiment, we found 45% average cyclization after
11 h (Figure 2), which compares favorably with the results
reported above. After a shorter reaction time (8 h) under the
same reaction conditions,13 the percent cyclization was 25%.
The effects of branching upon the extent of cyclization were
investigated from the polymerization of an A4B monomer unit
5 (Chart 1),14 which is a second-generation dendrimer that has
an ideally branched layer built into each of the repeat units.
Two MALDI-TOF spectra (Figure 3) of samples of 6, obtained
from homopolymerizations of 5, demonstrate that the extent of
cyclization is affected by the reaction conditions, suggesting
again that cyclization occurs during polymerization and not
during sample preparation or in the mass spectrometer. The
product-size distributions also suggest that sample fractionation
or fragmentation is not significant during the mass spectrometry
experiment.
Although cyclization was observed for each of the oligomeric
products that are composed of 4 and 6, macrocyclic products
of the monomers 1 and 5 were not observed. The enthalpies
of cyclization, which were determined from semiempirical AM1
calculations, showed that cyclization is disfavored for 1 by 109
kJ/mol, favored for 5 by 15 kJ/mol, and favored for the dimer
of 5 by 46 kJ/mol (Table 1). Considering also the entropic
contributions (the product of entropy and temperature, ∼-17
to -26 kJ/mol at 298 K),15 only the dimer of 5 is expected to
form a stable macrocycle.
We might expect that the same trends for cyclization would
occur if the various monomers and oligomers were subjected
to mass spectrometric conditions that lead to postsource
decompositions (PSD) and collisionally activated decomposi-
tions (CAD). We were able to do a PSD/CAD experiment only
for the A4B monomer 5 but were unable to observe loss of HF
even under forcing conditions of high laser power and introduc-
tion of argon gas in the collision cell. These results are
consistent with those observed for 5 in solution and also suggest
that the elimination of HF by a mass spectrometric process is
unlikely.
synthetic polymers is choice of matrix. Dithranol was one of
the chosen matrixes, because it is commonly used for this class
of compounds. The MALDI-TOF spectrum (Figure 1) of
hyperbranched polymer 4 from polymerization of 1 is composed
of peaks that correspond to Na+-cationized oligomers. The
cyclized products are clearly observable as ions that are 20 m/z
units less than the masses of each main ion, owing to the loss
of HF (from one A plus B chain ends) upon cyclization. For
three oligomers with masses in a range where the mass resolving
power is sufficient to give baseline separation, the peaks
corresponding to HF loss were 47% of the area of the peak
corresponding to the acyclic oligomer. The standard deviation
(N ) 30) was 11% (absolute) for a product isolated after 11 h
of reaction time, indicating reasonable reproducibility.
There is a question of whether the loss of HF took place
during the synthesis or as a consequence of the mass spectral
ionization. To rule out a cyclization caused by mass spectrom-
etry, the laser power was increased by 40%, with the expectation
that the percent cyclization would remain constant. The new
areas of the peaks corresponding to loss of HF were 50% with
a standard deviation of 8% (absolute). The two percentages,
47 and 50%, are not significantly different at the 99% confidence
level as determined by the standard t test.
To obtain further confidence in the percent cyclization value
and in the conclusion that the loss of HF occurs in the synthesis,
additional MALDI experiments were conducted with the matrix
4-phenyl-R-cyanocinnamic acid (PCC). PCC is a matrix that
we had developed previously for detection of DNA bases that
had been modified with polycyclic aromatic hydrocarbons,11 and
it was chosen because it resembles the repeat unit of the
oligomer. The appearance and the signal-to-noise ratio of the
MALDI mass spectrum of the oligomer mixture were nearly
identical to those obtained with dithranol. The percent cycliza-
tion was determined to be 56% with a standard deviation of
9% (absolute), which is slightly greater (at 99% confidence)
than the 46% cyclization value reported above.
A third matrix that we used is trans-3-indoleacrylic acid, a
matrix that has found use for aromatic polyether dendrimers.12
Sensitivity comparable to that found with the other two matrixes
was achieved by spiking with silver trifluoroacetate, which led
to desorption of silver-cationized oligomers. With this matrix
(13) The polymerization conditions included reaction of 1 as a 0.5 M
solution in tetrahydrofuran heated at reflux in the presence of sodium for
8 h, followed by purification by precipitation into water and then flash
chromatography.
(14) Stefanescu, A. D.; Wooley, K. L. Polym. Mater. Sci. Eng. 1997,
77, 218-9.
(15) The lower limit of 17 kJ/mol is characteristic for the cyclization of
six- to eight-membered alkanes, whereas the 26 kJ/mol upper limit is
obtained from simulations using the rotational isomeric state model (RIS)
cf: Winnik, M. A. Chem. ReV. 1981, 81, 491.
(11) George, M.; Wellemans, J. M. Y.; Cerny, R. L.; Gross, M. L.; Li,
K.; Cavalieri, E. L. J. Am. Soc. Mass Spectrom. 1994, 5, 1021.
(12) Leon, J. W.; Fre´chet, J. M. J. Polym. Bull. 1995, 35, 449-55.

Cyclization in Hyperbranched Polymer Syntheses
J. Am. Chem. Soc., Vol. 120, No. 39, 1998 10183
Chart 1
Copolymerization of the A2B (1) and A4B (5) monomers
allowed for evaluation of the behavior of competing polymer-
ization/cyclization reactions during the preparation of copolymer
7. The copolymer product distributions (relative abundances
of 1 and 5) agree with the statistically expected distribution for
incorporation of 1 and 5. For example, the series that results
from a combined degree of polymerization of six (labeled in
Figure 4), contained mass peaks of relative intensities e0.1,
0.39, 0.62, 0.43, and e0.1 for oligomers composed of (A2B)5-
(A4B)1, (A2B)4(A4B)2, (A2B)3(A4B)3, (A2B)2(A4B)4, and (A2B)1-
(A4B)5, respectively.
Under the polymerization conditions used, the main products
were the cyclized hyperbranched polymers (Figure 4). The
relative amounts of 1 and 5 within the copolymer also affect
the amounts of cyclized products, in which cyclization is more
prevalent with increasing content of 1. The effects of comono-
mer content were evaluated as a function of both the degree of
polymerization and the total number of pentafluorophenyl (A)
groups present within the structure. For the pentamer series,
(A2B)3(A4B)2 gave a ratio of cyclic:acyclic ) 2.1:1, whereas
(A2B)2(A4B)3 gave a lower ratio of cyclic:acyclic ) 1.6:1. The
heptamer (A2B)4(A4B)3 and the pentamer (A2B)1(A4B)4 both
contain 14 pentafluorophenyl groups; however, ratios of cyclic
to acyclic structures were 3.2:1 and 1.9:1, respectively. Similar
data were found for other combinations throughout the MALDI-
TOF spectrum (Table 2).
Conclusions
The insight from MALDI-TOF MS analysis of hyperbranched
polymers will lead to a better understanding of the effects of
polymer structure upon the physical properties. For example,
although macrocycles occupy smaller hydrodynamic volumes
than their linear counterparts, it is not clear how these differences
will be manifested, if at all, in a comparison of hyperbranched
vs hyperbranched macrocyclic structures. Further investigations
of the extent of cyclization under different reaction conditions
and postpolymerization treatment are in progress.
Experimental Section
General Procedures. The MALDI-TOF experiments were
carried out on a PerSeptive Biosystems, Inc., Voyager-DE RP
mass spectrometer (Cambridge, MA). A nitrogen laser (337-
nm, 20-kW peak laser power, 2-ns pulse width) was used to
desorb the sample ions. The instrument was operated in linear
delayed-extraction mode with an accelerating potential of 25
kV. Raw data were acquired with a Tektronix 520A digitizing
oscilloscope before being transferred to a personal computer
equipped with GRAMS/386 software (Galactic Industries). Peak
areas were measured with the GRAMS software.
When necessary, the laser power was varied by controlling
an iris in the beam path. The PSD experiments were conducted
as previously described.16 CAD experiments were conducted
under the same conditions as PSD except argon gas was
introduced to an indicated pressure (measured on Bayard-Alpert
gauge) of 9 × 10-6 Torr in a collision cell that is located after
the ion source. Results from approximately 80 laser shots were
signal averaged to give one spectrum.
Figure 3. MALDI-TOF mass spectra of 6, prepared under different
polymerization conditions to result in different extents of cyclization:
(a) 90% cyclized from polymerization of 5 at 0.3 M concentration in
THF using sodium suspension (particle size <1 mm, 2.9 M in 1:3
toluene/THF) and heating at 55 °C for 0.5 h; (b) 15% cyclized from
polymerization of 5 at 0.3 M concentration in THF using neat NaK at
25 °C for 25 h. The cyclized products are observed as resolved ion
peaks at -20 m/z units from their acyclic analogues. As shown in the
inset, the monomer 5 does not cyclize and is observed as the silver
complex at 1212.1 Da.
(16) Gooden, J. K.; Byun, J.; Chen, L.; Rogan, E. G.; Cavalieri, E. L.;
Gross, M. L. Int. J. Mass Spectrom. Ion Processes 1997, 169/170, 241-9.

10184 J. Am. Chem. Soc., Vol. 120, No. 39, 1998
Gooden et al.
Table 1. Results of the Molecular Mechanics (MMFF) and Quantum Semiempirical (AM1) Calculations (Values in kJ/mol)
dimer of 5
calculation method
MMFF
energy
1
1 cyclic
2.44
5
5 cyclic
4.14
acyclic
cyclic
8.45
106.90
65.98
stretch
bend
torsion
VdW
1.55
13.64
12.05
4.52
9.33
110.62
26.15
11.96
82.10
51.00
11.80
-4.73
0.010
46.78
46.02
-28.07
-32.01
-36.65
-12.26
-79.91
out of plane
0.062
5.91
0.53
0.054
0.30
total (strain)
-0.72
70.37
62.58
60.84
133.89
101.71
AM1
∆Hf
-1968.0
-1548.7
-3853.9
-3558.5
-7437.9
-7173.5
∆∆Hf
-419.3
+108.7
-295.4
-15.2
-264.4
-46.2
∆Hr ) ∆∆Hf - 310.6
Analytes were dissolved in tetrahydrofuran (THF) at 10 mg/
mL, and a small volume (e.g., 10 µL) of that solution was added
to an equal volume of a saturated solution of the matrix in THF.
An aliquot of the resulting solution (1 µL) was deposited on
the sample plate surface and allowed to air-dry. Silver-doped
samples were prepared by using silver trifluoroacetate in THF
to give a concentration of 2-7 mM in the matrix/analyte
solution.
Size exclusion chromatography (SEC) was conducted on a
Hewlett-Packard series 1050 HPLC pump with a Hewlett-
Packard 1047A refractive index detector; data analysis was done
by Viscotek (Houston, TX) Trisec SEC Software, v. 3.0. Two
5-µm Polymer Laboratories PLgel columns (300 × 7.7 mm)
connected in series in order of increasing pore size (500 Å,
mixed bed C) were used with THF (distilled from CaH2) as
solvent. The Mw and Mw/Mn values for the samples 4, 6, and
7 were calculated from SEC results, based upon calibration with
PS standards.
Figure 4. MALDI-TOF mass spectrum of 7. The compositions of each
of the oligomers were assigned; for example, the series of hexamers
constructed from differing numbers of A2B (1) and A4B (5) monomer
m/z units is labeled (/). The cyclized products are observed as resolved
ion peaks at -20 Da from their acyclic analogues.
IR spectra were obtained on a Mattson Polaris spectrometer
as thin films on NaCl disks. 1H NMR spectra were recorded
on solutions in CDCl3 or toluene-d8 on a Varian Unity 300-
MHz spectrometer with the solvent proton signal as standard.
13C NMR spectra were recorded at 75 MHz on solutions in
CDCl3 or toluene-d8 on a Varian Unity 300 spectrometer or at
125 MHz on a Varian 500 spectrometer, with the solvent carbon
signal as standard. 19F NMR spectra were recorded at 282 MHz
on solutions in CDCl3 or toluene-d8 on a Varian Unity 300
spectrometer with external CFCl3 as standard.
Table 2. MALDI-TOF m/z and Relative Intensity Values for a
Sample of the Copolymer 7, Composed of Different Combined
Numbers of 4 and 6 and at Different Overall Degrees of
Polymerization
[cyclic species‚Ag]+
m/z
[acyclic species‚Ag]+
m/z
DPn
species
4
6
calcd found intens calcd found intens
dimers
trimers
0
2
1
0
3
2
1
0
4
3
2
1
5
4
3
2
1
5
4
3
2
5
4
3
2
1
2
3
1
2
3
4
1
2
3
4
1
2
3
4
5
2
3
4
5
3
4
5
2276.1 2277.4
2152.6 2153.4
2756.4 2757.1
3360.2 3361.7
2632.9 2633.6
3236.7 3237.6
3840.5 3841.4
4444.3 4445.1
3113.2 3113.5
3717.0 3717.7
4320.8 4321.7
4924.6 4925.0
0.73 2296.1 2296.9
0.47 2172.6 2172.8
1.00 2776.4 2777.0
0.43 3380.2 3381.6
0.37 2652.9 2652.8
0.84 3256.7 3257.4
0.70 3860.5 3861.2
0.18 4464.3 4464.8 e0.1
0.19 3133.2 3134.0 e0.1
0.65 3737.0 3737.0
0.73 4340.8 4341.5
0.32 4944.6 4945.3
0.67
0.52
0.72
0.32
0.29
0.45
0.53
A2B Monomer, 3,5-Bis(pentafluorobenzyloxy)benzyl Al-
cohol (1). To a solution of 3,5-dihydroxybenzyl alcohol (22.98
g, 0.1640 mol) and 2,3,4,5,6-pentafluorobenzyl bromide (90.13
g, 0.3453 mol) in dry THF (200 mL) were added potassium
carbonate (49.82 g, 0.361 mol) and 18-crown-6 (4.33 g, 0.0164
mol). The reaction mixture was heated to 55 °C and allowed
to stir under dry argon for 4 days. After cooling to room
temperature, most of the THF was removed at reduced pressure
leaving a pale yellow paste that was partitioned between CH2-
Cl2 (100 mL) and water (100 mL). The organic layer was
separated, and the aqueous layer was extracted with CH2Cl2 (3
× 50 mL). The organic layers were combined, dried over
MgSO4, and evaporated to dryness. The product was then
purified by flash chromatography on silica gel, eluting with 50%
hexane/CH2Cl2 and gradually increasing to CH2Cl2 to give 1
as a white crystalline solid: yield 58.42 g (71%); Tg -0.5 °C;
mp 94-95 °C; IR 3600-3100, 3100-2850, 1660, 1595, 1528,
1512, 1458, 1433, 1382, 1307, 1288, 1161, 1062, 973, 942, 830,
tetramers
pentamers
hexamers
0.30
0.44
0.17
3593.5 3593.7 e0.1 3613.5 3612.3 e0.1
4197.3 4197.5
4801.1 4801.6
5404.9 5404.9
0.39 4217.3 4217.4
0.62 4821.1 4822.1
0.43 5424.9 5425.7
0.24
0.26
0.30
6008.7 6009.5 e0.1 6028.7 6030.0 e0.1
heptamers
octamers
4677.6 4677.8
5281.4 5281.4
5885.2 5885.8
6489.0 6489.4
5761.7 5761.6
6365.5 6365.3
0.15 4697.6 4696.3 e0.1
0.43 5301.4 5301.7
0.43 5905.2 5906.4
0.15 6509.0 6510.5 e0.1
0.21 5781.7 5781.2 e0.1
0.22 6385.5 6385.8 e0.1
0.13
0.14
6969.3 6969.0 e0.1 6989.3
1
770, 685 cm-1; H NMR (CDCl3) δ 1.96 (t, 1H, J ) 6 Hz,
Three MALDI matrixes, dithranol,17 trans-3-indoleacrylic
acid (IAA), and 4-phenyl-R-cyanocinnamic acid (PCC), were
used. The former two were purchased from Aldrich (Milwau-
kee, WI), and the latter was synthesized as previously reported.12
ArCH2OH), 4.64 (d, 2H, J ) 6 Hz, ArCH2OH), 5.08 (s, 4H,
C6F5CH2OAr), 6.47 (t, 1H, J ) 2 Hz, ArH), 6.64 (d, 2H, J )
(17) Juhasz, P.; Costello, C. W. Rapid Commun. Mass Spectrom. 1993,
7, 343-51.

Cyclization in Hyperbranched Polymer Syntheses
J. Am. Chem. Soc., Vol. 120, No. 39, 1998 10185
2 Hz, ArH); 13C NMR (CDCl3) δ 57.5 (C6F5CH2O), 64.9
(ArCH2OH), 101.3, 106.3 (ArCH), 109.9 (m, C6F5 ipso C),
for C49H24F20O7 (1104.7): C, 53.28; H, 2.19; F, 34.39.
Found: C, 53.12; H, 2.32; F, 34.73.
137.6 (d of m, 1JC-F ) 249 Hz, ArCF), 141.8 (d of m, 1JC-F
)
A2B Hyperbranched Polyfluorinated Poly(benzyl ether)s
4. To a 0.5 M solution of 1 (10.00 g, 19.99 mmol) in dry THF
(40 mL) at room temperature under argon were added small
pieces of sodium (0.51 g, 22.00 mmol, 1.1 equiv). The reaction
was heated at reflux for 10.5 h and then quenched by
precipitation of the polymer into water. The product was
isolated by centrifugation and further purified by flash chro-
matography, eluting with 30% hexanes/CH2Cl2, and increasing
to CH2Cl2 to give 4 as a transparent glass: yield 8.47 g (88%,
SEC Mw ) 7,900, Mw/Mn ) 2.8); Tg 52 °C; IR 3100-2900,
1719, 1658, 1599, 1524, 1510, 1502, 1461, 1431, 1379, 1312,
256 Hz, ArCF), 144.0 (Ar ipso C), 145.7 (d of m, 1JC-F ) 249
Hz, ArCF), 159.4 (Ar ipso C); 19F NMR (CDCl3) δ -142.5
(m, 4F, o-F), -152.7 (t, 2F, J ) 21 Hz, p-F), -161.7 (m, 4F,
m-F). Anal. Calcd for C21H10F10O3 (500.29): C, 50.42; H,
2.01; F, 37.97. Found: C, 50.40; H, 2.23; F, 38.12.
3,5-Bis(pentafluorobenzyloxy)benzyl Bromide (8). Carbon
tetrabromide (13.32 g, 39.9 mmol) in dry THF (10 mL) was
quickly poured into a stirred solution of 1 (10.05 g, 20.08 mmol)
and triphenylphosphine (10.51 g, 40.1 mmol) in dry THF (50
mL) cooled to 0 °C (ice bath). The mixture was then allowed
to stir, maintaining the cooling, under an atmosphere of dry
argon for 2 min, after which time it became milky. Water (50
mL) was immediately added together with CH2Cl2 (100 mL).
The organic layer was separated, and the aqueous layer was
extracted with CH2Cl2 (3 × 50 mL). The organic extracts were
combined and washed three times with water (50 mL) before
being dried (MgSO4) and evaporated to dryness. The product
was then purified by flash chromatography on silica gel, eluting
with 50% hexane/CH2Cl2, and increasing to 15% hexane/CH2-
Cl2 to give 8 as a white crystalline solid: yield 9.53 g (85%);
Tg -13 °C; mp 112-113 °C; IR 3100-2850, 1659, 1595, 1525,
1508, 1462, 1434, 1382, 1308, 1290, 1169, 1062, 973, 940, 835,
1
1292, 1164, 1059, 1005, 977, 830, 770, 685 cm-1; H NMR
(toluene-d8) δ 4.5-4.8 (br, 4H, C6F4CH2O and C6F5CH2O),
4.8-5.0 (br, 2H, ArCH2O), 6.5 (br, 1H, ArH), 6.7 (br, 2H, ArH);
13C NMR (toluene-d8) δ 57.3 (C6F4CH2O), 57.6 (C6F5CH2O),
75.9 (ArCH2O), 102.5, 107.5 (ArCH), 109.3 (m, C6F5 ipso C),
110.1 (m, C6F4 ipso CCH2O), 137.4 (d of m, J ) 250 Hz,
ArCF), 137.8 (C6F4 ipso COCH2), 139.1 (m, Ar ipso C), 141.9
(d of m, J ) 250 Hz, ArCF), 145.9 (d of m, J ) 250 Hz, ArCF),
146.3 (d of m, J ) 250 Hz, ArCF), 160.0 and 160.1 (Ar ipso
C), assignments confirmed by 19F-decoupled 13C NMR at 125
MHz; 19F NMR (CDCl3) δ -143.6 (2F, C6F5 o-F), -144.8 (2F,
C6F4 o-F), -153.59 (1F, p-F), -156.94 (2F, C6F4 m-F),
-162.62 (2F, C6F5 m-F); average of 45% cyclized by MALDI-
TOF.
1
770, 685 cm-1; H NMR (CDCl3) δ 4.41 (s, 2H, ArCH2Br),
5.10 (s, 4H, C6F5CH2O), 6.49 (t, 1H, J ) 2 Hz, ArH), 6.67 (d,
2H, J ) 2 Hz, ArH); 13C NMR (CDCl3) δ 32.9 (ArCH2Br),
57.5 (C6F5CH2O), 102.1, 108.7 (ArCH), 109.7 (m, C6F5 ipso
C), 137.6 (d of m, 1JC-F ) 251 Hz, ArCF), 140.3 (Ar ipso C),
“Cyclized” A4B Hyperbranched Polyfluorinated Poly-
(benzyl ether)s (6). A suspension of sodium (particle size <0.1
mm, 2.90 M) in 1:3 toluene/THF (0.070 mL, 0.20 mmol) was
added under argon at room temperature to a stirred solution of
5 (0.110 g, 0.100 mmol) in dry THF (0.260 mL). The reaction
mixture was heated to 55 °C for 0.5 h and then quenched by
precipitation of the polymer into 10 mL of water. The product
was centrifuged, dried, redissolved in THF/hexanes (2:3), and
further purified by filtration through silica gel. After the
evaporation of the combined eluents, 6 was obtained as a
transparent glass: yield 0.071 g (65%, SEC Mw ) 9,800, Mw/
Mn ) 2.0); Tg 55 °C; IR 3100-2880, 1657, 1597, 1523, 1506,
1457, 1434, 1379, 1312, 1295, 1162, 1058, 1007, 976, 836, 734,
142.0 (d of m, 1JC-F ) 254 Hz, ArCF), 145.8 (d of m, 1JC-F
)
246 Hz, ArCF), 159.2 (Ar ipso C); 19F NMR (CDCl3) δ -143.3
(m, 4F, o-F), -153.4 (t, 2F, J ) 20 Hz, p-F), -162.4 (m, 4F,
m-F). Anal. Calcd for C21H9F10BrO2 (563.18): C, 44.79; H,
1.61; F, 33.73; Br, 14.19. Found: C, 44.72; H, 1.61; F, 33.56;
Br, 14.00. m/z 562.3, 564.3 [M]+, 483.3 ([M]+ - Br), 181.0
[C7H2F5]+.
A4B Monomer, 3,5-Bis[3,5-bis(pentafluorobenzyloxy)ben-
zyloxy]benzyl Alcohol (5). To a solution of 3,5-dihydroxy-
benzyl alcohol (0.18 g, 1.3 mmol) and 8 (1.53 g, 2.71 mmol)
in dry acetone (10 mL) were added potassium carbonate (0.39
g, 2.8 mmol) and 18-crown-6 (0.034 g, 0.12 mmol). The
mixture was heated at 55 °C with vigorous stirring under dry
argon for 30 h. Most of the acetone was then removed on a
rotary evaporator, leaving a pale yellow paste. Water (50 mL)
and CH2Cl2 (50 mL) were added. The organic layer was
separated, and the aqueous layer was extracted with CH2Cl2 (3
× 50 mL). The organic extracts were combined, dried over
MgSO4, and evaporated to dryness. Flash chromatography
using 10% hexane/CH2Cl2 and increasing to CH2Cl2 as eluent
gave 5 as a white solid: yield 0.99 g (69%); Tg 30 °C; IR 3600-
3200, 3100-2850, 1658, 1595, 1526, 1504, 1457, 1375, 1308,
1
690 cm-1; H NMR (CDCl3) δ 4.7 (s, 2H (6%), ArCH2OH),
5.0 (br, 4H, ArCH2OAr), 5.1 (br, 8H, OC6F4CH2O and
C6F5CH2O), 5.2 (br, 2H, (90%), ArCH2OC6F4), 6.5 (br, 3H,
ArH), 6.7 (br, 6H, ArH); 13C NMR (CDCl3) δ 57.5 (OC6F4CH2O,
C6F5CH2O), 69.7 (ArCH2OAr), 76.0 (ArCH2OC6F4), 101.7,
102.3, 106.8 (ArCH), 109.8, 109.9 (C6F5 ipso C, OC6F4 ipso
CCH2O), 137.5 (J ) 264 Hz, OC6F4), 139.7 (Ar ipso CCH2O),
159.3, 159.5, 159.9 (Ar ipso CO); 19F NMR (CDCl3) δ -142.7
(br, 2F (69%), C6F5 o-F), -144.6 (br, 2F (31%), C6F4 o-F),
-153.0 (br, 1F (69%), C6F5 p-F), -156.5 (br, 2F (31%), C6F4
m-F), -162.0 (br, 2F (69%), C6F5 m-F). Calculated Mn: 4300
(19F NMR); 9000 (1H NMR, relative to peak at 5.2 ppm); 17 500
(1H NMR, relative to peak at 4.7 ppm); average of 90% cyclized
by MALDI-TOF.
1
1294, 1164, 1062, 970, 942, 836, 740, 685 cm-1; H NMR
(CDCl3) δ 1.81 (t, 1H, J ) 6 Hz, ArCH2OH), 4.66 (d, 2H, J )
6 Hz, ArCH2OH), 4.99 (s, 4H, ArCH2O), 5.11 (s, 8H,
C6F5CH2O), 6.52 (t, 2H, J ) 2 Hz, ArH), 6.53 (t, 1H, J ) 2
Hz, ArH), 6.62 (d, 2H, J ) 2 Hz, ArH), 6.71 (d, 4H, J ) 2 Hz,
ArH); 13C NMR (CDCl3) δ 57.5 (C6F5CH2O), 65.2 (ArCH2-
OH), 69.5 (ArCH2O), 101.4, 101.5, 105.7, 106.7 (ArCH), 109.9
(m, C6F5 ipso C), 137.6 (d of m, 1JC-F ) 246 Hz, ArCF), 139.8
(Ar ipso C), 141.7 (d of m, 1JC-F ) 268 Hz, ArCF), 143.6 (Ar
“Acyclic” A4B Hyperbranched Polyfluorinated Poly-
(benzyl ether)s (6). To a stirred solution of 5 (0.386 g, 0.349
mmol) in dry THF (1.100 mL) was added a drop of Na-K
alloy (21 mg, 0.210 mmol Na, 0.415 mmol K) under Ar at room
temperature. After 25 h of stirring at room temperature, an
aliquot of 0.050 mL was carefully quenched under argon at 0
°C in 0.5 mL of moist THF, and then the polymer was
precipitated into 10 mL of water. After the usual workup, the
polymer was separated as a transparent glass (SEC Mw ) 2900,
Mw/Mn ) 1.4). Calculated Mn: 2400 (19F NMR); 2900 (1H
1
ipso C), 145.8 (d of m, JC-F ) 250 Hz, ArCF), 159.3, 159.9
(Ar ipso C); 19F NMR (CDCl3) δ -142.4 (m, 8F, o-F), -152.6
(t, 4F, J ) 18 Hz, p-F), -161.6 (m, 8F, m-F). Anal. Calcd

10186 J. Am. Chem. Soc., Vol. 120, No. 39, 1998
Gooden et al.
NMR, relative to peak at 5.2 ppm); 2700 (1H NMR, relative to
peak at 4.7 ppm); average of <15% cyclized by MALDI-TOF.
A2B and A4B Hyperbranched Polyfluorinated Copoly-
(benzyl ether) (7). A 0.290 M solution of sodium tert-butoxide
in THF (0.364 mL, 0.106 mmol) was added under Ar at room
temperature to a stirred solution of 1 (0.025 g, 0.050 mmol)
and 5 (0.056 g, 0.050 mmol) in dry THF (0.136 mL). The
reaction mixture was stirred at room temperature for 12 h, and
then the polymer was separated by quenching the solution with
water, followed by the usual workup, to yield copolymer 7 as
a transparent glass: yield 0.075 g (95%, SEC Mw ) 14 200,
1
Mw/Mn ) 2.7); Tg 30 °C; H NMR (CDCl3) δ 4.66 (br d, J )
6 Hz, ArCH2OH), 5.0 (br, ArCH2OAr), 5.1 (br, OC6F4CH2O
and C6F5CH2O), 5.2 (br, ArCH2OC6F4), 6.5 (br, ArH), 6.7 (br,
ArH); 13C NMR (CDCl3) δ 57.5 (OC6F4CH2O, C6F5CH2O), 67.7
(C6H3CH2OH), 69.7 (ArCH2OAr), 75.9, 76.2 (ArCH2OC6F4),
101.7, 102.5, 105.8, 106.9, 107.4 (ArCH), 109.9 (C6F5 ipso C,
OC6F4 ipso CCH2O), 137.5 (J ) 253 Hz, OC6F4), 138.3 (t, J )
31 Hz, C6F4 ipso COCH2) 139.7 (br, Ar ipso CCH2), 159.3,
159.5, 159.9 (Ar ipso CO); 19F NMR (CDCl3) δ -142.7 (br,
C6F5 o-F), -144.5 and -144.6 (br, C6F4 o-F), -152.9 (br, C6F5
p-F), -156.5 (br, C6F4 m-F), -162.0 (br, C6F5 m-F), ∼2% tert-
butoxide substitution upon the p-F positions detected by 1H and
19F NMR; average of 60% cyclized by MALDI-TOF.
Molecular Modeling. Molecular mechanics (MMFF force
field) and semiempirical (AM1 force field) calculations were
performed with the Spartan version 5.0 software (Wavefunction,
Inc., Irvine, CA). All structures were minimized in vacuo with
the MMFF module, followed by geometry optimization and
calculation of the heats of formation using the AM1 module.
The AM1 calculations could not be performed on the dimer of
5 owing to the large number of atoms; therefore, only the cyclic
and corresponding acyclic core fragments 11,c (cyclized) and
11,a (acyclized, where X ) F and H) were used in the
calculations. The remainders of the structures, consisting of
the spectator side arms, were replaced by H atoms (hydrogena-
tion). The formation energy of the entire dimer was derived
from the heats of formation of the cyclic and acyclic core
fragments (11) and detached side arms (9), and the correspond-
ing heat of hydrogenation calculated from similar fragments
from monomer 5.
estimation of the hydrogenation heats was done by using the
equation
∆∆Hhydrogenation,c ) ∆Hf9 + ∆Hf10,c - ∆HfH - ∆Hf5,c ) -
2
14.2 kJ/mol
In a similar fashion, the hydrogenation heat of the corresponding
acyclic species gave -13.3 kJ/mol.
An average hydrogenation heat ∆∆Hhydrogenation ) -13.8 kJ/
mol was used in the calculation of the formation heat of the
cyclic dimers of 5 according to the reaction
(5)2,c + 2H2 f 11,c + 2 × 9
∆Hf5,c ) 2∆Hf9 + ∆Hf11,c - 2∆∆Hhydrogenation - ∆HfH ) -
2
7173.5 kJ/mol
In a similar fashion, the formation heat of the acyclic dimer of
5 gave -7437.9 kJ/mol (see Table 1).
Acknowledgment. Financial support for this research from
the U.S. Army Research Office (Grants DAAH04-96-1-0158
and DAAG55-97-1-0183) (K.L.W.) and the National Institutes
of Health (Grant P41RR00954) (M.L.G.) is gratefully acknowl-
edged. The authors extend thanks to Professor Allan H. Fawcett
(The Queen’s University of Belfast, Northern Ireland, U.K.) for
motivating discussions, and to Mr. Christopher G. Clark, Jr.,
and Mr. Omar El-Ghazzawy (Washington University) for
assistance with the semiempirical calculations.
AM1 calculations of the heats of formation of the species
9-11 and those of HF (-310.6 kJ/mol) and H2 (-21.7 kJ/mol)
were used to derive, by thermochemical manipulation of the
data, the final cyclization heats of reaction and the heats of
formation corresponding to the cyclic and acyclic dimers of 5.
For the hydrogenation/fragmentation reaction of the cyclic
species
5,c + H2 f 10,c + 9
JA9739547