Hyperbranched polyfluorinated benzyl ether polymers: Mechanism, kinetics, and optimization

Article abstract of DOI:10.1002/pola.27078

Highly fluorinated, hyperbranched polymers were synthesized from the polycondensation of AB₂ monomers, 3,5-bis[(pentafluorobenzyl)oxy] benzyl alcohol and 3,5-bis[(pentafluorobenzyl)-oxy]phenol with potassium carbonate base, and 18-crown-6 phase

Full text of DOI:10.1002/pola.27078

JOURNAL OF
POLYMER SCIENCE
WWW.POLYMERCHEMISTRY.ORG
ARTICLE
Hyperbranched Polyfluorinated Benzyl Ether Polymers: Mechanism,
Kinetics, and Optimization
Matthew J. Quast, Anja Mueller
Science of Advanced Materials, Central Michigan University, Mount Pleasant, Michigan 48858
Correspondence to: A. Mueller (E-mail: muell1a@cmich.edu)
Received 1 November 2013; accepted 12 December 2013; published online 9 January 2014
DOI: 10.1002/pola.27078
ABSTRACT: Highly fluorinated, hyperbranched polymers were
synthesized from the polycondensation of AB2 monomers, 3,5-
bis[(pentafluorobenzyl)oxy]benzyl alcohol and 3,5-bis[(penta-
fluorobenzyl)-oxy]phenol with potassium carbonate base, and
18-crown-6 phase transfer agent in a variety of polar aprotic
solvents. The regioselectivity of the polymerization was opti-
mized and was found to be temperature dependent. The new
polymerization technique produced higher molecular weight
polymer using safer conditions than earlier methods. The
resulting optimization was used to control substitution of
oxygen-bearing nucleophiles along nonactivated fluoroaryl sys-
tems in high yield. Water was found to induce side reactions
that generate a highly conjugated fluoroaryl phenol with low-
ered reactivity. The removal of a methylene spacer in the poly-
mer backbone of the hyperbranched polymer produced
a
polymer with greater thermal stability. The reaction conditions
for polymerization were found to be general for nucleophile-
C
V
bearing perfluorinated systems.
2014 Wiley Periodicals, Inc.
J. Polym. Sci., Part A: Polym. Chem. 2014, 52, 985–994
KEYWORDS: fluoropolymers; hyperbranched; kinetics (polym.);
polyethers; thermal properties
energy, making the material even more suitable for manufac-
turing thin films or membranes.
INTRODUCTION Perfluorinated polymers are an outstanding
building block for thin film and membrane technologies that
require materials with low surface energy, low wettability,
and high thermal stability. The chemical inertness and oxida-
tive/hydrolytic stability of CAF bonds provides an excellent
barrier against chemical and environmental wear.1–4 As a
result, fluorinated polymers are used in self-assembling mate-
rials such as superhydrophobic coatings, antifouling surfaces,
electrolyte membranes, and organic semiconductors.5–12 While
fluorinating agents may be used for the introduction of fluo-
rine, this often results in a random distribution of fluorinated
units along the polymer mainchain.13,14 Traditional fluorinated
polymers such as poly(tetrafluoroethylene) (PTFE) have stiff
main chains and are difficult to tailor, which limits their utility
toward nanometer-scale-ordered structures and specialized
applications.
Hyperbranched, perfluorinated polymers (HBFP) are highly
attractive because of their desirable properties. The hyper-
branched material described here contains two perfluoroaryl
groups suitable for reaction with various nucleophiles to
generate a range of functionalized materials, including poly-
mer additives or material modifiers. Functional density is
increased from branching, which enhances postpolymeriza-
tion modification capabilities. Industrially, the current
method of preparation for hyperbranched perfluorinated
polymer is not safe for large scale-up; here is presented an
alternative system for preparing HBFPs that yields improved
polymer properties from safer conditions.16
Attack by oxygen-, nitrogen-, or sulfur-containing nucleo-
philes along perfluorinated aromatics normally occurs at the
para-position with respect to the point of attachment.17,18
Additional substitution at the ortho-position of fluoroaryl
rings is also documented.19 Computational models have been
designed for determining the principal site of nucleophilic
substitution. The position of attack directly reflects the sta-
bility of the Meisenheimer complex associated with nucleo-
philic aromatic substitution (SNAr) processes (Fig. 1).20 The
energy of this intermediate can be lowered via activating
Hyperbranched polymers have lower glass transition temper-
atures and enhanced flexibility in comparison to linear poly-
mers, which makes them more favorable building blocks
toward nanometer-scale-ordered structures. It is well known
that branching in hyperbranched polymers produces a mate-
rial with lowered solution viscosities and enhanced solubil-
ity, which help permit simple casting of membranes and thin
films.7,15 Additionally, fluorination effectively lowers surface
Additional Supporting Information may be found in the online version of this article.
C
V
2014 Wiley Periodicals, Inc.
WWW.MATERIALSVIEWS.COM
JOURNAL OF POLYMER SCIENCE, PART A: POLYMER CHEMISTRY 2014, 52, 985–994
985

JOURNAL OF
POLYMER SCIENCE
ARTICLE
WWW.POLYMERCHEMISTRY.ORG
FIGURE 1 Meisenheimer complex associated with nucleophilic aromatic substitution in perfluorinated aromatics.
groups in perfluorinated aromatics, as seen in systems bear-
ing an electron withdrawing groups such as benzophenones,
acetophenones, and oxadiazoles.17,21–23
reactions. The extent of substitution at various positions in
the perfluoroaryl group was assessed. The versatility of the
new method has been demonstrated for the polymerization
of AB2 monomers-bearing phenol-based nucleophiles.
The extent of nucleophilic aromatic substitution for nonacti-
vated perfluorinated aromatics using oxygen-containing
nucleophiles has been investigated in this research. This was
done to establish conditions to maximize para-selectivity
during SNAr polymerization to generate HBFPs. In addition,
promotion of the polymerization by various bases in the
presence of phase transfer agent has been explored. A new
phase transfer agent/base system for polymerization has
been established. Herein is described the synthesis condi-
tions required for quantitative conversion of oxygen-bearing
perfluorinated AB2 monomers to the corresponding hyper-
branched polymers (Fig. 2). Reaction of a model compound
was used to assess which species lead to crosslinking
EXPERIMENTAL
Materials
3,5-Dihydroxybenzyl alcohol (99%), 2,3,4,5,6-pentafluoroben-
zyl bromide (99%), 18-crown-6 (99%), resorcinol (99%),
and phloroglucinol (99%) were purchased from Sigma
Aldrich. All reagents were stored in a desiccator and were
used as received. Potassium carbonate (reagent grade-Sigma
Aldrich) was used for preparation of 1 and 4. Analytical
grade potassium carbonate powder (Sigma-Aldrich,
199.99%) was used for all other syntheses. Dry solvents:
acetone (HPLC grade, Fisher Sci.), acetonitrile (HPLC grade,
FIGURE 2 Synthesis of hyperbranched, perfluorinated polymer-bearing benzyl alcohol (5) and phenol (7) nucleophile.
986
JOURNAL OF POLYMER SCIENCE, PART A: POLYMER CHEMISTRY 2014, 52, 985–994

JOURNAL OF
POLYMER SCIENCE
WWW.POLYMERCHEMISTRY.ORG
ARTICLE
Fisher Sci.), and toluene (EMD) were stored over molecular
sieves. Chloroform-d (99.81 atom% D-ACROS) and deuter-
ated dimethyl sulfoxide (99.51 atom% D-ACROS) were used
as received.
and the solvent removed under reduced pressure. After
evaporation, the white solid was recrystallized from 90%
hexanes/dichloromethane: yield 3.57 g (76%).
Tm 82–85 ꢀC, decomposition onset 179.5 ꢀC; 1H NMR (300
MHz, CDCl3) d 5.11 (s, 4H), 6.59 (t, J 5 2.3 Hz, 1H), 6.62–
6.66 (dd, J 5 2.4 and 8.2 Hz, 2H), 7.25 (t, J 5 8.2 Hz, 1H); 13C
NMR (500 MHz, CDCl3) d 57.7, 102.6, 108.3, 110.2, 130.6,
136.1, 139.5, 140.3, 143.7, 144.3, 147.6, 159.4; 19F NMR
(300 MHz, CDCl3) d 2143.0 (m, 4F, ortho-F), 2153.3 (m, 2F,
para-F), 2162.3 (m, 4F, meta-F); IR (ATR, cm21) 2957, 2890,
Characterization
All spectra were recorded at 298 K. FTIR spectra were
recorded using a Thermoscientific Nicolet iS50 ATR and ana-
lyzed using TA Instruments Universal Analysis 2000 software.
1H, 13C, and 19F NMR spectra were recorded using a Varian
Mercury 300 MHz or Varian Inova Unity 500 MHz spectrom-
eter. CDCl3 was used as solvent and TMS as reference
[d(1H) 5 0.00 ppm; d(13C) 5 77.0 ppm]. 19F NMR spectra
were referenced to external C6F6 standard [d(19F)5 2162.9
ppm].
1659, 1590, 1521, 1503, 1380, 1179, 1155, 1058, 937 cm21
GCMS calculated for C20H8F10O2 470.0365, found 470.0417.
;
3,5-Bis[(4-(benzyloxy)22,3,5,6-tetrafluorophenyl)oxy]
benzene (2)
To a mixture of 3,5-bis[(pentafluorobenzyl)-oxy]benzene (0.2
g, 0.43 mmol, 1.0 equiv), 18-crown-6 (0.034 g, 0.13 mmol,
0.3 equiv) and benzyl alcohol (0.138 g, 0.132 mL, 1.3 mmol,
3 equiv) in acetonitrile (15 mL) was added K2CO3 (0.176 g,
1.3 mmol, 3 equiv) and the mixture was vigorously stirred at
room temperature for 24 h. The solvent was removed under
reduced pressure and the white residue was taken up in
chloroform (50 mL) and partitioned between water (50 mL)
and 1 M potassium chloride (50 mL, two times). The organic
layer was dried over magnesium sulfate, the solvent removed
under reduced pressure and a white solid was recrystallized
from 90% methanol/acetone: yield 0.146 g (53%).
Thermal decomposition was examined using a TA Instru-
ments TGA Q500 with a 10 ꢀC/min heating ramp under
nitrogen. DSC measurements were recorded using a TA
Instruments DSC Q1000 with 10 ꢀC/min heating rate in
nitrogen. Glass transition temperature was determined from
cyclic heating ramps after an isothermal step well above the
glass transition temperature of the material.
Mass spectra were recorded using a Waters Micromass MS
Technologies LCT Premier XE unit. GCMS spectra were
recorded using a Waters Micromass GCT Premier instrument.
MALDI-TOF spectra were recorded using a Bruker Daltonics
autoflex unit and using Compass FlexControl software.
Tm 93–95 ꢀC; decomposition onset 269.1 ꢀC; 1H NMR (300
MHz, CDCl3) d 5.04 (s, 4H), 5.26 (s, 4H), 6.56–6.64 (m, 3H),
7.22 (t, 1H), 7.32–7.46 (m, 4H); 13C NMR (500 MHz, CDCl3)
d 57.6, 76.3, 102.2, 107.9, 108.5, 128.3, 128.6, 128.9, 130.2,
135.4, 137.6, 140.2, 142.2, 139.8, 146.8, 159.3; 19F NMR
(300 MHz, CDCl3) d 2144.20 (m, 4F, ortho-F), 2156.03 (m,
4F, meta-F); IR (ATR, cm21) 3040, 2957, 2899, 1655, 1594,
1488, 1377, 1250, 1127, 1039, 932; HRMS (ESI): calcd. for
(M1H1): 647.1463 found 647.3107.
Molar mass distribution and PDI was determined by gel per-
meation chromatography (GPC) equipped with a Viscotek
VE1122 solvent delivery system, PAS102 and PAS103 col-
umns (Polyanalytik), and Perkin Elmer series 200 differential
refractive index detector. All samples were injected at a con-
centration of 3 mg/mL in 100 mL volumes and THF was
used as the mobile phase. Mass distribution was analyzed
using OmniSEC 4.6 software (Malvern Instrument). To
ensure accuracy of molar mass distributions among samples
the GPC instrument assembly was calibrated with external
polystyrene standards (Shodex, Showa Denko K.K.). Percent
monomer conversion was determined by 1H NMR by using
the ratio between peak areas for the benzyl alcohol proton
to that for the internal aromatic protons.
(2-((3-(2,4-Bis(benzyloxy)23,5,6-trifluorobenzyl-oxy)
phenoxy)methyl)24,6-difluorobenzene-1,-3,5-triyl)tris(oxy)
tris(methylene)tribenzene (3)
To
a mixture of 3,5-bis[(pentafluorobenzyl)oxy]-benzene
(0.25 g, 0.53 mmol, 1.0 equiv), 18-crown-6 (0.14 g, 0.53
mmol, 1.0 equiv) and benzyl alcohol (0.38 g, 0.364 mL, 3.5
mmol, 6.6 equiv) in acetonitrile (15 mL) was added K2CO3
(0.483 g, 3.5 mmol, 6.6 equiv), and the mixture was vigo-
rously stirred at reflux for 24 h. The solvent was removed
under reduced pressure and the orange residue was taken
up in chloroform (50 mL), extracted with 1 M potassium
chloride (50 mL, two times), and dried over magnesium sul-
fate. The residue was purified by flash chromatography elut-
ing with 80% hexane/ethyl acetate, increasing polarity to
ethyl acetate, to yield a colorless oil: yield 0.11 g (25.2%).
Synthesis
3,5-Bis[(pentafluorobenzyl)oxy]benzene (1)
A mixture of resorcinol (2.20 g, 20.0 mmol), 18-crown-6
(0.529 g, 0.20 mmol, 0.1 equiv), and 2,3,4,5,6-pentafluoro-
benzyl bromide (6.34 mL, 42.0 mmol, 2.1 equiv) was dis-
solved in anhydrous acetone (100 mL). To the solution was
added finely ground K2CO3 (5.80 g, 42.0 mmol, 2.1 equiv),
and the mixture was stirred at room temperature under
nitrogen for 4 days. The solvent was removed under reduced
pressure and the solid taken up in CH2Cl2 and partitioned
between water (50 mL, two times) and 1 M aqueous potas-
sium chloride solution (50 mL, two times). The organic layer
was dried over anhydrous MgSO4. The solution was filtered
1
ꢀ
Tg 212.93 C; H NMR (300 MHz, CDCl3) d 4.95 (s, 2H), 4.97
(s, 2H), 5.05 (s, 2H), 5.08 (s, 4H), 5.26 (s, 4H), 5.29 (s, 2H),
6.57–6.69 (m, 3H), 7.18–7.52 (ArH, 25H); 19F NMR 2144.2
(m, ortho-F), 2148.0 (dd, J 5 3.79 and 10.32, ortho-F),
WWW.MATERIALSVIEWS.COM
JOURNAL OF POLYMER SCIENCE, PART A: POLYMER CHEMISTRY 2014, 52, 985–994
987

JOURNAL OF
POLYMER SCIENCE
ARTICLE
WWW.POLYMERCHEMISTRY.ORG
TABLE 1 Initial Conditions Used in Stoichiometry Optimization
2156.0 (m, meta-F), 2156.0 (dd, J 5 8.71 and 22.2 Hz, meta-
F); IR (ATR, cm21) 3065, 3033, 2953, 2891, 1640, 1592,
1492, 1481, 1454, 1375, 1281, 1258, 1175, 1146, 1052, 989;
MALDI-TOF MS calcd for C48H36F6O6K1 861.2053 found
of 5
Conc.
(M)
K2CO3
(equiv)a
18-crown-6
(equiv)a
Conv.
(%)
Entries
861.3570
55H43F5O7K1 949.2566 found 949.452 (pentabenzyl alcohol
addition).
(tetrabenzylalcohol
addition);
cald
for
C
A1
A2
A3
B1
B2
0.8
0.8
0.8
0.8
0.8
1.0
1.2
1.4
1.2
1.2
0.3
0.3
0.3
0.2
0.1
89
98
98
94
89
3,5-Bis[(pentafluorobenzyl)oxy]benzyl Alcohol (4)
The synthesis and work up of (4) was adapted from previ-
ous methods.16 A dry round bottomed flask equipped with a
gas inlet adapter was charged with a magnetic stirrer,
2,3,4,5,6-pentafluorobenzyl bromide (39.67 g, 0.152 mol, 2.1
equiv), 3,5-dihydroxybenzyl alcohol (10.14 g, 0.0724 mol, 1
equiv), finely ground K2CO3 (21.01 g, 0.152 mol, 2 equiv)
and 18-crown-6 (1.92 g, 7.2 mmol, 0.1 equiv) and allowed to
react under vigorous stirring in acetone (400 mL) at room
temperature under nitrogen. After 4 days, the solvent was
removed under reduced pressure. The residue was parti-
tioned between CH2Cl2 (300 mL) and water (200 mL, two
times), the aqueous layer was back extracted with CH2Cl2
(100 mL), and all CH2Cl2 layers were combined. The CH2Cl2
extracts were extracted with 4 M KCl solution (150 mL, two
times) to help extract 18-crown-6. The organic layer was
dried over anhydrous MgSO4 and concentrated under
reduced pressure. The product, 4, was recrystallized from
70% hexane/CH2Cl2: yield 29.1 g (80%).
a
Equivalents to monomer.
2144.40 (m, ortho-F), 2152.92 (m, para-F), 2156.53 (m,
meta-F), 2161.94 (m, meta-F); IR (ATR, cm21) 2955, 1657,
1596, 1523, 1493, 1455, 1430, 1374, 1291, 1137, 1051, 937.
3,5-Bis[(pentafluorobenzyl)-oxy]phenol (6)
To
a dry three-necked 100-mL round bottomed flask
equipped with a gas inlet adapter and pressure equalizing
addition funnel was charged a magnetic stirrer, phloroglu-
cinol (1.5 g, 11.9 mmol, 1.0 equiv) potassium carbonate
(1.81 g, 13.09 mmol, 1.1 equiv), 18-crown-6 (0.314 g, ca. 0.1
equiv) and acetone (75 mL). The mixture was stirred for 45
min at room temperature before the drop-wise addition of
2,3,4,5,6-pentafluorobenzyl bromide (3.105 g, 11.9 mmol, 1.0
equiv), and the mixture was stirred for 3 days. The reaction
mixture was filtered, solvent was removed under reduced
pressure, and the beige/orange solid was taken up in ethyl
acetate (75 mL) and partitioned between 1 M HCl (50 mL),
water (50 mL, two times) and saturated potassium chloride.
The organic layer was dried over anhydrous MgSO4 and the
solvent removed under reduced pressure. The solid was
purified by flash column chromatography with eluting with
80% hexanes/ethyl acetate increasing polarity to ethyl ace-
tate to yield a white solid: yield 0.48 g (16.6%).
1
ꢀ
ꢀ
Tg 4 C; Tm 98–100 C; H NMR (300 MHz, CDCl3) d 4.67 (d,
J 5 3.72 Hz, 2H), 5.10 (t, J 5 1.5 Hz, 4H), 6.50 (t, J 5 2.30 Hz,
1H), 6.66 (d, J 5 2.32 Hz, 2H); 13C NMR (500 MHz, CDCl3) d
57.69, 65.17, 70.43, 101.5, 106.4, 110.1, 136.1, 139.5, 140.3,
143.7, 144.2, 147.6, 159.5; 19F NMR d 2143.0 (dd, J 5 8.19
and 22.0 Hz, ortho24F), 2153.3 (t, J 5 20.7 Hz, para22F),
2162.3 (dt, J 5 10.5 and 20.8, meta-4F); IR (ATR, cm21
)
3363, 2891, 1659, 1597, 1524, 1507, 1456, 1434, 1383,
1314, 1290, 1164, 1059, 977, 941, 835, 774, and 689 cm21
;
HRMS (ESI): calcd for (M1H1): 501.0543; found 501.1920.
Tg 1.1 ꢀC; Td 5 181.7 ꢀC; 1H NMR (300 MHz, CDCl3) d 5.04
(t, J 5 1.47 Hz, 4H), 6.12 (d, J 5 2.13 Hz), 6.17 (t, J 5 2.13 Hz,
2H); 13C NMR (300 MHz, CDCl3) d 57.42, 94.61, 96.01,
109.8, 135.8, 139.2, 140.1, 143.5, 144.0, 147.3, 157.5, 159.9;
19F NMR d 2142.29 (dd, J 5 9.56 and 15.36 Hz, ortho-F),
2152.51 (t, J 5 20.68 Hz, para-F), 2162.01 (dt, J 5 7.81 and
13.59 Hz), meta-F; IR (ATR, cm21) 3384, 2962, 1658, 1597,
1522, 1502, 1466, 1380, 1309, 1146, 1053, 974, 938; GCMS:
calcd for C20H8F10O3 (M1) 486.0314; Found 486.0320.
Synthesis Optimization of Hyperbranched Polyfluorinated
Poly(benzyl ether)s (5)
The general method for the polymerization optimization of 1
was to a solution of 1 (0.05–1.25 M) under nitrogen, add
K2CO3 (1.0–1.4 equiv) and 18-crown-6 (0.10–0.30 equiv) in
the chosen solvent (dry) and vigorously stir for 24 h at
room temperature (see Table 1). The reaction mixture was
quenched with excess dichloromethane and dried over anhy-
drous MgSO4 followed by filtration. The solvent removed
under reduced pressure and product was fully characterized
without further purification.
Polyfluorinated Poly-(benzyl-oxyphenol) Polymers (7)
To a dry flask under nitrogen was added 3,5-bis[(pentafluor-
obenzyl)oxy]phenol (0.1 g, 0.21 mmol, 1.0 equiv), potassium
carbonate (0.0341 g, 60.25 mmol, 1.2 equiv), 18-crown-6
(0.0163 g, 0.06 mmol, 0.3 equiv), dry toluene (0.25 mL, 0.8
M), and a magnetic stirrer. The mixture quickly turned pur-
ple and was vigorously stirred for 24 h. After 2 h, a small
aliquot was taken for MALDI-TOF analysis. After 24 h, the
mixture was taken up in dichloromethane (50 mL) and parti-
tioned between 1 M HCl (40 mL), water (40 mL), saturated
From a mixture of monomer (4) (0.25 g, 0.50 mmol, 1.0
equiv), 18-crown-6 (0.040 g, 0.15 mmol, 0.30 equiv), K2CO3
(0.083 g, 0.60 mmol, 1.2 equiv) in acetonitrile (0.50 mL, 1.25
M) stirred for 24 h and by precipitation into 70:30 MeOH/
CHCl3, yield 80%, Mw 5 99,800; Mw/Mn 5 1.86, from GPC
calibrated with PS standards; Tg 57.6 ꢀC; Td 5 315.5 ꢀC; 1H
NMR (300 MHz, CDCl3) d 5.08–5.10 (m, 4H), 5.21 (s, 2H),
19
6.55 (s, 1H), 6.72 (s, 2H);
F NMR d 2142.76 (m, ortho-F),
988
JOURNAL OF POLYMER SCIENCE, PART A: POLYMER CHEMISTRY 2014, 52, 985–994

JOURNAL OF
POLYMER SCIENCE
WWW.POLYMERCHEMISTRY.ORG
ARTICLE
potassium chloride (30 mL), and the organic layer dried
over magnesium sulfate. The solvent was removed under
reduced pressure to afford an off-white yellowish solid. The
product was characterized without any further purification;
yield 84 mg (86%).
Mw 5 54,423, Mw/Mn 5 2.54; Tg 80.9 ꢀC; Td 5 341.0 ꢀC; 1H
NMR (300 MHz, CDCl3) d 5.04 (t, J 5 1.47 Hz, 4H), 6.12 (d,
J 5 2.13 Hz), 6.17 (t, J 5 2.13 Hz, 2H); 13C NMR (300 MHz,
CDCl3) d 57.42, 94.61, 96.01, 109.8, 135.8, 139.2, 140.1,
143.5, 144.0, 147.3, 157.5, 159.9; 19F NMR d 2142.29 (dd,
J 5 9.56 and 15.36 Hz, ortho-F), 2152.51 (t, J 5 20.68 Hz,
para-F), 2162.01 (dt, J 5 7.81 and 13.59 Hz, meta-F); IR
(ATR, cm21) 3384, 2962, 1658, 1597, 1522, 1502, 1466,
1380, 1309, 1146, 1053, 974, 938.
FIGURE 3 Synthetic pathway for penta-fluorinated model (1),
benzyl alcohol-bearing monomer (4), and phenol-bearing
monomer (6).
RESULTS AND DISCUSSION
2
(dimer,
C
40H15F18O5
,
calcd m/z 5 917.0638, found m/
Previous work demonstrated the synthesis for hyper-
branched perfluorinated polymers from sodium/toluene sus-
pensions or sodium potassium alloy, which prevent any safe
scale up of the polymerization.16 Therefore, potassium car-
bonate was investigated here as an alternative base. Initial
polymerization experiments using potassium carbonate were
successful, although a phase transfer agent was required to
achieve high conversions. Rarely, red gelatinous side prod-
ucts were observed—which was a common side product of
earlier polymerizations using sodium metal. A model com-
pound was used to identify the red side products, which are
characteristic of pregelation processes in hyperbranched per-
fluorinated systems. Furthermore, the new polymerization
system was versatile toward polymerization of other per-
fluorinated polymers.
z 5 916.9954). This suggested that the anion was being gen-
erated from proton abstraction along the hydrocarbon back-
bone. While unexpected, this proton abstraction may be
possible at the benzyl ether position. Deprotonation at the
benzyl ether position would produce a carbanion, which is
stabilized by extended conjugation with both aromatic rings
(Fig. 4). Furthermore, a carbanion generated at the benzyl
ether position would be stabilized inductively with the
neighboring oxygen and neighboring electron-deficient per-
fluorinated ring. When the benzyl ether is deprotonated, con-
jugation between aromatic rings combined with the electron
withdrawing characteristics of perfluorinated segments
results in a low-energy, highly stable ionic species that is red
in color. These results suggest that a reactive hydroxide spe-
cies that was capable of crosslinking perfluorinated rings
was produced by hydroxide substitution along the perfluori-
nated ring. Hydroxide was generated from the in situ associ-
ation of water with the potassium carbonate/crown ether
system. Additional base (present as K2CO3) then deproto-
nated the newly formed arylfluoro-phenol, creating a nucleo-
phile attacking a para-fluorine on another model compound.
Model Experiments
A model compound (1) was exposed to a variety of bases in
presence of 18-crown-6 phase transfer agent. This allowed
for the source of discoloration and sequential gelation during
polymerization of perfluorinated AB2 monomers to be identi-
fied for the first time. It was expected that the source of dis-
coloration could be attributed to extended conjugation
resulting from interlinking reactions between pentafluoro-
phenyl rings. The model (1) was chosen because the absence
of a benzyl alcohol group allows for study of the nucleophilic
aromatic substitution reaction on the pentafluorophenyl ring
exclusively. Furthermore, steric hindrance and activation
energy of the perfluorinated rings of both the model (1) and
the monomers (4) and (6) were considered similar (Fig. 3).
In a second experiment, the model was reacted with potas-
sium hydroxide in presence of 18-crown-6 in acetonitrile at
room temperature to examine the susceptibility and regiose-
lectivity of perfluorinated aromatics to hydroxide substitu-
tion. In this case, the hydroxide was directly introduced into
the system rather than generated in situ from the association
of K2CO3 with water. Discoloration ensued almost immedi-
ately after the addition of potassium hydroxide to the model.
Negatively charged mass spectral analysis revealed that
hydroxide was incorporated along the perfluorinated ring.
Because this experiment did not rely on in situ hydroxide
generation, and no (non-nucleophilic) K2CO3 base was pres-
ent, substitution processes were enhanced, while interlinking
processes leading to oligomerization were suppressed. This
allowed for more detailed characterization of the regioselec-
tivity for hydroxide substitution.
Side Product Identification
Reaction of the model with potassium carbonate/18-crown-6
in acetonitrile with water (10 equiv) induced reddening in
24 h at reflux, or over several days at room temperature.
Negatively charged mass spectral analysis revealed the for-
mation hydoxy-substituted model as well as ether-linked
oligomers of the model (Fig. 4, also GPC data, not shown).
Signals for the dimer and trimer of the model in the HRMS
spectrum show that the negative ion being generated corre-
sponds to oligomers with no terminal hydroxyl groups
The 19F NMR revealed that the hydroxide was incorporated
at the para position of the perfluorinated ring and upon
WWW.MATERIALSVIEWS.COM
JOURNAL OF POLYMER SCIENCE, PART A: POLYMER CHEMISTRY 2014, 52, 985–994
989

JOURNAL OF
POLYMER SCIENCE
ARTICLE
WWW.POLYMERCHEMISTRY.ORG
FIGURE 4 Interlinking processes from hydroxide substitution in nonactivated perfluorinated aromatics (anion depicted at benzyl
ether position, stabilized by extended conjugation, confirmed by ESI).
treatment of the hydroxide-substituted model with acetic acid
the two new fluorine signals immediately shifted downfield
(Fig. 5). Further treatment with potassium carbonate illus-
trated the proton absorption at the phenol was reversible.
The downfield shift suggests that the hydroxide substituted
perfluorinated ring existed as an anion stabilized by extended
conjugation. This stabilizing effect slows kinetics of the inter-
linking processes due to lowered nucleophilicity associated
with extended conjugation. The effect of pH was also demon-
strated by UV–vis spectroscopy where, immediately after acid-
ification with acetic acid, the absorption bands at 404 and
486 nm disappeared (Fig. 6), suggesting that the conjugation
length decreased upon proton uptake. These data show that
the colored side product is a result of the combination of
para-hydroxy-substituted fluoraryl rings and fluoroaryl ether-
linked oligomers. Under appropriate conditions, the hydroxyl
can be deprotonated and react with another pentafluoro-
phenyl group (Fig. 4). During a polymerization, this would
lead to crosslinking and thus gelation. Deprotonation of the
benzyl ether proton results in
a stabilized anion with
extended conjugation that is red in color. These results sug-
gest that a polymerization should be performed in absence of
water and would be best in a solvent that is immiscible with
water or less susceptible to water uptake.
FIGURE 5 19F NMR of hydroxide-substituted model compound.
990
JOURNAL OF POLYMER SCIENCE, PART A: POLYMER CHEMISTRY 2014, 52, 985–994

JOURNAL OF
POLYMER SCIENCE
WWW.POLYMERCHEMISTRY.ORG
ARTICLE
positions. Alkyl groups are para- and ortho- directing while
oxygen groups are ortho- and meta- directing in perfluori-
nated aromatics.24 This combined activation results in benzyl
ether substitutions occurring at the ortho- position of the
perfluorinated ring, which was expected from activation
predictions.20
These results suggested that all polymerizations should be
performed at room temperature and in a dry atmosphere to
avoid crosslinking and enhance para-fluorine selectivity in
the polycondensation of perfluorinated AB2 monomers.
Optimization of Polymer Yield and Molecular Weight
The synthesis of monomer (4) was modified from previous
work16 by the addition of a saturated potassium chloride
wash during the workup. This step was used in each experi-
ment to help draw 18-crown-6 into the aqueous layer and to
ensure polymers were potassiated for mass spectral analysis.
FIGURE 6 Effect of adjusting the pH with acetic acid on
hydroxide substituted model (reversible).
The polymerization optimization used the above results to
achieve high purity. The reagent stoichiometry and concen-
tration of the polymerization are herein optimized for maxi-
mum monomer conversion, maximum molecular weight, and
minimal polydispersity. It is known that the reactivity of the
para-fluorine toward nucleophilic aromatic substitution of
strong and weak nucleophiles is dependent on tempera-
ture.20 The group of Zhao et al. reported that the pentafluor-
ophenyl group exhibited temperature-based stepwise
reactivity toward nucleophilic aromatic substitution; how-
ever, this was more reactive than the monomer (4).23 When
monomer (4) was exposed to similar conditions in DMF
using potassium fluoride as the base no polymerization
occurred and slight discoloration ensued. This was attributed
to the poor basicity of the fluoride ion. The addition of phase
transfer agent (18-crown-6, in excess) induced polymeriza-
tion, but only 60% conversion was seen after one week. The
poor conversion and slow kinetics associated with potassium
fluoride-catalyzed polymerization of our monomer is also
due to the decreased acidity of the benzyl alcohol relative to
the phenol used by Zhao et al. In another report, Xu pre-
pared oligomers of pentafluorobenzyloxy fluorophenones
from potassium carbonate and 18-crown-6 in toluene at
Benzyl Alcohol Substitution Along the Model (2) and (3)
To assess the reactivity along the perfluorinated ring during
a polymerization at room temperature or reflux, the model
was exposed to excess benzyl alcohol in presence of base/
phase transfer catalyst mixture. Benzyl alcohol was chosen
since this nucleophile has similar electronic and steric fea-
tures as the monomer (Fig. 7). Characterization by 19F NMR
illustrated that only the para-fluorine carbon centers were
susceptible to nucleophilic aromatic addition reactions at
room temperature, which resulted in the dibenzyl alcohol
substituted model (Fig. 7, Supporting Information Figs. SI1
and SI2). In comparison, when the model was exposed to
excess benzyl alcohol in presence of base/phase transfer cat-
alyst mixture at reflux, extensive substitution occurred along
the perfluorinated ring (Fig. 7, Supporting Information Figs.
SI3 and SI4). Analysis by MALDI-MS in positive ion reflectron
mode showed that primarily
a tetrasubstituted model
formed. Only a minor amount of model substituted with 5
equiv was seen at reflux. No product substituted with 6
equiv was seen in the mass spectrum, which is likely a result
of steric hindrance around the fluoroaryl ring. Substitution
beyond the para-position was found to occur at ortho-
FIGURE 7 Addition of benzyl alcohol (excess) to the model at room temperature and reflux in acetonitrile.
WWW.MATERIALSVIEWS.COM
JOURNAL OF POLYMER SCIENCE, PART A: POLYMER CHEMISTRY 2014, 52, 985–994
991

JOURNAL OF
POLYMER SCIENCE
ARTICLE
WWW.POLYMERCHEMISTRY.ORG
redden; therefore, 1.2 equiv of potassium carbonate was
used for the remainder of the study. Reducing the amount of
phase transfer agent (entry B, Table 1) reduced the percent
conversion of monomer in polymerization. Quantitative
monomer conversion was achieved from 0.3 equiv of 18-
crown-6 phase transfer agent. These results suggest that the
amount of phase transfer agent determines the quantity of
active nucleophiles in the system.
Interestingly, varying the concentration of the monomer dur-
ing a polymerization alters the structure of the polymer (Fig.
8). This change may be attributed to cyclization. Cyclization
occurs when a highly branched polymer closes on itself by a
unimolecular reaction. As the concentration of a polymeriza-
tion mixture increases, the probability of intramolecular
cyclization decreases and the probability of bimolecular reac-
tions increases. The effect of monomer concentration on
molecular weight is shown in Table 2. The preliminary
model studies demonstrated that the nucleophilic aromatic
substitution along (1) was para-specific at room temperature
and only at reflux was further substitution possible. This
suggests the three new signals in the 19F NMR can be attrib-
uted to cyclization during a polymerization, which has been
previously demonstrated in HBFPs.26 This assignment was
confirmed by MALDI-TOF (Supporting Information Figs. SI5
and 6). The data confirms that, as concentration is reduced,
the probability of intramolecular cyclization increases. Like-
wise, at 0.05-M monomer concentration, only the cyclic poly-
mer was seen and, at 0.8-M monomer concentration, almost
no cyclic polymer was seen relative to the acyclic structure.
FIGURE 8 Effect of varying the concentration in a polymeriza-
tion mixture using 1.2 equiv K2CO3 and 0.3 equiv 18-crown-6.
room temperature in good yield.22 Herein, we investigate
methods similar to work by Xu toward the preparation of
hyperbranched perfluorinated systems.
Initially, the addition of potassium carbonate (1 equiv, in
absence of phase transfer agent) to (4) in acetonitrile
resulted in polymerization with little byproduct formation.
Additional potassium carbonate up to 10 equiv, or reducing
the particle size of potassium carbonate yielded larger
molecular weight, however, even at long reaction time (>3
days), only oligomers were obtained.
That demonstrates that a phase transfer agent is required.
The addition of 18-crown-6 to the reaction mixture was
found to significantly improve polymerization rates, as seen
for similar substitutions.24,25 The base/phase transfer agent
system was optimized for maximum conversion of 4 and
purity using a multiparameter approach by varying potas-
sium carbonate stoichiometry, minimalizing 18-crown-6 stoi-
chiometry, varying monomer concentration, and varying
solvents at near ideal conditions (Table 1).
Polymerization Kinetics
Interpreting kinetics of phase transfer-assisted nucleophilic
aromatic substitutions is difficult due to the complex influ-
ence of cations and anions on the activity of charged and
neutral species involved in a polymerization.27 The addition
of cation complexing agents are known to activate substitu-
tion processes by reducing ion association of the crown
ether complex with the alkoxide ion, generating a “free” alk-
oxide ion with greater reactivity.28 For simplicity, the kinetics
of nucleophilic substitution in polymerizations were studied
in acetonitrile, acetone, and toluene from a dilute (0.1 M)
monomer mixture by monitoring monomer conversion over
time. Kinetic plots of ln([M]0/[M]) versus time revealed
pseudo first-order kinetics, which are characteristic of phase
transfer catalyzed nucleophilic aromatic substitution (Fig.
9).29 Polymerization rate increased with increasing solvent
The percent conversion during a polymerization was moni-
tored via 1H NMR using unreacted benzyl alcohol groups as
a reference and was measured over a variety of conditions.
Initial results suggested that 0.8-M monomer concentration
yielded large molecular weight and nearly quantitative
monomer conversion. The effect of the potassium carbonate
equivalence (entry A) was measured in presence of excess
phase transfer catalyst (0.3. equiv) (Table 1). The difference
in conversion percent between 1.4 and 1.2 equiv of potas-
sium carbonate was negligible, and polymerizations using
1.4 equiv of potassium carbonate were more likely to
TABLE 2 Effect of Concentration on Molecular Weight of (5) in Acetonitrile
Entries
Conc. (M)
K2CO3 (equiv)a
18-crown-6 (equiv)a
Conversion (%)
Mn (kDa)
Mw/Mn
C1
C2
C3
C4
0.50
0.80
1.25
1.75
1.2
1.2
1.2
1.2
0.3
0.3
0.3
0.3
98.7 6 1.9
98.8 6 1.0
97.2 6 0.5
95.3 6 1.6
39.2 6 1.7
52.7 6 3.7
55.2 6 3.9
46.8 6 1.6
2.45 6 0.13
2.64 6 0.45
1.89 6 0.02
1.75 6 0.10
a
Equivalents to monomer.
992
JOURNAL OF POLYMER SCIENCE, PART A: POLYMER CHEMISTRY 2014, 52, 985–994

JOURNAL OF
POLYMER SCIENCE
WWW.POLYMERCHEMISTRY.ORG
ARTICLE
cyclization probability during a polymerization was assessed.
MALDI-MS revealed that at high concentration a mixture of
cyclic and acyclic polymer was present (Fig. 10). At low con-
centration, only the cyclic structure was observed (Support-
ing Information Fig. SI7). These results showed instead that
removal of one carbon along the benzyl alcohol had insignifi-
cant effect on cyclization probability. The cyclized polymer
revealed a similar signal (Supporting Information SI8) in the
19F NMR as that for (5) (Fig. 8).
Thermal Properties of Hyperbranched Polymers
It was anticipated that removing the benzyl alcohol from the
monomer in place for a phenol would increase the stability
of the polymer. Thermogravimetric analysis revealed that
before polymerization the decomposition onset of the benzyl
alcohol-bearing monomer (Td 5 218.5 ꢀC) was higher than
that of the phenol-bearing monomer (Td 5 213.6 ꢀC). After
polymerization, the thermal decomposition of both materials
FIGURE 9 Monomer conversion versus time for 0.1 M poly-
merizations in acetonitrile, acetone, and toluene.
ꢀ
polarity (see dielectric constants). The rate determining step
in SNAr reactions is the formation of the carbon-nucleophile
bond of a tetrahedral geometry corresponding to a nega-
tively charged Meisenheimer intermediate (Fig. 1).27 In more
polar solvent, the solvation shell around of the nucleophile is
reduced, which effectively increases energy and reactivity of
the alkoxide ion, ultimately increasing the rate.30 The nega-
tive charge in this intermediate is delocalized in the electron
deficient perfluorinated ring. Kinetic data demonstrated that
the nucleophilic aromatic substitution along the fluoroaryl
rings occurred most readily in polar aprotic solvents. The
rate increased with increasing polarity. For similar systems,
solvent should be chosen based on the solubility of the prod-
uct and kinetics can be enhanced by increasing polarity.
increased to 315.5 and 341.0 C, respectively. The phloroglu-
cinol polymer possessed a considerably higher decomposi-
tion onset, which was likely a result of increased stability of
the phenyl ether linkage when compared with the benzyl
ether. The glass transition temperature of the phloroglucinol
polymer was significantly higher as well when compared
with the benzyl ether polymer (Fig. 11) due to the reduced
Versatile Polymerization
To demonstrate the versatility of the potassium carbonate/
18-crown-6 system in the nucleophilic aromatic substitution
of perfluorinated aromatics a second, phenol-bearing AB2
monomer system was polymerized (Fig. 2). Phloroglucinol
was used as precursor to prepare the phenol-bearing mono-
mer in a single step. Low yields for the phloroglucinol mono-
mer synthesis may be attributed to the equivalent electronic
environment at each phenol center, which prohibits any
regioselective control of the pentafluorobenzyl bromide sub-
stitution process. The substitution extent was therefore
manipulated by adjusting the stoichiometry of phloroglucinol
to pentafluorobenzyl bromide (1 equiv of phenol to 3 equiv
of pentafluorobenzyl bromide). The disubstituted phloroglu-
cinol was obtained in relatively good yields (16.6%) com-
pared with similar phloroglucinol etherifications that
additionally required tedious protection/deprotection steps
(19%).31
The phenol-bearing AB2 system, missing a methylene spacer
when compared with the benzyl ether polymer, was expected
to be less susceptible to cyclization. By exposing the phenol-
based AB2 monomer to the optimum potassium carbonate/
18-crown-6 stoichiometry in a concentrated solution (0.8 M)
of toluene and dilute solution of acetonitrile (0.1 M), the
FIGURE 10 MALDI-MS spectrum of polymer (7) from a 0.8 M
monomer polymerization after 2 h illustrating primarily acyclic
polymer. (a) Full spectrum; (b) enhanced scale.
WWW.MATERIALSVIEWS.COM
JOURNAL OF POLYMER SCIENCE, PART A: POLYMER CHEMISTRY 2014, 52, 985–994
993

JOURNAL OF
POLYMER SCIENCE
ARTICLE
WWW.POLYMERCHEMISTRY.ORG
4 S. V. Gangal, Perfluorinated polymers, tetrafluoroethylene-
perfluorovinyl ether copolymers, in styrene polymers to toys,
In Encyclopedia of Polymer Science and Engineering, 2nd Ed.;
H. F. Mark, N. M. Bikales, C. G. Overberger, G. Menges, Eds.;
Wiley Interscience: New York, 1989; Vol. 16, pp 677–679.
5 B. J. Basu, V. D. Kumar, ISRN Nanotechnology, 2011, 2011,
803910.
6 P. Graham, M. Stone, A. Thorpe, T. G. Nevell, J. Tsibouklis,
J. Fluorine Chem. 2000, 104, 29–36.
7 D. Gan, W. Cao, Z. Wang, J. Fluorine Chem. 2002, 116, 59–63.
8 J. Shen, T. Hogen-Esch, J. Am. Chem. Soc. 2008, 130, 10866–
10867.
9 D. W. Shin, S. Y. Lee, C. H. Lee, K. Lee, C. H. Park, J. E.
McGrath, M. Zhand, R.
B Moore, M. D. Lingwood, L. A.
Madsen, Y. T. Kim, I. Hwang, Y. M. Lee, Macromolecules 2013,
46, 7797–7804.
10 Y. Xiu, D. W. Hess, C. P. Wong, In A Novel Method to Pre-
pare Superhydrophobic, Self-Cleaning and Transparent Coat-
ings for Biomedical Applications, Electronic Components and
Technology Conference, Reno, NV, May 29–June 1, 2007.
FIGURE 11 Effect of number average molecular weight on
glass transition temperature (᭹ benzyl ether HBFP – 5) (᭺ phe-
nyl ether HBFP – 7).
11 H. Jeong, B. Lim, D. Khim, M. Han, J. Lee, J. Kim, J. Yun, K.
Cho, J. Park, D. Kim, Adv. Mat. 2013, 25, 6416–6422.
flexibility of the phenol linkages and smaller free volume
associated with a more linear structure. In both polymers,
the effect of molecular weight on glass transition tempera-
ture leveled off near 20k Mn. This was attributed to the
reduced concentration of chain ends associated with higher
molecular weights.
12 C. S. Gudipati, C. M. Greenlief, J. A. Johnson, P.
Prayongpan, K. L. Wooley, J. Polym. Sci. Part A: Polym. Chem.
2004, 42, 6193–6208.
13 M. E. Arnold, K. Nagai, R. J. Spontak, B. D. Freeman, D.
Leroux, D. E. Betts, J. M. DeSimone, F. A. DiGiano, C. K.
Stebbins, R. W. Linton, Macromolecules 2002, 35, 3697–3707.
14 M. A. Hillmyer, T. P. Lodge, J. Polym. Sci. Part A: Polym.
Chem. 2002, 40, 1–8.
CONCLUSIONS
15 M. Seiler, Fluid Phase Equilib. 2006, 241, 155–174.
In summary, a facile, alternative synthesis for hyperbranched
perfluorinated polymers with greater purity, greater molecu-
lar weight, and lower polydispersity index was developed.
The extent of substitution along nonactivated perfluorinated
aromatics can be controlled with temperature to achieve
hierarchical reactivity using 18-crown-6 phase transfer agent
and potassium carbonate base system. Cyclization probability
and molecular weight of benzyl ether and phenyl ether per-
fluorinated AB2 polymers were governed by concentration.
Perfluorinated aromatics are susceptible to substitution by
hydroxide, which in basic polar aprotic media can further
react to yield crosslinked tetrafluorinated aryl ethers. This
16 A. Mueller, T. Kowalewski, K. L. Wooley, Macromolecules
1998, 31, 776–786.
17 S. Buscemi, A. Pace, R. Calabrese, N. Vivona, P.
Metrangolo, Tetrahedron 2001, 57, 5865–5871.
18 J. Ogikubo, J. L. Worlinsky, Y. Fu, C. Bruckner, Tetrahedron
Lett. 2013, 54, 1707–1710.
19 I. R. Hardcastle, A. M. Barber, J. H. Marriott, M. Jarman, Tet-
rahedron Lett. 2001, 42, 1363–1365.
20 J. Baker, M. Muir, Can. J. Chem. 2010, 88, 588–597.
21 P. D. Buttero, R. Gironda, M. Moret, A. Papagni, M. Parravicini,
S. Rizzato, L. Miozzo, Eur. J. Org. Chem. 2011, 2265–2271.
22 L. Xu, R. W. Giese, J. Fluorine Chem. 1994, 67, 47–51.
method is
nucleophile-bearing perfluorinated systems.
a general method that can be used for
23 Z. Zhao, X. Ma, A. Zhang, N. Song, J. Polym. Sci. Part A:
Polym. Chem. 2011, 49, 2423–2433.
24 B. Davis, Anal. Chem. 1977, 49, 832–834.
25 S. Montanari, C. Paradisi, G. Scorrano, J. Org. Chem. 1991,
56, 4274–4279.
ACKNOWLEDGMENTS
Support from the Science of Advanced Materials graduate pro-
gram at the University of Central Michigan is gratefully
acknowledged.
26 J. K. Gooden, M. L. Gross, A. Mueller, A. D. Stefanescu, K.
L. Wooley, J. Am. Chem. Soc. 1998, 120, 10180–10186.
27 C. A. Bunton, L. Robinson, J. Am. Chem. Soc. Perk. Trans. II
1968, 90, 5972–5979.
28 C. Paradisi, G. Scorrano, Nucleophilicity, Advances in Chem-
istry Series 215; American Chemical Society: Washington DC,
1987; pp 339–354.
REFERENCES AND NOTES
1 R. D Chambers, In Fluorine in Organic Chemistry; Wiley: New
York, 1973; Chapter 1, pp 3–5.
29 C. Mariani, G. Modena, G. P. Pizzo, G. Scorrano, J. Am.
Chem. Soc. Perk. Trans. II 1979, 1187–1193.
2 B. E. Smart, Characteristics of C-F Systems, In Organofluorine
Chemistry: Principles and Commercial Applications; B. E.
Banks, B. E. Smart, J. C. Tatlow, Eds.; Plenum Press: New York
and London, 1994; Chapter 3, pp. 57–88.
30 C. Reichardt, Solvents and Solvent Effects in Organic Chem-
istry, 3rd ed.; Wiley-VCH: Weinheim, 2003.
31 P. M. Tadross, C. D. Gilmore, P. Bugga, S. C. Virgil, B. M.
Stoltz, Org. Lett. 2010, 12, 1224–1227.
3 R. F. Brady, Chem. Br. 1990, 26, 427.
994
JOURNAL OF POLYMER SCIENCE, PART A: POLYMER CHEMISTRY 2014, 52, 985–994