Thermoreversible crosslinking of polyethylene enabled by free radical initiated functionalization with urethane nitroxyls

Article abstract of DOI:10.1016/j.polymer.2009.11.039

This paper describes the functionalization of polyethylene with urethane derivatives of 2,2,6,6-tetramethylpiperidinyloxy (TEMPO), resulting in thermoreversible crosslinking of the polymer, which exhibits thermoplastic characteristics at sufficiently elevated temperatures. The urethane TEMPO adducts were synthesized by the reaction of 4-hydroxy TEMPO with various diisocyanates using appropriate catalysts, and were subsequently grafted to polyethylene using free-radical chemistry. A model study conducted on urethane isopropanol adducts confirmed that endothermic urethane reversion occurred with aromatic and alkyl diisocyanates at temperatures greater than 170 °C, and no exothermic urethane decomposition was observed. In contrast, urethane TEMPO adducts underwent exothermic decomposition at elevated temperatures, probably because of participation of the free nitroxyl species in urethane decomposition. Hence, the urethane TEMPO adducts were most effective when used as crosslinkers below their decomposition points.

Full text of DOI:10.1016/j.polymer.2009.11.039

Polymer 51 (2010) 153–163
Contents lists available at ScienceDirect
Polymer
journal homepage: www.elsevier.com/locate/polymer
Thermoreversible crosslinking of polyethylene enabled by free radical
initiated functionalization with urethane nitroxyls
,
b
a
b
Bharat Indu Chaudhary a , Thomas H. Peterson , Eric Wasserman , Stephane Costeux ,
*
´
John Klier b, Andrew J. Pasztor Jr. b
a The Dow Chemical Company, 171 River Road, Piscataway, NJ 08854, USA
b The Dow Chemical Company, Midland, MI 48674, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
This paper describes the functionalization of polyethylene with urethane derivatives of 2,2,6,6-tetra-
methylpiperidinyloxy (TEMPO), resulting in thermoreversible crosslinking of the polymer, which exhibits
thermoplastic characteristics at sufficiently elevated temperatures. The urethane TEMPO adducts were
synthesized by the reaction of 4-hydroxy TEMPO with various diisocyanates using appropriate catalysts,
and were subsequently grafted to polyethylene using free-radical chemistry. A model study conducted on
urethane isopropanol adducts confirmed that endothermic urethane reversion occurred with aromatic
and alkyl diisocyanates at temperatures greater than 170 ꢀC, and no exothermic urethane decomposition
was observed. In contrast, urethane TEMPO adducts underwent exothermic decomposition at elevated
temperatures, probably because of participation of the free nitroxyl species in urethane decomposition.
Hence, the urethane TEMPO adducts were most effective when used as crosslinkers below their
decomposition points.
Received 28 July 2009
Received in revised form
17 November 2009
Accepted 18 November 2009
Available online 26 November 2009
Keywords:
Thermoreversible
Crosslinking
Polyethylene
Ó 2009 Elsevier Ltd. All rights reserved.
1. Introduction
A key disadvantage of peroxide-based crosslinking (as well as
other conventional crosslinking chemistries, including silane and
Polyolefins such as ethylenic polymers are often crosslinked in
order to increase resistance to flow and high temperature defor-
mation, as well as to enhance other properties. The methods used
to crosslink polyolefins include the use of organic peroxides at
elevated temperatures which generate polymer radicals that
subsequently combine to form carbon–carbon bonds [1–9].
For the specific case of polyethylene, it has been demonstrated
that crosslinking via peroxides can be controlled through the
reaction of carbon-centered radicals with persistent nitroxyl radi-
cals such as 2,2,6,6-tetramethyl-1-piperidinyloxy (TEMPO) and its
derivatives which, if additional functionality is present, also enables
functionalization of the polymer [10]. The mechanism through
which crosslinking or ‘‘cure’’ is influenced by persistent radicals
derives from their ability to intercept incipient polymeryl radicals
formed via hydrogen atom abstraction by peroxides to yield poly-
meryl alkoxyamines [10]. When persistent radicals comprising
more than one nitroxyl (such as PROSTABÒ 5415, also known as
‘‘bis-TEMPO’’) are used, crosslinks result and the polymer is trans-
formed from a thermoplastic to a thermoset.
radiation technologies) is that the materials become thermoset,
and cannot be processed as thermoplastics after manufacture.
Therefore, it would be advantageous to identify and develop
a strategy that enables polymer crosslinking in a reversible manner
wherein the subsequent application of a chemical or physical
stimulus permits controlled reversion of the thermoset back to its
antecedent thermoplastic form. This strategy may take advantage
of a variety of chemistries, for instance, by incorporating a reactive
group into the polymer architecture through grafting or copoly-
merization of a functional comonomer and subsequently coupling
with difunctional crosslinking agents that react in a reversible
manner. For the case of thermoreversible crosslinking, reversion to
the non-crosslinked thermoplastic state may be achieved by the
application of high temperatures wherein the entropic gains of
decrosslinking overcome the enthalpic driving force of crosslinking.
Entropy values typical for fragmentation reactions are in the range
of 20–30 cal molꢁ1. Thus, by a judicious choice of the crosslinking
chemistry, one can, in theory incorporate reactions where the
entropic/enthalpic balance effects decrosslinking in a temperature
range useful for polymer processing.
In this paper, we describe a novel approach toward thermor-
eversible crosslinking by taking advantage of the reversibility of the
reactions of isocyanates with alcohols that can be grafted to the
* Corresponding author. Tel.: þ1 732 563 5137; fax: þ1 732 563 5218.
E-mail address: bichaudhary@dow.com (B.I. Chaudhary).
0032-3861/$ – see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.polymer.2009.11.039

154
B.I. Chaudhary et al. / Polymer 51 (2010) 153–163
polymer backbone through radical-mediated nitroxyl coupling.
Specifically, we have synthesized diurethane crosslinking agents
that are end-capped with graftable nitroxyl groups. The urethane
linkage, derived from reactions of isocyanates with polyols, is
known to be thermoreversible at temperatures between 120 ꢀC and
250 ꢀC, depending on the nature of the alcohol and isocyanate
precursors [11]. In the present work, urethane TEMPO adducts were
synthesized by reactions of 4-hydroxy TEMPO with various diiso-
cyanates, and these were grafted to polyethylene to impart ther-
moreversibility to the crosslinked polymer. A schematic of such
thermoreversible crosslinking with urethane bis-TEMPO (UBT)
derivatives is shown in Fig. 1. Although this concept has been
demonstrated with polyethylene and peroxides, it is potentially
adaptable to other polymers (such as propylene polymers) and
with other free-radical generators (such as electron beam
radiation).
were purchased from Honeywell, Burdick and Jackson or Aldrich
and were further purified through an in-house solvent clean-up
system.
Hexamethylene
diisocyanate
(HMDI),
methyl-
enediphenylisocyanate (MDI), triethylamine and dibutyltin dilau-
rate (DBTDL) were obtained from Aldrich. PAPIÒ 901 (polymeric
MDI) was obtained from The Dow Chemical Company. 4-Hydroxy
TEMPO refers to 4-hydroxy-2,2,6,6-tetramethyl-1-piperidinyloxy,
obtained from AH Marks or Aldrich. The urethane nitroxyls
synthesized (HMDI UBT, MDI UBT, and PAPI-TEMPO adducts) are
illustrated in Fig. 2. Infrared spectroscopy of HMDI UBT, MDI UBT
and PT adducts was performed using a Mattson Galaxy 5000
spectrometer.
2.2.2. Synthesis
Preparation of HMDI UBT. To a stirred solution of HMDI (8.00 g,
49.7 mmol) in methylene chloride (100 mL) was added 4-hydroxy
TEMPO (17.55 g, 101.9 mmol, 2.05 equiv). The mixture was stirred
for 2 h at room temperature and DBTDL (100 mL, 0.174 mmol,
2. Experimental materials and procedures
0.0035 equiv) was added. Stirring was continued at room temper-
ature for an additional 48 h. After that time, the solvent was
removed under reduced pressure to yield a viscous orange oil. The
oil was macerated with a metal spatula while repeatedly washing
the resulting paste with hexanes until a peach-tan colored solid
was obtained. The solid was vigorously stirred with hexanes over-
night and filtered under vacuum. The solid was then dried under
vacuum to give a pale peach-colored powder (23.62 g) in 93% yield.
Infrared spectroscopic analysis of the solid indicated complete
disappearance of the –NCO band at 2262 cmꢁ1, hence complete
consumption of the isocyanate residues in the sample. IR (Thin
Film) 3332 (N–H), 2930 (C–H), 1703 (C]O), 1520 cmꢁ1. HRMS: Calc.
514.3730, Anal. 514.3796 HMDI UBTH2: HMDI UBT was character-
ized as its diamagnetic reduced derivative by heating a toluene
solution of HMDI UBT and 1.05 equiv of 9,10-dihydroanthrcene
overnight at 100 ꢀC (Fig. 3). The orange-peach color of the HMDI
2.1. Polymer
The polymer used in this investigation was low-density poly-
ethylene (LDPE). It was additive-free, to start with, and its proper-
ties were as follows: Melt index (measured in accordance with
ASTM D-1238; 190 ꢀC: I2 with 2.16 kg load) of 2.4 dg/min; density
(determined according to ASTM D-792) of 0.9200 g/cc. Detailed
characterization of this polymer was published in a previous paper
[10].
2.2. Synthesis of urethane TEMPO adducts
2.2.1. General
Unless otherwise indicated, all reactions were performed under
nitrogen in a dry box or on a vacuum line. Anhydrous solvents used
Hydrogen abstraction
A
•
R
OH
HO
O
O
R
R
R
R
R
O
•
O
Δ
•
•
H
H
H
H
Crosslinking
B
permanent
crosslink
•
•
X
O
C
O
H
N
N
O
N O
R'
N
H
R'
HO
N O
O N
O
O N
OH
N
C
O
O
thermoreversible crosslink
O
•
•
O
NH
N O
R'
HN
O
O N
X =
O
Fig. 1. Schematic of thermoreversible crosslinking with urethane chemistry.

B.I. Chaudhary et al. / Polymer 51 (2010) 153–163
155
O
H
N
O
N O
O N
O
N
H
O
HMDI UBT
HMDI
•
O
N
O
O
MDI
O N
O
NH
2
O
N O
NH
OH
MDI UBT
2/n PAPI 901
H2
C
HN
O
PT adduct
2/n
O
N
O
•
n
Fig. 2. Chemical structures of HMDI UBT, MDI UBT and PAPI-TEMPO adduct.
UBT solution bleached during reaction resulting in a colorless
solution at the conclusion of reaction. HMDI UBTH2 was isolated as
a white powder and was purified by recrystallization from a hot
toluene solution. The derivatization was determined to be quanti-
(100 mL, 0.167 mmol, 0.0042 equiv) was added via syringe. Stirring
was continued for 6 h at room temperature after which the addi-
tion of pentane caused the separation of a viscous orange oil. The
supernatant was decanted and set aside and the oil was twice
washed with diethyl ether, followed by pentane, resulting in
precipitation of a beige solid. The solid was washed with ether and
dried in vacuo to afford 16.2 g of a pale peach-colored powder. An
additional 3.0 g of material was recovered from the wash super-
natants by addition of pentane, giving a total yield of 19.2 g (80%).
Infrared spectroscopic analysis of the solid indicated complete
disappearance of the –NCO band at 2274 cmꢁ1, IR (thin Film) 3295
(N–H), 2967 (C–H), 1715 (C]O), 1533 cmꢁ1. HRMS: Calc. 594.3417,
Anal. 594.3486. HMDI UBTH2: MDI UBT was characterized as its
tative by 1H NMR analysis. 1H NMR (THF-d8)
d 1.13 (s, 12 H), 1.14 (s,
12 H), 1.26–1.55 (m, 12 H), 1.86 (m, 4 H), 3.04 (AB q, J ¼ 6.6 Hz, 4 H),
4.87 (m, 2 H), 6.11 (br s, 2 H), 6.53 (s, 2 H) ppm. 13C NMR (THF-d8)
20.90, 27.61, 31.22, 33.19, 41.61, 45.83, 59.55, 156.75 ppm. HRMS:
Calc. 512.3574, Anal. 512.3647.
Preparation of MDI UBT. A solution of 4-hydroxy TEMPO (15.56 g,
90.31 mmol, 2.26 equiv) in 3:1 v:v Et2O-toluene (100 mL) was
added to a solution of MDI (10 g, 40.0 mmol) in toluene (30 mL).
After 45 min of stirring at room temperature, dibutyltin dilaurate
OH
N
OH
N
O
N
O
N
O
O
O
O
toluene
100 °C
O
N
NH
O
O
N
H
H
NH
O
overnight
HMDI UBTH2
HMDI UBT
O
N
O
N
OH
N
OH
N
O
O
O
O
toluene
100 °C
O
O
O
HN
NH
O
HN
NH
overnight
MDI UBT
MDI UBTH2
Fig. 3. Derivatization of HMDI UBTH2 and MDI UBTH2.

156
B.I. Chaudhary et al. / Polymer 51 (2010) 153–163
diamagnetic reduced derivative by heating a toluene solution of
MDI UBT and 1.05 equiv of 9,10-dihydroanthrcene overnight at
100 ꢀC (Fig. 3). The orange-red color of the HMDI UBT solution
bleached during reaction resulting in a pale-yellow solution at the
conclusion of reaction. MDI UBTH2 was isolated as a white powder
and was purified by recrystallization from a hot toluene solution.
The derivatization was determined to be quantitative by 1H NMR
analysis. 1H NMR (THF-d8) 1.15 (s, 12 H), 1.18 (s, 12 H), 1.48 (t,
J ¼ 12.3 Hz, 4 H), 1.93 (dt, J ¼ 12.3, 1.4 Hz, 2 H), 1.97 (dt, J ¼ 12.3,
1.4 Hz, 2 H), 3.83 (s, 2 H), 5.00 (m, 2H), 6.58 (s, 2H), 7.02 (AB q,
J ¼ 8.5 Hz, 4 H), 7.34 (AB q, J ¼ 8.5 Hz, 4 H), 8.50 (s, 2 H) ppm. 13C
NMR (THF-d8) 20.58, 33.09, 41.31, 45.50, 59.42, 119.01, 129.86,
136.39, 138.74, 153.94 ppm. HRMS: Calc. 596.3574, Anal. 596.3639.
Preparation of PAPI-TEMPO adduct (PTadduct). A mixture of PAPIÒ
901 (5 mL, 46.2 mmol of isocyanate equivlants) and 4-hydroxy
TEMPO (9.95 g, 57.8 mmol, 1.25 equiv) in 50 mL THF was refluxed
for 23 h. After cooling to room temperature, the volatile materials
were removed under reduced pressure to yield a viscous brown
liquid. The liquid was repeatedly washed with diethyl ether and
then with hexanes to yield a granular solid. Filtration and drying in
vacuo produced a pale tan solid (10.5 g) representing a 75% yield. IR
formed over the course of several days. The crystals were collected
on a frit and dried in vacuo to give 2.65 g of product in 76 % yield. 1H
NMR (CDCl3)
3.16 (br m 4 H), 4.65 (br s 2 H), 4.91 (septet, JH–H ¼ 6.2 Hz 2 H) ppm.
13C {1H} NMR (CDCl3)
22.45, 26.59, 30.23, 40.90, 68.96,156.44 ppm.
d
1.23 (d, JH–H ¼ 6.2 Hz 12 H), 1.34 (m 4 H), 1.50 (m 4 H),
d
HRMS: Calc. 288.2049, Anal. 288.2128.
Preparation of MDI-iPrOH. In a 100 mL RBF equipped with
a
magnetic stir bar methylenediphenyldiisocyanate (2.00 g,
7.99 mmol) and dibutyltin dilaurate (DBTDL) (25 L, .044 mmol,
m
0.0035 equiv) were dissolved in isopropanol (30 mL). The MDI had
only partial solubility in isopropanol and the slurry was stirred 48 h
at ambient temperature. After that time the volatile materials were
removed in vacuo to provide a white powder. The powder was
dissolved in 20 mL of CH2Cl2 and was layered with hexanes. Long
fibrous crystals formed over the course of several days. The white
solid was collected on a frit and was dried in vacuo to give 0.67 g of
desired product in 23% yield from a single crop. 1H NMR (CDCl3)
d
1.32 (d, JH–H ¼ 6.2 Hz 12 H), 3.91 (br s 2 H), 5.04 (septet, JH–
¼ 6.2 Hz 2 H), 6.66 (br s, 2 H), 7.12 (AB quartet, JH–H ¼ 8.7 Hz 4 H),
H
7.32 (AB quartet, JH–H ¼ 8.6 Hz 4 H),ppm. 13C {1H} NMR (CDCl3)
d
22.31, 40.76, 68.82, 100.11, 119.00, 136.22, 136.27, 153.40 ppm.
(Thin Film) 3311 (N–H), 2988 (C–H), 1722 (C]O), 1533 cmꢁ1
.
HRMS: Calc. 370.1893, Anal. 370.1968.
2.3. Thermal studies of urethane TEMPO adducts
2.5. Thermal investigation of HMDI-iPrOH and MDI-iPrOH
model compounds
Thermal transitions of the PT adduct and HMDI UBT were
determined by differential scanning calorimetry (DSC) in a closed
pan, using Thermal Analysis (TA) Model 2920. The sample was
heated from ꢁ50 ꢀC to 150 ꢀC at a rate of 10 ꢀC/min (1st heat);
cooled to ꢁ50 ꢀC at a rate of 10 ꢀC/min; and heated to 300 ꢀC at
a rate of 10 ꢀC/min (2nd heat).
The thermal stability of the PT adduct was investigated by
Thermal Gravimetric Analysis (TGA) using TA Instruments High
Resolution TGA 2950 version 5.4A. TGA testing was conducted
under nitrogen by raising the temperature from 20 ꢀC to 800 ꢀC at
a rate of 10 ꢀC/min, to determine the weight loss as a function of
temperature.
The thermal stability of the PT adduct was also studied by hot
cell FT-IR. Samples were analyzed via transmission between 25 mm
round KBr disks. Hot cell FT-IR data were collected on a Nicolet
Magna 750 FT-IR spectrometer running on Omnic E.S.P. software
and fitted with a Thermo Electron Model 0019-200 heated cell and
a DTGS detector. Temperature was controlled using a Scientific
Instrument Inc (SIS) Model CT-101 temperature controller. The
assembled cell, including clean KBr disks was placed in the nitrogen
purged sample compartment, heated to 150 ꢀC and used as the
background spectrum. Acquisition parameters were set at 64 co-
added scans, 4 cmꢁ1 resolution and triangular apodization was
employed. The cell was removed, partially disassembled, sample
inserted, reassembled, and placed back in the spectrometer. Spec-
tral collection for the PT adduct started at 150 ꢀC and raised in 10 ꢀC
increments to 280 ꢀC. Samples were given 5 min at each tempera-
ture to equilibrate prior to spectral collection. All data were plotted
in absorbance relative to the empty cell.
Thermal transitions of these urethane adducts were investi-
gated by closed capillary DSC. A method for using glass capillaries
to prevent sample loss has been previously described [12,13]. A
portion of the sample was placed in a glass capillary and while the
sample was frozen (using a cold finger at liquid nitrogen temper-
atures) the glass capillary was flame sealed. Special silver capillary
holders were used for the measurement as well as temperature and
heat flow calibration. The sample was re-weighed after the DSC
scan to assure that no sample mass was lost.
TGA/DSC/MS experiments on the model compounds were con-
ducted in open pans using TA Instruments 2960 SDT running
operating system V3.0F and with a MS interface via a heated sample
line. After loading, the sample was ramped rapidly to 160 ꢀC under
a stream of nitrogen where it was held in isothermal mode for
a period of 6 h during which mass change, heat flow and selected
ions were monitored.
2.6. Melt blending of peroxides and urethane TEMPO adducts
with polyethylene
Peroxides and/or urethane TEMPO adducts (ranging in
concentration from 1 wt% to 4 wt%) were blended with poly-
ethylene using a Brabender mixing bowl. The polymer and nitroxyls
(except peroxide) were mixed for 3 min at 125 ꢀC, followed by
additional 4 min mixing with peroxide at that temperature, in the
mixing bowl. The temperature was low enough to avoid significant
decomposition of the peroxide, while still melting the polymer. The
free-radical generators studied were: 2,5-bis(tert-butylperoxy)-
2,5-dimethylhexane [LuperoxÒ 101 (L101)] and 2,5-bis(tert-butyl-
peroxy)-2,5-dimethyl-3-hexyne [LuperoxÒ 130 (L130)]. These two
peroxides were used (instead of more commonly available dicumyl
peroxide) in order to minimize the amounts of methyl radicals
spontaneously generated from peroxide decomposition, since these
can also be trapped by nitroxyls thereby leading to decreased
nitroxyl trapping of carbon-centered polymer radicals [10]. The
nitroxyls evaluated were HMDI UBT, MDI UBT, and PT adduct
(Fig. 2).
2.4. Model studies of urethane adducts with isopropanol (iPrOH)
Preparation of HMDI-iPrOH. A 100 mL RBF equipped with
a magnetic stir bar was charged with isopropanol (30 mL). Hexam-
ethylene diisocyanate (2.04 g, 2.00 mL 12.2 mmol) and dibutyltin
dilaurate (DBTDL) (25 mL, 0.0425 mmol, 0.0035 equiv) were added
successively. The mixture was stirred 48 h at ambient temperature.
After 48 h, the volatile materials were removed in vacuo to provide
a white powder. The powder was dissolved in 20 mL of CH2Cl2 and
the solution was layered with hexanes. Long fibrous crystals were

B.I. Chaudhary et al. / Polymer 51 (2010) 153–163
157
2.7. Crosslinking kinetics and solid/melt state properties
of crosslinked polyethylene
N
N
X
X
ð
uliÞ2
uli
1 þ ðuliÞ2
G0ð
u
Þ ¼
gi
and G00ð
u
Þ ¼
gi
1 þ ðuliÞ2
i ¼ 1
i ¼ 1
The kinetics of free-radical (peroxide) grafting of urethane
TEMPO adducts to polyethylene were studied at temperatures of
160 ꢀC, 180 ꢀC and 200 ꢀC using a Moving Die Rheometer from
Alpha-Technologies (Model MDR 2000), set at 100 cycles/min and
deformation amplitude of 0.5ꢀ. Details of the experimental proce-
dure have been described elsewhere [10]. Comparisons were made
with peroxide crosslinked polyethylene without any urethane
TEMPO adducts in the composition. The products from the MDR
experiments were subjected to extraction in boiling decalin or
cyclohexane in order to measure the gel fraction. This method,
ASTM D-2765, Method A, determines the percentage of a sample
that is soluble in decalin (190 ꢀC boiling point) or cyclohexane
(80 ꢀC boiling point) over 6 h. The material which is insoluble, and
therefore does not pass through a 120 mesh screen, is the cross-
linked portion of the material (and is reported as the gel content).
Specimens were prepared for the following property evalua-
tions by compression molding at 160 ꢀC until complete decompo-
The Relaxation Spectrum Index [15], a measure of the breadth of
the relaxation index, was calculated as the ratio of the first two
moments of the relaxation spectrum:
N
N
X
X
gi
g
ili
RSI ¼ lII
3. Results and discussion
3.1. Synthesis of urethane TEMPO adducts
=
lI
;
where lI
¼
and lII
¼
i ¼ 1 gi=li
gi
i ¼ 1
The synthesis of the targeted urethane TEMPO adducts was
effected using standard procedures developed for isocyanate-
alcohol coupling reactions. In general, secondary alcohols could be
efficiently reacted with aliphatic isocyanates using dibutyltin
dilaurate as a catalyst. Aromatic isocyanates could be reacted using
triethylamine as a catalyst, though reactions are generally faster
with the tin catalyst. Molecular characterization of the urethane
nitroxyls was attempted by NMR, but the species are paramagnetic
and so the NMR spectra could not be used to characterize structure.
Instead, the paramagnetic urethane TEMPO adducts derived from
HMDI and MDI were converted to diamagnetic species by reaction
with 9,10-hydroanthracene. Heating toluene solutions of HMDI UBT
or MDI UBT at 100 ꢀC overnight with a slight excess of 9,10-dihy-
droanthracene afforded hydrogen atom transfer and quantitative
conversion to the corresponding TEMPOH species, permitting
characterization by 1H and 13C NMR spectroscopy.
sition of peroxide, followed by aging for
1 week at room
temperature, to fully crosslink those compositions that contained
peroxide and/or urethane TEMPO adducts.
Dynamic modulus was measured as a function of temperature
using a dynamic mechanical spectrometer (DMS). The DMS testing
on samples (prepared by compression molding at 160 ꢀC for
120 min) was performed on an ARES controlled strain rheometer
(TA Instruments) equipped with dual cantilever fixtures for torsion
testing. A 1.5 mm plaque was pressed and cut into a bar of
dimensions 32 mm ꢂ 12 mm. The sample was then clamped at both
ends between the fixtures separated by 10 mm (grip separation DL)
and subjected to successive temperature steps from 30 ꢀC to 220 ꢀC
(2 ꢀC per step). At each temperature, the torsion moduli G0 and G00
were measured at an angular frequency of 10 rad/s, the strain
amplitude being maintained between 0.1% and 3% to ensure that
the torque was sufficient and that the measurement remained in
For the polymeric aromatic system PAPI 901, 3 h reaction time
with triethylamine as the catalyst was insufficient to achieve
complete conversion of the starting materials (at an initial
concentration of approximately 1 M). Infrared spectroscopy
provided a convenient means of monitoring the progress of the
reaction. After 3 h, a significant quantity of PAPI 901 remained, as
was evident from the intense band at 2266 cmꢁ1, and so the reac-
tion was maintained for 20 h whereupon a second sampling indi-
cated complete conversion (Fig. 4). The product was isolated from
the reaction mixture as a viscous oil. However, as with its two
congeners, the initially isolated product could be induced to solidify
by maceration of the oil with a non-solvent (e.g., ether or alkane).
This procedure also permitted the removal of unreacted 4-hydroxy
TEMPO from the reaction mixture. Infrared analysis of the resin
purified in this manner indicated that no unreacted TEMPO or
isocyanate functionality remained (Fig. 5).
An initial attempt to prepare the analogous product between
MDI and 4-hydroxy TEMPO was made without a catalyst; however,
when no reaction was evident after a period of 45 min, dibutyltin
dilaurate was added. The addition of the tin catalyst led to imme-
diate reaction as indicated by a rise in the pot temperature. After 6 h
of stirring without the application of an external heat source, the
reaction was complete by IR spectroscopy. In contrast to the PAPI
901 adduct, the MDI adduct remained soluble in the reaction
medium. The addition of pentane induced precipitation of the
crude product as a viscous oil that could be treated similarly to the
PAPI adduct to obtain a purified solid product.
the linear regime. The tangent of the loss angle, tan(d), was
obtained as G00/G0 at every temperature. An initial static force of 10 g
was maintained (auto-tension mode) to prevent slack in the sample
when thermal expansion occurred. As a consequence, the grip
separation
DL increased with temperature, particularly above the
melting or softening point. The test stopped at the maximum
temperature or when the gap between the fixtures reached 65 mm.
Temperature-dependent probe penetration experiments were
performed using a TA instrument Thermo-Mechanical Analyzer
(TMA) on samples (prepared by compression molding at 160 ꢀC for
120 minutes). The sample was cut into an 8 mm disk (thick-
ness ¼ 1.5 mm). A 1 mm diameter cylindrical probe was brought to
the surface of the sample and a force of 1 N (102 g) was applied. As
the temperature was varied from 30 ꢀC to 220 ꢀC at a rate of 5 ꢀC/
min, the probe penetrated into the sample due to the constant load
and the rate of displacement was monitored. The test ended when
the penetration depth reached 1 mm.
Dynamic oscillatory shear measurements were conducted using
a
Weissenberg Rheogoniometer on samples (prepared by
compression molding at 160 ꢀC for 100 min) at temperatures
ranging from 140 ꢀC to 220 ꢀC, to determine if the urethane-
crosslinked compositions would flow at elevated temperatures and
to compare melt rheological properties with those of uncrosslinked
polyethylene. Each frequency sweep from 0.1 to 100 rad/s took
about 15 min to complete (including about 3 min initially to
equilibrate at test temperature). The data were then fit by the
commercial IRISÔ software using the method of Baumgaertel and
Winter [14] to determine the discrete relaxation spectrum (gi, li) by
The adduct derived from the aliphatic isocyanate HMDI also
required the use of the tin catalyst for its preparation. Stirring HMDI
and 4-hydroxy TEMPO together in methylene chloride solution
with 0.35 mol percent tin catalyst for a period of 96 h effected
complete conversion. In subsequent experiments, we found that
the reaction is essentially complete within 48 h. In contrast to the
work of Lizzotte et al. [16], the HMDI urethane bis-TEMPO could be
fitting the dynamic moduli G0 and G00 over the range of frequency
u:

158
B.I. Chaudhary et al. / Polymer 51 (2010) 153–163
The MDI-TEMPO exhibited a melt feature at 145 ꢀC and a minor
endotherm with an onset at 170 ꢀC. Continued heating above
190 ꢀC led to two large exothermic events with peak temperatures
identical to that for the HMDI-TEMPO adduct. Similarly, the PT
adduct showed a melt transition at 93 ꢀC followed by large exo-
therm of (259.7 J/g) with onset and peak temperatures of 180 ꢀC
and 232 ꢀC, respectively. As before, we interpret the exothermic
thermal event to correspond to a decomposition reaction involving
the urethane functional group.
In order to determine if the potentially reactive nitroxyl group
was contributing to the decomposition reaction, we also prepared
and examined model compounds (Fig. 6) derived from the reaction
of isopropanol (iPrOH) with HMDI and MDI. In contrast to the
urethane TEMPO agents, the urethane-iPrOH agents did not exhibit
exothermic decomposition up to 250 ꢀC (Fig. 7). HMDI-iPrOH
exhibited two reversible melt features at 85 ꢀC and 105 ꢀC. Begꢁin1-
ning at 179 ꢀC, the onset of an endothermic event (24.1 kcal mol
)
was observed continuing with a peak temperature of 234 ꢀC. For
MDI-iPrOH, a single melt feature at 152 ꢀC was accompanied by the
endothermic reversion (33 kcal molꢁ1
) with onset and peak
Fig. 4. Infrared spectral monitoring of the reaction between PAPI 901 and 4-hydroxy
TEMPO.
temperatures of 230 ꢀC and 273 ꢀC, respectively. With consider-
ation of the putatively greater stability of urethanes based on
aliphatic isocyanates in comparison to urethanes derived from
aromatic isocyanates, we were somewhat surprised that endo-
thermic reversion of HMDI-iPrOH occurred at a lower temperature
than that of MDI-iPrOH. However, further support for the greater
lability of iPrOH in HMDI-iPrOH in comparison to MDI-iPrOH was
obtained through TGA/DSC/MS studies of the model compounds. In
these experiments, the samples were ramped to 160 ꢀC and held for
6 h while heat flow, mass loss and selected masses were monitored.
Mass spectral data during the experiment indicated that mass loss
was occurring by loss of iPrOH. Unfortunately, at 160 ꢀC, the low
rate of reversion precluded detection of any change in the heat load
of the sample. After 6 h, the HMDI-iPrOH sample experienced a 36%
mass loss compared to only 4% mass loss for the iPrOH-MDI
compound. The theoretical mass loss for HMDI-iProH and MDI-
iPrOH arising from reversion and iPrOH volatilization were 42% and
32%, respectively. Therefore, despite literature claims concerning
relative stabilities, reversion in the aliphatic urethane of our system
appears more facile than reversion in the aromatic system.
The reason for the discrepancy in DSC behavior between the
TEMPO adducts and model compounds became apparent when 4-
hydroxy TEMPO itself was analyzed by DSC. Following a melt at
52 ꢀC, 4-hydroxy TEMPO exhibited two large exothermic events
with an initial onset temperature of 178 ꢀC. These decomposition
events were superimposable with those observed for the urethane
TEMPO adducts indicating that the nitroxyl functionality was
responsible for the discrepancy in thermal behavior between the
urethane TEMPO adducts and the urethane-iPrOH model
compounds. Note that, since the DSC scans were done in sealed
glass capillaries to prevent weight loss (and none was found),
sublimation of the sample can be excluded. This result places
a practical constraint on the application of urethane TEMPO agents
isolated as a solid when the oil was washed exhaustively with
a non-solvent.
3.2. Thermal stabilities of urethane TEMPO reagents
and model studies
In order to gain insight into the thermochemical attributes of
the urethane-based crosslinking agents, we examined the thermal
stabilities of the urethane TEMPO adducts and urethane model
compounds using DSC and TGA/DSC/MS techniques. DSC analysis of
HMDI-TEMPO (2nd heat) indicated an endothermic feature at
122 ꢀC corresponding to the melt of the crystalline solid followed
by two large exotherms (74 kcal molꢁ1) with an onset at 190 ꢀC and
peak temperatures of 237 and 250 ꢀC. Since the thermal reversion
of a urethane group would be manifest as an endothermic feature
(ca. 20 kcal molꢁ1), we have interpreted the large exotherm
therefore to correspond to an unknown decomposition of the
HMDI-TEMPO agent. Support for this assertion was obtained from
the corresponding cooling cycle which showed irreversibility by
the lack of the expected exo- and endotherm features.
For the reagents derived from aromatic isocyanates, MDI-
TEMPO and PAPI-TEMPO, slightly different behavior was observed.
O
O
O
O
O
O
HN
NH
O
O
N
H
N
H
HMDI-iPrOH
MDI-iPrOH
Fig. 5. Infrared spectral analysis of the purified reaction product (PT adduct) between
Fig. 6. Model compounds (MDI-iPrOH and HMDI-iPrOH) synthesized from iso-
PAPI 901 and 4-hydroxy TEMPO.
propanol and diisocyanates.

B.I. Chaudhary et al. / Polymer 51 (2010) 153–163
159
Fig. 7. Closed capillary DSC studies of model compounds MDI-iPrOH and HMDI-iPrOH.
for grafting and crosslinking applications requiring that the grafting
step must be conducted at temperatures below 180 ꢀC to avoid
significant TEMPO decomposition. It should be noted that DSC data
was obtained under ‘‘artificially-concentrated’’ conditions and that
similar decompositions may not be relevant under the relatively
dilute concentrations of urethane TEMPO encountered in polymer
processing. Nevertheless, when interpreting the rheological data
that follow, one must consider that some free nitroxyl-bearing
species may exist at elevated temperatures, leading to decompo-
sition and an overall reduction in crosslinking efficiency.
homolysis of a benzylic-TEMPO bond (BDE ¼ 30 kcal molꢁ1) would
be 9 h at 100 ꢀC, the half-life for homolysis of a secondary carbon-
TEMPO bond would be 57 h at 200 ꢀC. Since LDPE possesses
primary, secondary and tertiary C–H bonds through which
peroxide-mediated grafting of TEMPO reagents may occur, it is
possible to generate all three types of alkoxyamines in the present
studies. However, in model compound studies, we have deter-
mined that with the use of LuperoxÒ 101 and LuperoxÒ 130 as
initiators, greater than 95% of the TEMPO grafts reside at the more
thermally stable primary or secondary positions on the polymer
Testing of the PT adduct by hot cell FT-IR, in which the
temperature was varied from 160 ꢀC to 260 ꢀC in 10 ꢀC increments,
showed a decrease in urethane linkages starting from 180 ꢀC and
the appearance of what could be interpreted as an isocyanate band
starting around 200 ꢀC, thus providing qualitative evidence of
urethane reversion at these elevated temperatures (Fig. 8).
1.2
180 ºC
200 ºC [1 h]
260 ºC
b
1.1
1.0
0.9
0.8
0.7
0.6
0.5
0.4
0.3
3.3. Peroxide and urethane crosslinking of LDPE
2450
2350
2250
An important consideration in the crosslinking of polymers with
nitroxyl-bearing species, especially in light of nitroxyl decomposi-
tion, is the stability of the graft, since thermal reversion and/or
decomposition of the resulting alkoxyamine is known to occur
under certain conditions. For example, Otsuka et al. have exploited
the lability of nitroxyl-benzylic carbon bonds in styrenic polymers
wherein reversible homolysis of the C–O bond provided the basis
for macroinitators [17], molecular weight control [18] and cross-
linking [19]. However, without sufficient stabilization of incipient
carbon radical, C–O bond homolysis does not appreciably occur at
temperatures below 200 ꢀC. While the homolytic bond dissociation
energies (BDE) of benzylic-TEMPO bonds have been measured to
range from 25–31 kcal molꢁ1 enabling homolysis at temperatures
as low as 100 ꢀC [20], computational estimates of bond dissociation
energies for primary and secondary alkoxyamines derived from
TEMPO fall in the range of 40–44 kcal molꢁ1 [21]. In contrast,
calculations for tertiary alkoxyamines predicted significantly less
stability (BDE ¼ 33 kcal molꢁ1). Therefore, while the half-life for
a
-1
Frequency, cm
Fig. 8. FTIR spectra of PAPI-TEMPO adduct taken at various temperatures. The spec-
trum at 200 ꢀC was acquired after holding the sample for 1 h at that temperature. The
feature marked ‘‘a’’ indicates increased absorbance attributed to the concentration of
isocyanate groups, while ‘‘b’’ marks the loss of absorbance at the frequency charac-
teristic of the carbamate (urethane) C]O stretch.

160
B.I. Chaudhary et al. / Polymer 51 (2010) 153–163
Table 1
chain (unpublished results). Consequently, polymer processing at
temperatures below 200 ꢀC should not result in graft loss from
alkoxyamine decomposition [10].
Gel contents (weight percent) of crosslinked polyethylene.
Weight percent
gels (decalin
extraction)
Weight percent
gels (cyclohexane
extraction)
Unless otherwise noted, the crosslinking experiments in the
moving die rheometer (MDR) were conducted at a temperature of
160 ꢀC in order to minimize exothermic decomposition and
urethane reversion of the TEMPO adducts at higher temperatures.
Under these conditions, decompositions of the PT adduct, MDI UBT
and HMDI UBT are unlikely to have occurred, based on the DSC
results in which exothermic activity was not observed below 170–
190 ꢀC. The results of moving die rheometer experiments con-
ducted at 160 ꢀC to evaluate the effectiveness of grafting and
crosslinking with the PT adduct are plotted in Fig. 9. The addition of
4 wt% PT adduct (to polymer compositions containing 1.0 wt% L130
peroxide and 1.5 wt% L130 peroxide) resulted in an unusual flat
induction period. As shown previously [10], the induction period
corresponds to the initial trapping of polymeryl radicals (and
adventitious methyl radicals generated from secondary peroxide
decomposition reactions) by the PT adduct generating grafted
urethane TEMPO moieties. Once the concentration of free urethane
TEMPO reagent is depleted, the pendant grafted nitroxyl group can
participate in radical trapping, leading to crosslinking and an
expected rise in torque. In the present case, no crosslinks were
formed in the initial 25–30 min. Thereafter, radical trapping was
increasingly due to the pendant nitroxyl moieties on polymer-
bound PT adduct, leading to an increase in MDR torque.
The ultimate degree of crosslinking was measured by the final
MDR torque as well as the gel content measured by extraction in
boiling decalin or cyclohexane (Table 1). The gel contents were not
significantly different with and without PT adduct, at a fixed
concentration of peroxide. This indicates that the PT adduct func-
tioned as a crosslinker and did not just inhibit carbon–carbon
crosslinking. Of course, it is possible that some of the PT adduct may
have been lost during mixing by mechanical promoted degradation
and/or volatilization. Noteworthy is the observation that in the
presence of PT adduct, the ultimate torques were less than those
obtained with peroxide alone. The reason for this is not clear, but it
is possible that with the use of a polyfunctional crosslinking agent,
the ultimate crosslink architecture (e.g. star-type polymers) could
look very different from what would result from simple peroxide
carbon-carbon crosslinking or even a dimeric TEMPO adduct. This
difference may manifest itself in a lower overall crosslink density as
measured by ultimate MDR torque.
160 ꢀC/4 h in MDR
1.0 wt% L130
1.5 wt% L130
82
87
74
86
Not available
Not available
Not available
91
1.0 wt% L130 þ 4 wt% PT adduct
1.5 wt% L130 þ 4 wt% PT adduct
160 ꢀC/80 min in MDR
1.25 wt% L101
91
1
61
85
63
79
1.25 wt% L101 þ 3 wt% HMDI UBT
1.25 wt% L101 þ 3 wt% MDI UBT
equal weight percent, the molar concentration of nitroxyl varied
slightly. The flat induction period observed with PT adduct (Fig. 9)
was absent with MDI UBT and HMDI UBT, probably due in part to
the faster decomposition rate of L101 in comparison to L130.
Despite the fact that both MDI UBT and PAPI-TEMPO effect cross-
linking through aromatic/aliphatic urethane linkages, they differed
significantly in their crosslinking behavior. Unlike PAPI-TEMPO, the
MDI UBT yielded approximately the same ultimate torque as the
composition with peroxide alone (Fig. 10). In contrast, the gel
content determined by extraction in boiling decalin was substan-
tially lower with MDI UBT as crosslinker. This is consistent with
urethane reversion during the elevated temperatures of the
extraction procedure and that lower extractables from polymers
crosslinked with PAPI-TEMPO derived from inherently more stable
crosslinks.
Surprisingly, HMDI UBT yielded very different crosslinking
results, exhibiting ultimate torque values less than half of those of
its aromatic congeners (Fig.10). Enthalpic considerations would not
have predicted this, since the literature reports that urethanes
derived from aliphatic alcohols and aliphatic isocyanates are stable
to 250 ꢀC while urethanes derived from aliphatic alcohols and
aromatic isocyanates are only stable to about 200 ꢀC [11]. In
contrast, the lower degree of crosslinking achieved at 160 ꢀC using
HMDI UBT is supported by our model compound studies which
showed that reversion occurred at a lower temperature and with
a lower net enthalpy for the urethane formed from a secondary
alcohol and an aliphatic isocyanate in comparison to the urethane
derived from an aromatic isocyanate, However, the degree of
reversion that must have occurred at 160 ꢀC is somewhat surprising
if bond thermodynamics is the lone consideration, since even the
HMDI-iPrOH model compound did not show reversion onset until
The crosslinking of LDPE with L101 peroxide and various
urethane bis-TEMPO derivatives at 160 ꢀC is shown in Fig. 10, and
the corresponding gel contents are given in Table 1. The theoretical
molecular weights of the HMDI UBT and MDI UBT were 512 g/mol
and 595 g/mol, respectively. Since the UBT reagents were used at
0.35
0.30
0.25
0.5
0.4
0.3
1.25 wt% L101
0.20
1.25 wt% L101 + 3 wt% HMDI UBT
0.15
1.25 wt% L101 + 3 wt% MDI UBT
0.2
0.1
0.0
1.0 wt% L130
0.10
0.05
0.00
1.5 wt% L130
1.0 wt% L130 + 4 wt% PT
1.5 wt% L130 + 4 wt% PT
0
20
40
Time (minutes)
60
80
0
40
80 120
Time (minutes)
160
200
240
Fig. 10. Crosslinking of LDPE with urethane bis-TEMPO derivatives at 160 ꢀC (with
L101 peroxide).
Fig. 9. Crosslinking of LDPE with PT adduct at 160 ꢀC (with L130 peroxide).

B.I. Chaudhary et al. / Polymer 51 (2010) 153–163
161
nearly 180 ꢀC. One possible explanation is that urethane reversion
100
MDR @
180°C
at elevated temperature results in free HMDI or MDI that can
evaporate from the melt pool resulting in an irreversible loss of
crosslinking density. Given the comparatively lower reversion
temperature and greater volatility of HMDI, it is reasonable to
suppose that it would evaporate faster than the MDI from the melt
pool leading to lower torque values.
MDR @
200°C
80
60
40
20
0
MDR @
160°C
It is interesting to note that the composition containing PT
adduct and L130 peroxide crosslinked at approximately the same
rate, and to the same ultimate torque, as one containing HMDI UBT
and L130 peroxide (Fig. 11). Since HMDI UBT led to slower and less
effective crosslinking than MDI UBT, it can be inferred that the PT
adduct was also slower and less efficient than MDI UBT at cross-
linking the polymer. This may be due to slower diffusion of the
bulkier PT adduct, particularly after its ‘‘free’’ concentration was
depleted due to grafting to the polymer, such that the pendant
nitroxyls on polymer-bound adduct were not as readily accessible.
The gel contents of fully crosslinked specimens out of MDR, as
determined by 6 h extraction in boiling decalin (190 ꢀC) or cyclo-
hexane (80 ꢀC), are plotted in Fig. 12. The tests were also conducted
on the uncrosslinked polyethylene (‘‘no peroxide, no TEMPO’’). The
crosslinking temperatures in the MDR were 160 ꢀC, 180 ꢀC or
200 ꢀC. As expected, the gel content of the uncrosslinked poly-
ethylene was measured as zero in both solvents. The compositions
crosslinked with peroxide alone yielded similar gel contents after
extraction in both solvents and the gel contents did not vary
significantly with MDR test temperature. This confirms that
permanent (stable) crosslinks were obtained with peroxide alone.
On the other hand, the addition of the PT adduct or urethane bis-
TEMPO derivatives generally led to lower values of gel content
measured by decalin extraction in comparison with the measure-
ments obtained by cyclohexane extraction, showing that many of
the crosslinks that were present at 80 ꢀC had broken over the long
period of time at 190 ꢀC, further substantiating the observation that
urethane reversion occurred at elevated temperatures. In partic-
ular, the difference in gel contents with the two solvents was
dramatically different with HMDI UBT, presumably due to the
relatively higher amounts of the aliphatic isocyanate generated
upon urethane reversion and the associated drive in the equilib-
rium towards more urethane reversion. The fact that the gel
contents of HMDI UBT crosslinked polyethylene measured by dec-
alin extraction were only 1wt% indicates that it contained very few
(if any) carbon–carbon crosslinks. This further suggests that the
nitroxyls were extremely efficient at trapping polymer radicals and
functionalizing the polymer. With the PT adduct and MDI UBT, the
crosslinking temperature had an effect on the measured gel
no peroxide, no TEMPO
1.25 wt% L101
1.5 wt% L130
1.25 wt% L101 + 3 wt% HMDI UBT
1.25 wt% L101 + 3 wt% MDI UBT
1.5 wt% L130 + 4 wt% PT
0
20
40
60
Gel Content (%) by Decalin Extraction
80
100
Fig. 12. Effect of extraction solvent (temperature) on gel content of uncrosslinked and
crosslinked LDPE.
temperature of 200 ꢀC. This is consistent with any decomposition of
the nitroxyls that might have occurred at this temperature (as
shown before in the DSC analyses), prior to grafting to the polymer,
as well as loss of diisocyanates at the elevated temperatures in
MDR. Although the PT adduct had earlier been observed to be a less
efficient crosslinker than MDI UBT, the resulting crosslinks did
appear to be more thermally stable as evidenced by the results
plotted in Fig. 12.
The results of DMS testing and rate of probe penetration (by
TMA) testing are presented in Figs.13–15. Note that the gel contents
reported in these figures are from cyclohexane extraction after
80 min in MDR at 160 ꢀC. Interestingly, the modulus of LDPE
crosslinked with MDI UBT decreased to a plateau above the melting
point of the polymer, and only started decreasing again above
200 ꢀC, indicating that this was the onset temperature for signifi-
cant urethane reversion. This reversion temperature was consistent
with that observed in model compound studies. No such reversion
was observed in the LDPE sample crosslinked with L101 peroxide
alone, confirming that the reversion was not due to polymer
degradation. The expected decrease in elasticity of the LDPE
crosslinked with MDI UBT during reversion was confirmed by the
increase in tan(d) (Fig. 14). The HMDI UBT crosslinked specimen did
not show evidence of a plateau above the crystalline melting point,
but instead showed a continuous linear decline in modulus up to
contents, with the lowest gel contents obtained at
a MDR
109
LDPE
+ 1.25 wt% L101
108
0.35
0.30
0.25
0.20
0.15
+ 1.25 wt% L101 + 3 wt% HMDI UBT
+ 1.25 wt% L101 + 3 wt% MDI UBT
107
More elastic
85% gels
106
79% gels
105
63% gels
0.10
1.25 wt% L101 + 3 wt% HMDI UBT
1.25 wt% L101 + 3 wt% MDI UBT
1.0 wt% L130 + 3 wt% HMDI UBT
1.0 wt% L130 + 4 wt% PT
104
0.05
0.00
0% gels
103
0
10
20
Time (minutes)
30
40
30
80
130
Temperature (°C)
180
230
Fig. 11. Crosslinking of LDPE with PT Adduct and urethane bis-TEMPO derivatives at
180 ꢀC.
Fig. 13. Temperature-dependent dynamic storage modulus (G0) of uncrosslinked and
crosslinked LDPE.

162
B.I. Chaudhary et al. / Polymer 51 (2010) 153–163
100000
LDPE
LDPE @ 140°C
LDPE @ 220°C
+ 1.25 wt% L101 + 3 wt% HMDI UBT @ 220°C
+ 1.25 wt% L101 + 3 wt% MDI UBT @ 220°C
10
1
+ 1.25 wt % L101
+ 1.25 wt % L101 + 3 wt% HMDI UBT
+ 1.25 wt% L101 + 3 wt% MDI UBT
10000
1000
100
0.1
0.01
RSI
29
LDPE at 140°C
LDPE at 220°C
15
82
442
+ 1.25 wt% L101 + 3 wt% HMDI UBT at 220°C
+ 1.25 wt% L101 + 3 wt% MDI UBT at 220°C
More elastic
0.1
1
Frequency (rad/sec)
10
100
30
80
130
Temperature (°C)
180
230
Fig. 16. Melt viscosity and relaxation spectrum index (RSI) of uncrosslinked and
crosslinked LDPE.
Fig. 14. Temperature-dependent dynamic tan(d) of uncrosslinked and crosslinked
LDPE.
a desirable means of improving melt strength (a key requirement
for extrusion foaming, blow molding, thermoforming and other
plastic processing). The oscillatory shear data were transformed
into relaxation spectra, which in turn were used to compute values
of relaxation spectrum index (RSI) [15]. The higher RSI values and
shear-thinning characteristics of urethane-crosslinked LDPE were
indicative of rheology modification of the polymer through
coupling of polymer chains, leading to a broadening of the distri-
bution of relaxation times towards long times.
230 ꢀC. The decreasing modulus and absence of a plateau indicates
significant urethane reversion up to the maximum test tempera-
ture of 220 ꢀC, in agreement with greater reversion of the aliphatic
urethane predicted by model compound studies. Of course, it is
reasonable to assume that some carbon-carbon crosslinks may also
have been present, in addition to the urethane crosslinks. The
results on rate of probe penetration (Fig. 15) followed the same
order as the dynamic modulus data (Fig. 13).
The shear-thinning behavior determined from dynamic oscil-
latory shear measurements is given in Fig. 16. Except for uncros-
slinked LDPE (which was tested at two different temperatures), all
others were only run at 220 ꢀC. The specimen crosslinked with L101
peroxide (with 85% gels in cyclohexane) could not be run since it
was thermoset. The composition crosslinked with MDI UBT (with
79% gels in cyclohexane and remarkably high modulus up to about
190 ꢀC) showed considerable shear-thinning at 220 ꢀC, such that
the high-shear viscosity of this composition approached that of
uncrosslinked LDPE at 220 ꢀC (even though its low-shear viscosity
was close to that of uncrosslinked LDPE at 140 ꢀC). That is, this
4. Conclusions
Three carbamate-linked dinitroxyl or polynitroxyl compounds
were synthesized and subsequently used for free-radical initiated
functionalization of polymers. The syntheses were effected simply
by the reaction of 4-hydroxy TEMPO with various diisocyanates
using triethylamine or a tin compound as catalyst. The reactions
were monitored by infrared spectroscopy and the products were
isolated as tan or peach-colored powders in good yields (75–93%).
The di- or polynitroxyl urethanes were used in combination
with peroxide to couple or crosslink polyethylene in the molten
state. The mechanism involved peroxide-induced formation of
a polymer radical, followed by rapid trapping of the radical via the
nitroxyl end group. In this way the bifunctional or polyfunctional
nitroxyl adducts were able to provide crosslinking sites, in addition
to carbon-carbon crosslinking arising from combination of polymer
radicals generated as a consequence of peroxide-induced hydrogen
abstraction. Urethane reversion was detected in the adducts as well
as the crosslinks at elevated temperatures. The types of urethane
groups in the crosslinks were observed to influence the rate and
degree of crosslinking, as well as kinetics of urethane reversion.
Remarkably high plateau modulus was achieved in aromatic/
aliphatic urethane-crosslinked polyethylene up to a temperature of
about 190 ꢀC. This urethane-crosslinked polyethylene, as well as
another crosslinked to lesser extent, both behaved like thermo-
plastics at sufficiently elevated temperatures. Thus, not only were
thermoset-like properties achieved with urethane-crosslinked
compositions, but the high degree of shear-thinning in the ther-
moplastic state was indicative of polymer chain coupling, which is
a desirable means of improving melt extensional properties. It is
anticipated that di- or polynitroxyl urethanes can also be used to
reversibly couple or crosslink propylene polymers and other poly-
mers. Although thermoreversible crosslinking has been demon-
strated, the ability of the systems to withstand multiple heating and
cooling cycles has not been ascertained.
crosslinked composition behaved like
a thermoplastic above
200 ꢀC, confirming that urethane reversion had occurred. The
composition crosslinked with HMDI UBT exhibited lower viscosity
and less shear thinning than that crosslinked with MDI UBT,
consistent with its relatively lower gel content. Both crosslinked
compositions showed evidence of polymer chain coupling, which is
2.5
LDPE
Gel Contents (%)
+ 1.25 wt% L101 + 3 wt% HMDI UBT
LDPE -------------------- 0
2
+ 1.25 wt% L101 + 3 wt% MDI UBT
+ L101 + 3 wt%
HMDI UBT ------------ 63
+ 1.25 wt% L101
1.5
+ L101 + 3 wt%
MDI UBT -------------- 79
1
+ L101 ----------------- 85
0.5
0
95
100
105
110
115
Sample temperature (C)
120
125
Fig. 15. Temperature-dependent rate of probe penetration in uncrosslinked and
crosslinked LDPE.

B.I. Chaudhary et al. / Polymer 51 (2010) 153–163
163
[6] Abraham D, George KE, Francis DJ. Die Angewandte Makromolekulare Chemie
1992;200:15–25.
This work also demonstrates that polymers can be functional-
ized with nitroxyls in the presence of free-radical generators, thus
providing the ability to graft functional groups onto a variety of
polymers. These functional groups might include hydroxyl, amine,
carboxyl, urethane and any other group that can be attached via
a permanent or reversible linkage to the 4-position of a TEMPO
derivative.
[7] Abraham D, George KE, Francis DJ. J Appl Polymer Sci 1998;67:789–97.
[8] Isac SK, George KE. J Appl Polymer Sci 2001;81:2545–9.
[9] Andersson LHU, Hjertberg T. Polymer 2006;47:200–10.
[10] Chaudhary BI, Chopin LJ, Klier J. Polymer Eng Sci 2007;47:50–61.
[11] Yang WP, Macosko CW, Wellinghoff ST. Polymer 1986;27:1235–40.
[12] Whiting LF, Labean MS, Eadie SS. Thermochim Acta 1988;136:231.
[13] Tou JC, Whiting LF. Thermochim Acta 1980;42:21.
[14] Baumgaertel M, Winter HH. Rheologica Acta 1989;28:511–9.
[15] Wasserman SH. SPE ANTEC 1997;55:1129.
References
[16] Lizotte JR, Anderson SG, Long TE. J Polymer Sci Part A Polymer Chem 2004;42:
1547–56.
[17] Higaki Y, et al. Macromolecules 2003;36:1494.
[1] Dorn M. Adv Polymer Technol 1985;5:87–97.
[18] Ostsuka H, et al. Macromolecules 2007;40:1429.
[19] Higaki Y, Ostsuka H, Takahara A. Macromolecules 2006;39:2121.
[20] Skene WG, et al. Macromolecules 1998;21:9103.
[21] Marsal P, et al. J Phys Chem A 1999;103:2899.
[2] Gulmine JV, Akcelrud L. J Appl Polymer Sci 2004;94:222–30.
[3] Bremner T, Rudin A, Haridoss S. Polymer Eng Sci 1992;32:939–43.
[4] Lazar M, et al. J Polymer Sci Part A Polymer Chem 2004;42:675–88.
[5] Wu G, et al. J Polymer Sci Part A Polymer Chem 1999;37:1541–8.