Multicyclic poly(ether sulfone)s of phloroglucinol forming branched and cross-linked architectures

Article abstract of DOI:10.1021/ma021569i

K₂CO₃-promoted polycondensations of phloroglucinol and 4,4′-difluorodiphenyl sulfone (DFDPS) in DMSO yielded mixtures of linear oligomers due to numerous side reactions. In contrast, polycondensations of silylated phloroglucinol under similar conditions were not plagued by such side reactions but involved various cyclization reactions with high efficiency. The MALDI-TOF mass spectra revealed the formation of various cyclic, bicyclic, and multicyclic species of complex structure up to 12000 Da. Therefore, cross-linking only occurred when a large excess (+30 mol %) of DFDPS was added. The resulting branched and cross-linked polymers mainly consisted of cyclic building blocks. Attempts to introduce sulfonic acid groups by alkylation of pendant OH groups with sultones gave only a moderate degree of substitution, due to the high degree of cyclization. For comparison two polycondensations of silylated 5-methylresorcinol were studied.

Full text of DOI:10.1021/ma021569i

Macromolecules 2003, 36, 4337-4344
4337
Multicyclic Poly(ether sulfone)s of Phloroglucinol Forming Branched
and Cross-Linked Architectures
Ha n s R. Kr ich eld or f,*,† Detlev F r itsch ,‡ La li Va k h ta n gish vili,† a n d Ger t Sch w a r z†
Institut fu¨r Technische und Makromolekulare Chemie, Bundesstrasse 45, 20146 Hamburg, Germany,
and GKSS Forschungszentrum Geesthacht GmbH, Max-Planck-Strasse, 21502 Geesthacht, Germany
Received October 9, 2002; Revised Manuscript Received April 11, 2003
ABSTRACT: K2CO3-promoted polycondensations of phloroglucinol and 4,4′-difluorodiphenyl sulfone
(DFDPS) in DMSO yielded mixtures of linear oligomers due to numerous side reactions. In contrast,
polycondensations of silylated phloroglucinol under similar conditions were not plagued by such side
reactions but involved various cyclization reactions with high efficiency. The MALDI-TOF mass spectra
revealed the formation of various cyclic, bicyclic, and multicyclic species of complex structure up to 12 000
Da. Therefore, cross-linking only occurred when a large excess (+30 mol %) of DFDPS was added. The
resulting branched and cross-linked polymers mainly consisted of cyclic building blocks. Attempts to
introduce sulfonic acid groups by alkylation of pendant OH groups with sultones gave only a moderate
degree of substitution, due to the high degree of cyclization. For comparison two polycondensations of
silylated 5-methylresorcinol were studied.
Sch em e 1
In tr od u ction
The original purpose of this work was synthesis and
functionalization of branched poly(ether sulfone)s based
on phloroglucinol (structure 1). Either free phloroglu-
cinol or silylated phloroglucinol should be polycondensed
with 4,4′-difluorodiphenyl sulfone, DFDPS (Scheme 1).
Finally, sulfonic acid groups should be introduced to
obtain polymers which might be used as acidic mem-
branes or catalysts. Yet, because of unexpected side
reactions and extensive cyclizations, the course of the
polycondensations became the main focus of this study.
Whereas in the classical theory of polycondensation1,2
cyclizations do not play a significant role, Stockmayer
and co-workers3,4 have demonstrated 5 decades ago that
cyclic oligomers are formed by “backbiting degradation”
in thermodynamically controlled polycondensations in-
volving efficient equilibration reactions. Quite recently
we have demonstrated5-7 for a variety of kinetically
controlled polycondensations, KCPs, (no equilibration
reactions), that cyclization competes with propagation
at any concentration and at any stage of the polymer-
ization process. K2CO3-promoted polycondensations of
free or silylated diphenols with DFDPS at temperatures
e 140 °C belong to the group of KCPs.7 The influence
of cyclization reactions on the structure and properties
of networks has also been discussed in recent theories
of network formation.8 However, cyclization was mainly
discussed as an origin of loops resulting from the high
concentration of functional groups in large branched
polymers shortly before or during the gelation process.
The results obtained from polycondensations of silylated
phloroglucinol in this work present another scenario.
For better understanding, the results obtained from
polycondensations of silylated 5-methylresorcinol, a
stereochemical equivalent of phloroglucinol, will be
discussed first.
chased from Aldrich Co. (Milwaukee, WI) and used as received.
4,4′-Difluorodiphenyl sulfone, DFDPS, was synthesized as
described in the literature.9 K2CO3 p.a. grade was purchased
from Merck KG (Darmstadt, Germany) and dried at 65 °C over
P4O10 in vacuo. N-Methylpyrrolidone (NMP, a gift of Bayer AG,
Leverkusen, Germany) was twice distilled over P4O10 in vacuo.
Exp er im en ta l Section
Ma ter ia ls. Hexamethyldisilazane, phloroglucinol, 5-meth-
ylresorcinol, fluorobenzene, and chlorosulfonic acid were pur-
† Institut fu¨r Technische und Makromolekulare Chemie.
‡ GKSS Forschungszentrum Geesthacht GmbH.
10.1021/ma021569i CCC: $25.00 © 2003 American Chemical Society
Published on Web 05/15/2003

4338 Kricheldorf et al.
Macromolecules, Vol. 36, No. 12, 2003
Ta ble 1. P olycon d en sa tion s of P h lor oglu cin ol a n d
DF DP S in DMSO/Xylen ea
Silyla tion of P h lor oglu cin ol. Phloroglucinol (0.3 mol) and
hexamethyldisilazane (0.5 mol) were refluxed in toluene for
20 h and concentrated in vacuo. The product was isolated by
distillation in a vacuum of 0.02 mbar. Yield: 82%. 1H NMR
(CDCl3/TMS): δ ) 0.25 (s, 27 H), 6.01 (s, 3 H) ppm.
b
expt
no.
molar excess of
DFDPS and K2CO3
yield
(%)
ηinh
(dL/g)
1
2
3
4
5
6
7
8
9
0
10
20
30
40
50
60
70
80
76
70
78
90
78
88
95
97
90
0.08
0.08
0.08
0.07
0.07
0.07
0.08
0.08
0.09
5-Methylresorcinol (0.2 mol) was silylated analogously with
1
0.3 mol hexamethyldisilazane. Yield: 91%. H NMR (CDCl3/
TMS): δ ) 0.29 (s, 18 H), 6.18 (m, 1 H), 6.32 (m, 2 H) ppm.
P olycon d en sa tion of Silyla ted 5-Meth ylr esor cin ol. (A)
With Equ im ola r Stoich iom etr y. Silylated resorcinol (20
mmol), DFDPS (20 mmol), K2CO3 (20 mmol), and dry NMP
(50 g) were weighed under dry nitrogen into a 150 mL three-
necked flask equipped with flat-blade stirrer, gas-inlet tube,
and drying tube. Under a slow stream of dry purified nitrogen,
the reaction mixture was stirred for 48 h at 140 °C. After
cooling, the reaction mixture was poured into water (500 mL),
and the precipitated polymer was filtered off, washed inten-
sively with warm water, and dried at 80 °C in vacuo. Yield:
86%. ηinh ) 0.43 dL/g (with c ) 2 g/L at 20 °C in CH2Cl2).
(B) With Excess DF DP S. Silylated resorcinol (20 mmol),
DFDPS (22 mmol), and K2CO3 (22 mmol) were polycondensed
as described above. Yield: 81%. ηinh ) 0.16 dL/g (with c ) 2
g/L in CH2Cl2).
P olycon d en sa tion s of P h lor oglu cin ol (Ta ble 1). Phlo-
roglucinol (20 mmol), DFDPS (20 mmol), and K2CO3 (21 mmol)
were weighed into a 150 mL three-necked flask equipped with
flat-blade stirrer and distillation head. DMSO (100 mL) and
xylene (30 mL) were added. The reaction mixture was stirred
for 6 h at a bath temperature of 170 °C, whereby most of the
xylene and liberated water were distilled off. Finally, the
reaction mixture was poured into water, and the precipitated
polymer was filtered off, intensively washed with warm water,
and dried at 80 °C in vacuo.
a
b
Time: 6 h. Bath temperature: 170 °C. Measured at 20°C
with c ) 2 g/L in CH2Cl2/TFA (volume ratio 4:1).
Ta ble 2. P olycon d en sa tion s of Silyla ted P h lor oglu cin ol
a n d DF DP S in NMP a t 140 °C
molar excess
of DFDPS
and K2CO3
a
expt
no.
time yield ηinh
Mn(VPO)b Mw(LS)c
(h)
(%) (dL/g)
(Da)
(Da)
1
2
3
4
5
6
7
8
0
0
24
48
24
48
24
48
24
48
79
81
81
82
93
94
0.15
0.15
0.17
0.17
0.19
0.20
1900
2000
49 000
65 000
10
10
20
20
30
30
2200
2500
2400
100 000
250 000
200 000
cross-linked
cross-linked
a
Measured at 20 °C with c ) 2 g/L in CH2Cl2/TFA (volume ratio
b
4:1). Measured in. c Measured.
P olycon d en sa tion s of Silyla ted P h lor oglu cin ol (Ta ble
2). Silylated phloroglucinol (20 mmol), DFDPS (20 mmol) K2-
CO3 (20 mmol), and dry NMP (25 g) were weighed under dry
nitrogen in a 150 mL three-necked flask equipped with flat-
blade stirrer, gas-inlet tube, and drying tube. The reaction
mixture was stirred at 140 °C for 48 h under a slow stream of
dry N2. After cooling, the reaction mixture was poured into
water, and the precipitated polymer was isolated by filtration,
intensively washed with water, and dried at 80 °C in vacuo.
P olycon d en sa t ion s via t h e P seu d o-H igh -Dilu t ion
Meth od (Ta ble 3). Silylated phloroglucinol (20 mmol) dis-
solved in dry NMP (50 mL) and DFDPS (20 mmol) dissolved
in dry NMP (50 mL) were added dropwise and simultaneously
to a stirred suspension of dry K2CO3 (20.2 mmol) in dry NMP
(50 mL) at 140-145 °C. The time required for complete
addition was 1 h. Stirring at 145 °C was continued for 3 h.
The cold reaction mixture was then precipitated into water,
and the precipitated polymer was filtered off, washed with
water, and dried at 80 °C in vacuo.
Ta ble 3. P olycon d en sa tion s of Silyla ted P h lor oglu cin ol
a n d DF DP S by th e P seu d o High Dilu tion Meth od a
b
expt
no.
molar excess
of DFDPS and K2CO3
yield
(%)
ηinh
(dL/g)
1
2
3
4
0
10
20
30
77
79
91
0.14
0.20
0.23
cross-linked
a
b
In NMP at 140 °C, total reaction time 24 h Measured at 20
°C with c ) 2 g/L in CH2Cl2/TFA (volume ratio 4:1)
performed with a SK-18V6 column combination of Polymer
Standard Service GmbH having an inner diameter of 4.6 mm.
DMF containing NH4 acetate served as eluent. A combina-
tion of a RI detector (J asco, J apan) and a viscosity detector
(WGE, Dr. Bures GmbH) was used for the evaluation via
universal calibration based on polystyrene standards (WING-
PC software 6.01 of PSS was also used).
Alk yla tion w ith Su lton es. Silylated phloroglucinol (10
mmol), DFDPS (11 or 12 mmol), K2CO3 (11 or 12 mmol), and
dry NMP (60 g) were weighed into a 150 mL three-necked flask
and polycondensed at 140 °C for 24 h. After the temperature
was lowered to 100 °C, 1,4-butane sultone (11 mmol) and K2-
CO3 (11 mmol) were added, and the reaction mixture was
stirred at 100 °C for 24 h. The resulting slurry was then poured
into water, and the precipitated polymer was filtered off,
washed with water, and dried at 80 °C in vacuo. The dry
polymer was dissolved in a mixture of CH2Cl2 and trifluoro-
acetic acid (1:1 by volume) and precipitated into methanol.
Mea su r em en ts. The inherent viscosities were measured
with an automated Ubbelohde viscometer thermostated at 20
°C (Tables 1-3) or 25 °C (Table 4).
The MALDI-TOF mass spectra were recorded with a
Bruker Biflex III equipped with a nitrogen laser in the
reflectron mode. An acceleration voltage of 20 kV was used,
and a cutoff range of 500, 1000, or 3000 Da. The irradiation
targets were prepared from chloroform solution, with dithranol
as matrix and K-trifluoroacetate as dopant.
The light-scattering measurements were performed with a
triple-angle light-scattering detector “mini DAWN” from Wyatt
Technology in combination with ASTRA 4.70 chromatography
software.
Resu lts a n d Discu ssion
P olycon d en sa tion s of 5-Meth ylr esor cin ol. Sily-
lated 5-methylresorcinol was polycondensed with DFDPS
in N-methylpyrrolidone at 140 °C over a period of 48 h.
These reaction conditions were found in a previous
study7 dealing with silylated bisphenol A to allow for
nearly quantitative conversion and to avoid side reac-
tions and equilibrations almost completely. Two differ-
ent feed ratios were used for comparison with analogous
polycondensations of phloroglucinol. The first experi-
ment was performed with exactly equimolar amounts
of both monomers, and a poly(ether sulfone), PES, with
an inherent viscosity of 0.43 dL/g was isolated. For the
second polycondensation, a 10 mol % excess of DFDPS
was used, and the resulting PES had the low viscosity
The VPO measurements were conducted in DMF at 95 °C
with an “OSMOMATE 070” apparatus of Genotec GmbH.
Benzil was used for calibration. The SEC measurements were

Macromolecules, Vol. 36, No. 12, 2003
Multicyclic Poly(ether sulfone)s 4339
Ta ble 4. Ma sses of Cyclic Sp ecies (In clu d in g On e K Ion ) Id en tified in th e MALDI-TOF Ma ss Sp ectr u m of F igu r es 4 a n d
5 (P ES No. 1, Ta ble 3)
deg of
polymerization
of the cyclesa
C
B1C
B2C
B3C
B4C
B5C
B6C
B7C
B8C
2
3
4
5
6
7
8
9
719.8
1060.2
1400.5
1740.9
2081.2
2421.6
2761.9
3102.3
1224.2
1614.8
1955.1
2295.5
2635.8
2976.2
3316.6
3656.9
3997.3
4337.6
4877.9
2169.3
2509.7
2850.1
3190.4
3530.8
3871.5
4211.5
4551.9
4892.3
5232.6
5572.9
5913.2
6255.6
3064.3
3404.5
3745.0
4085.5
4425.8
4766.1
5106.4
5446.8
5786.1
6126.4
6466.7
6807.0
7147.3
7487.6
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
4299.5
4639.8
4980.2
5320.5
5660.9
6001.3
6341.6
6681.9
7022.2
7362.5
7072.0
8042.0
8382.3
8722.6
9063.0
5534.8
5875.1
6215.5
6555.8
6896.1
7236.4
7576.8
7917.1
8257.4
8597.8
8938.1
9278.4
9618.8
9959.1
10299.4
6770.2
7110.5
7451.3
7791.7
8132.0
8472.3
8812.6
9153.0
9493.3
9833.6
10174.0
10414.3
10754.6
11095.0
11435.3
8346.2
8686.5
9026.9
9367.2
9707.5
10047.9
10388.2
10728.5
11068.9
11408.2
11748.6
10262.1
10602.5
10942.8
11283.2
11623.5
a
This DP includes OH-terminated side chains, such as those of BC3′ in Scheme 1 or B2C8 and B3C11 in Scheme 2.
Sch em e 2
of 0.16 dL/g as expected of a significant imbalance of
the stoichiometry.
The MALDI-TOF mass spectrum (MS) of the high
molar mass sample displayed the peaks of cyclic oligo-
mers and polymers (2C in Scheme 1) as the predomi-
nant species up to 10 000 Da (Figure 1). At higher
masses, linear chains of structure 2La or 2Lb (Scheme
2) were prevailing. The absence of 2Lc chains together
with the presence of 2Lb chains indicates that the
stoichiometry in the reaction mixture was not perfect,
possibly due to a minor side reaction of DFDPS. This
observation agrees well with our previous results con-
cerning bisphenol A,7 where the most perfect stoichi-
ometry, the highest molecular weights, and the largest
extent of cyclization were obtained with a slight excess
(∼ 1 mol %) of DFDPS in the feed. The MALDI-TOF
MS of the low molar mass sample revealed again the
formation of cycles, but only the peaks of the trimer and
tetramer (2C3 and 2C4) were predominant and the
peaks of larger cycles vanished above 4000 Da. As
expected from the excess of DFDPS in the feed, the 2Lc
chains were now the predominant species above 1500
Da (Figure 2).
In summary, the MS showed the pattern of reaction
products expected for a normal kinetically controlled
polycondensation with one exception. The peak of the
cyclic dimer was barely detectable in contrast to analo-
gous polycondensations of silylated bisphenol A or
silylated 4-tert-butylcatechol (to be reported in a future
publication). Obviously, the cyclization of the linear
dimer is for sterical and entropical reasons unvaforable.
condition on the onset of gelation. The polycondensation
of free diphenols in DMSO combined with an aromatic
solvent for the azeotropic removal of water is the
standard method for the preparation of PES, which is
also used for the technical production of PES. When
P olycon d en sa tion s of F r ee P h lor oglu cin ol. For
the preparation of branched PES derived from phloro-
glucinol (1), two synthetic methods were compared
(Scheme 1) to study the influence of the reaction

4340 Kricheldorf et al.
Macromolecules, Vol. 36, No. 12, 2003
F igu r e 2. MALDI-TOF mass spectrum of the PES prepared
from silylated 5-methylresorcinol with a 10 mol % excess of
DFDPS.
F igu r e 1. MALDI-TOF mass spectrum of the PES prepared
from equimolar quantities of silylated 5-methylresorcinol and
DFDPS.
polycondensation of phloroglucinols in DMSO were
conducted in the presence of toluene, only low molar
masses were obtained (ηinh < 0.1 dL/g). Therefore,
toluene was replaced by xylene, hoping that a higher
reaction temperature (see Table 1) will result in higher
conversions and higher molar masses. However, the
viscosities listed in Table 1 demonstrate that this
modification of the reaction conditions was successless.
Yet, the most surprising result was that increasing
excess of DFDPS did neither enhance the molecular
weights nor generate cross-links. The MALDI-TOF MS
of all samples listed in Table 1 (examplarily demon-
strated in Figure 3) displayed the same peaks, only the
intensity ratios varied. The expected linear or cyclic
species were not found, and a comparison of the masses
summarized in Figure 3 with those listed in Table 4 for
PESs prepared from silylated phloroglucinol did not
show any overlapping. In other words, the successless
polycondensations of free phloroglucinol in DMSO were
a consequence of intensive side reactions, which were
not analyzed in detail.
P olycon d en sa tion s of Silyla ted P h lor oglu cin ol.
Whereas polycondensations of bisphenol A and silylated
bisphenol A with DFDPS gave nearly identical results,
the present work proved that polycondensations of free
and silylated phloroglucinol take a completely different
course. The molecular weights of PES samples prepared
from silylated phloroglucinol were about three times
higher and gelation occurred at certain feed ratios
(Table 2). Surprising was in this case that no cross-
linking took place at a 1.0:1.0 feed ratio despite almost
100% conversion of the C-F groups (as evidenced by
F igu r e 3. MALDI-TOF mass spectrum of a PES prepared
from phloroglucinol with a 30 mol % excess of DFDPS (no. 4,
Table 1).
MALDI-TOF MS). Even an excess of 10 or 20 mol % of
DFDPS did not suffice for gelation. When four polycon-
densations were repeated using the pseudo-high-dilu-
tion method, the results were similar to those obtained
at high concentration (Table 3). Obviously, the conden-
sation steps were too slow at 140 °C, so that the
concentration increased upon monomer addition over a
period of 1 h. Again no cross-linking occurred with an
excess of 10 or 20 mol % of DFDPS, but the increase of
the molecular weight was steeper. The MALDI-TOF
MS of the PES samples listed in Table 2 differed from
those of Table 3 insofar as weak peaks of noncyclic
species were detectable up to masses above 5000 Da,

Macromolecules, Vol. 36, No. 12, 2003
Multicyclic Poly(ether sulfone)s 4341
F igu r e 5. MALDI-TOF mass spectrum (higher mass spec-
trum) of a PES prepared from a silylated phloroglucinol with
a feed ratio of 1.0:1.0 via the “pseudo-high-dilution method”
(no. 1, Table 3).
F igu r e 4. MALDI-TOF mass spectrum (lower mass range)
of a PES prepared from silylated phloroglucinol with a feed
ratio of 1.0:1.0 via the “pseudo-high-dilution method” (no. 1,
Table 3).
Sch em e 3
whereas such peaks were almost absent in the MS of
the products listed in Table 3. All mass peaks of cyclic
species in both series of PES samples were identical;
only the intensity ratio varied as discussed below.
The MALDI-TOF MS of sample no. 1, Table 3, had
the best signal-to-noise ratio of all samples (mass peaks
up to 12 000 Da were detectable), and therefore, this
spectrum is exemplarily discussed in more detail. All
mass peaks were assigned (Table 4) and Figures 4 and
5 display 92 assignments. Only two weak peaks of linear
species, namely the dimer and trimer of structure 1Lb
were found (see Lb in Figure 4). All other mass peaks
originated from species containing at least one cycle.
Since no C-F terminated species were detected in the
MS of Figures 4 and 5, these MS prove an almost
quantitative conversion of the C-F groups. Therefore,
the isomeric structures presented in Schemes 3 and 4
are exclusively based on OH functionalities. The role of
diphenyl sulfone units is 3-fold. First, they form the
linear or cyclic repeating units, second, they connect to
cycles (intermolecular bridge) or they form a bridge
across a cycle (intramolecular bridge). Therefore, the
nomenclature BxCY was selected to count the total
number of “bridge units” (x) and the total number of
repeating units (Y). All the mass peaks assigned in this
way are listed in Table 4.
Interestingly, the peaks of the cyclic dimer (C2) and
of the bicyclic dimer BC2 (bridged cyclic dimer, Scheme
3) were barely detectable. This finding agrees with the
low concentration of the cyclic dimer in the polyconden-
sations of silylated 5-methylresorcinol. Therefore, it
needs to be emphasized that the peak of BC3, the
bridged cyclic trimer (see Scheme 3), was clearly detect-
able in all MS The peak intensity of BC3 was relatively

4342 Kricheldorf et al.
Sch em e 4
Macromolecules, Vol. 36, No. 12, 2003
F igu r e 6. MALDI-TOF mass spectrum of a PES prepared
from silylated phloroglucinol with a 1.0:1.2 feed ratio (no. 5,
Table 2).
Sch em e 5
silylated 5-methylresorcinol and phloroglucinol. In the
former case, the excess of DFDPS considerably reduces
the extent of cyclization and yields chains terminated
by two C-F groups, whereas in the case of phloroglu-
cinol the excess of DFDPS intensifies ring closing
reactions favoring the formation of bridged cycles.
For the proper understanding of the entire polymer-
ization process, it should be emphasized that formation
of small cycles results from two different processes. The
first one is the direct formation of cycles from end-to-
end reactions of oligomers. Most of these cycles was
detected up to C9 in the MALDI-TOF MS (Table 4 and
Figures 4 and 5). The second process yielding cyclic
building blocks is the “backbiting” of C-F chain ends
as illustrated in Scheme 6. This kind of “backbiting”
neither cleaves the chain nor intercepts the polycon-
densation process, because the resulting cycles may still
contain OH groups. The coexistence of both cyclization
pathways explains the high concentration of cyclic
species in all PES isolated from polycondensations of
silylated phloroglucinol. The consequence for the struc-
ture of the resulting branched or cross-linked PESs is
illustrated in Scheme 7. Most building blocks are small
macrocycles.
Molecu la r Weigh t Mea su r em en ts. Three analyti-
cal methods were used to obtain reliable information
on the molar masses of the PESs listed in Table 2, but
the results were not satisfactory. Polystyrene, even
using the universal calibration, is not a trustworthy
standard for the unusual structure of the PES samples
under investigation. However, the SEC measurements
revealed at least the expected high polydispersities
(PDs) which fell into the range of 4 (no. 1) to 10 (no. 6).
Vapor pressure osmometry (VPO, see Table 2) gave very
low number-average molecular weights (Mns). These low
values certainly underestimate the real Mns, possibly
because the OH groups of the PESs bind water via H
bonds which survives a normal drying process. Light
scattering (LS, see Table 2) gave extremely high weight-
average molecular weights (Mws), which clearly over-
estimate the real values. One reason for the overesti-
mation results from the fact that numerous oligomers
present in these samples do not scatter light. Associa-
tion via H bonds may be another source of errors.
However, the PS-calibrated SEC measurements in NH4
acetate DMF solutions gave similarly high Mw values
low in Figure 4, but it increased significantly with
higher DFDPS feeds. For the bicyclic structure of BC3,
no isomer exists. Therefore, the structure of BC3 proves
that the formation of dimeric cycles is much more
favorable when bridging of a preformed cycle is involved
(Scheme 5) compared to the formation of C2 from the
linear dimer. Another important aspect of BC3 is its
monofunctional character. This means that the normal
cyclization reactions competing with the propagation in
polycondensations of silylated phloroglucinol automati-
cally produce chain terminators, even when side reac-
tions destroying functional groups are absent.
Scheme 3 also includes the various isomers based on
C4 units. Here it is noteworthy that two BC4 species
may exist (BC4 and BC4′) which in contrast to BC3
can react as difunctional building blocks. In the case of
cycles containing two bridge units (Scheme 4) two
aspects are of interest. For B2C5, two isomers can be
formulated which have two structural features in com-
mon. First, they contain rings corresponding to cyclic
dimers, and second, both are monofunctional. This
means that in addition to BC3 cyclic species containing
two or more bridge units, such as B2C5, B3C7, or B4C9,
may also play the role of chain stoppers. The formulas
B2C8 and B2C11 in Scheme 4 should illustrate that
three- and higher-membered rings may act as star
centers having further rings in their star arms. When
the feed of DFDPS is increased beyond the 1.0:1.0 ratio,
the formation of bridged cycles including bicycles is
highly favored. The MS of PES no. 5, Table 2 (Figure
6), demonstrates exemplarily that the BC3/C3, BC4/
C4, and BC5/C5 ratios together with the content of
B2C5 have significantly increased.
The comparison of Figure 6 with Figure 2 illustrates
the striking difference between polycondensations of

Macromolecules, Vol. 36, No. 12, 2003
Multicyclic Poly(ether sulfone)s 4343
Sch em e 6
Sch em e 7
Sch em e 8
as the LS measurements, although the NH4 acetate
should break up the H bonds. Yet, these high Mw values
disagree with the low solution viscosities, which are
more closely correlated with Mw than with Mn. The
MALDI-TOF MS of polymers having a broad molecular
weight distribution do not give any useful information
on Mn or Mw, unless a sample is fractionated into
numerous narrow fractions. In summary, the compari-
son of five series of measurements (incl. viscosities)
demonstrates that it is rather difficult to obtain a
satisfactory description of molecular weights and mo-
lecular weight distributions of this type of polymers.
Alk yla tion w ith Su lton es. According to the original
purpose of this work, namely functionalization of
branched PESs with sulfonate groups, the alkylation
with sultones was studied. On the basis of the experi-
ence with alkylation of other aromatic polyethers,10,11
the direct alkylation of the trimethylsilyl-protected
crude reaction products (structure 3) seemed to be most
promising (Scheme 8). The reaction conditions and
results obtained from 1,3-propane sultone and 1,4-
butane sultone are summarized in Table 5. PESs
prepared with a 10 and 20 mol % excess of DFDPS were
compared. The degrees of substitution (DS) determined
to understand but reproducible. Even in the best case,
not more than 70% of the theoretically available OSiMe3
groups were alkylated. This limitation is most likely a
consequence of steric hindrance. The relatively high
viscosity values have a dual origin. First, a fractionation
took place upon precipitation into methanol. Second, the
charged polymers certainly possess and show the typical
solution properties of polyelectrolytes. In summary, the
introduction of sulfonate groups via sultones was in
principle successful, whereas treatment with concen-
trated sulfuric acid at 20-25 °C caused cross-linking
1
by means of H NMR spectroscopy were calculated on
the basis of OSiMe3 groups available after complete
conversion of all DFDPS. The following trends were
found. 1,4-Butane sultone may give higher DS values
than the propane sultone. An excess of sultone reduces
the DS instead of enhancing it. This result is difficult

4344 Kricheldorf et al.
Macromolecules, Vol. 36, No. 12, 2003
Ta ble 5. Syn th eses of P oly(eth er su lfon e)sa fr om
rubbers). They were not based on MALDI-TOF analy-
ses of the sol phase and were not designed to analyze
a2 + b3 polycondensation.
Silyla ted P h lor oglu cin ol a n d Alk yla tion s w ith Su lton esb
excess ofc
DFDPS
(mol %)
excess ofc
sultone
(mol %)
d
expt
no.
yield
ηinh
The approach of the present work is an alternative
to Flory’s method. It is based on reaction conditions
allowing for more than 99% conversion with variation
of the a2/b3 ratio. This approach in combination with
MALDI-TOF analyses of the sol phase proves the
absence of significant side reactions and a predominant
influence of various cyclization reactions on the course
of the polycondensation. Particularly interesting is the
formation of numerous small cycles and bicycles includ-
ing monofunctional compounds. These monofunctional
bicycles and multicycles can act as chain terminators
and explain the low molecular weights of the products
before gelation. The observation of numerous low molar
mass species prior to gelation was also reported by
Kienle and Flory and attributed to the low conversion
(e75%).
Finally, it should be mentioned that the formation of
bicyclic and multicyclic oligomers observed in this work
agrees with a recent report of Colquhoun and co-workers
on syntheses of cyclic oligo(ether ketone)s.14 Those
authors isolated a bicyclic dimer from the polyconden-
sation of bisphenol A with 1,3,5-tris(4-fluorobenzoyl)-
benzene under pseudo high dilution.
sultone
(%) (dL/g) DSe
1
2
3
4
5
6
7
8
9
10
10
20
20
20
10
10
20
20
20
1,3-propane-
1,3-propane-
1,3-propane
1,3-propane
1,3-propane
1,4-butane-
1,4-butane-
1,4-butane-
1,4-butane-
1,4-butane-
10
20
10
20
50
10
20
10
20
50
77
76
83
85
88
74
76
75
76
82
0.65
0.67
1.17
1.18
1.30
0.58
0.62
1.22
1.28
1.30
55
50
55
46
45
68
63
58
45
45
10
a
b
In NMP at 140 °C/48 h. In NMP at 100 °C/24 h. c Relative
d
to silylated phloroglucinol. Measured at 25 °C with c ) 2 g/L in
DMSO. e Degree of substitution as determined by 1H NMR spec-
troscopy.
and treatment with trimethylsilyl chlorosulfonate at
20-25 °C did not effect any substitution.
Con clu sion
Under conditions favoring a nearly quantitative reac-
tion of the C-F bonds, the polycondensations of silylated
5-methylresorcinol and silylated phloroglucinol showed
the expected high cyclization tendencies. However, in
both cases, the formation of cyclic dimers proved to be
unfavorable. Somewhat unexpected but characteristic
for the polycondensations of silylated phloroglucinol was
the bridging of preformed cycles (yielding bicycles and
possibly multicycles) and the formation of cyclic building
blocks by “backbiting” of C-F groups. As a consequence
of this high cyclization tendency, the resulting branched
PESs and networks mainly consist of cyclic and bicyclic
oligomers as building blocks (Scheme 7).
In the classical experiments and theories of Flory
concerning the gelation in polycondensations of a3 + b2
monomers, equivalent mixtures (a ) b) were studied
and the minimum conversion was determined leading
to gelation. In a clean polycondensation free of cycliza-
tion and side reactions, gelation should occur at 71 (
1% conversion. The experimental results obtained by
Kienle12 and Flory2,13 from various polyester syntheses
were gelation points around 77 ( 2%, and Flory
interpreted this difference as due to the influence of a
few cyclization steps. However, Flory ignored other side
reactions, such as the formation of ether and vinyl
groups, which may also occur in acid-catalyzed (poly)-
esterification reactions at high temperatures. Therefore,
no reliable information about cyclization existed at that
time. More advanced theories of networks and gelation8
also include the formation of a few loops before and
during the gelation process. These modern theories were
designed to improve the understanding of physical and
mechanical properties of loose networks (e.g., synthetic
Ack n ow led gm en t. We wish to thank Dr. H. -J .
Ziegler (GKSS Forschungszentrum Geesthacht GmbH,
Geesthacht, Germany) for the SEC measurements.
Refer en ces a n d Notes
(1) Collected papers of W. H. Carothers on polymerization; Mark,
H., Whitby G. S., Eds.; Wiley-Interscience: New York, 1940.
(2) Flory, P. J . Principles of Polymer Chemistry; Cornell Univer-
sity Press: Ithaca NY, 1953; Chapters VIII and IX.
(3) J acobson, H.; Stockmayer, W. H. J . Phys. Chem. 1950, 54, 0.
(4) J acobson, H.; Beckmann, C. O.; Stockmayer, W. H. J . Chem.
Phys. 1950, 18, 1607.
(5) Kricheldorf, H. R.; Rabenstein, M.; Maskos, M.; Schmidt, M.;
Macromolecules 2001, 34, 713.
(6) Kricheldorf, H. R.; Bo¨hme, S.; Schwarz, G. Macromolecules
2001, 34, 8879.
(7) Kricheldorf, H. R.; Bo¨hme, S.; Schwarz, G.; Kru¨ger, R.-P.;
Schulz, G. Macromolecules 2001, 34, 8886.
(8) Stepto, R. F. T. In Polymer Networks; Stepto, R. F. T., Ed.;
Blackie Academic & Professional: London, 1998; Chapter 2.
(9) Huntress, E. H.; Carter, F. H. J . Am. Chem. Soc. 1940, 62,
511.
(10) Kricheldorf, H. R.; Vakhtangishvili, L.; Fritsch, D. J . Polym.
Sci., Part A: Polym. Chem. 2002, 40, 2967.
(11) Fritsch, D.; Vakhtangishvili, L.; Kricheldorf, H. R. J . Mac-
romol. Sci.sPure Appl. Chem. 2002, A39, 1335.
(12) Kienle, R. H.; van der Meulen, F. A.; Petke, F. E. J . Am.
Chem. Soc. 1939, 61, 2258 and 2268.
(13) Flory, P. J . Chem. Rev. 1946, 39, 137.
(14) Colquhoun, H. M.; Arico, F.; Williams, D. J . Chem. Commun.
2001, 2574.
MA021569I