-
Molecular Characterization of Maleic Anhydride-Functionalized
Polypropylene
B. DE ROOVER,'* M. SCLAVONS,' V. CARLIER,' 1. DEVAUX,' R.
LECRAS,' and A. MOMTAZ'
'Laboratoire de Physique et de Chimie des Hauts PolymPres, 1
Croix d u Sud, 6-1348 Louvain-la-Neuve, Belgium, 2SoIvay S.A.
Direction Centrale des Recherches, Laboratoire Central, 31 0 rue de
Ransbeek, B-1 120 Bruxelles, Belgium
SYNOPSIS
This work deals with the molecular characterization of maleic
anhydride melt-functionalized polypropylene ( PP-g-MA) . The
functionalization mechanism, the nature, the concentration, and the
location of grafted anhydride species onto the polypropylene chain
are discussed. The polypropylene functionalization was performed
using a pre-heated Brabender Plas- tograph ( 190C, 4 min of mixing
time). Several concentrations of maleic anhydride and organic
peroxide were used for this study. In those experimental
conditions, the organic peroxide undergoes an homolytic rupture and
carries out a polypropylene tertiary hydrogen abstraction. The
resulting macroradical undergoes a 0-scission leading to a radical
chain end which reacts with maleic anhydride. When a termination
reaction occurs at this first step a succinic type anhydride chain
end is obtained. However, oligomerization of maleic anhydride is
found to occur more frequently leading to poly( maleic anhydride)
chain end. Concentration of both anhydride types and minimal length
of the grafted poly (maleic anhydride) were determined. A fraction
of maleic anhydride does not react with polypro- pylene or
homopolymerize leading to nongrafted poly (maleic anhydride). 0
1995 John Wiley & Sons, Inc. Keywords: polypropylene 0-scission
melt functionalization maleic anhydride poly( maleic anhydride)
INTRODUCTION
literature Survey
Maleic anhydride-functionalized polypropylene is of considerable
importance for application as a copol- ymer precursor in polymer
blends, as an adhesion promoter with glass or carbon fibers, and
even as a processing aid for recycling of plastics waste '-three
domains which received considerable attention in recent years.
Generally reported functionalization procedure consists in grafting
maleic anhydride in the presence of organic peroxide either in the
melt '-lo or in the solid state, 11,12 or in ~olu t ion .~ . '~ Few
particular methods were also reported: sus- pension method using
water l4 or toluene, l5 ene- reaction process, l6 melt process
where maleic an- hydride and peroxide are solubilized in a
s01vent.l~ Use of additive to overcome some side reactions is
mentioned.l7-l9 Grafting onto atactic polypropylene
* To whom all correspondence should be addressed. Journal of
Polymer Science: Part A Polymer Chemistry, Vol. 33, 829-842 (1995)
0 1995 John Wiley & Sons, Inc. CCC 0887-624X/95/050829-14
is also reported13 as well as functionalization of polypropylene
during its synthesis." Nevertheless, those studies need previous
complete characteriza- tion of the functionalization mechanism and
of the structure of PP-g-MA to be significant. Finally, sev- eral
articles concerning polyethylene and ethylene- propylene rubber
maleic anhydride grafting can also be considered as pertinent
referen~es.~l-~'
In any case, the most widespread method is the melt state
process often called "reactive extrusion method." A definite
molecular characterization of PP-g-MA resulting from the reactive
extrusion method has not been realized thus far and some
controversy remains about the concentration, lo- cation, and nature
o the grafted anhydride species. Anhydride can, from a chemical
viewpoint, be grafted at the chain end, along the chain or included
in the backbone.
The occurrence of homopolymerization of maleic anhydride during
the grafting of polypropylene is a particular subject of
controversy. Homopolymeri- zation of maleic anhydride could lead to
grafted or ungrafted poly (maleic anhydride) (PMA) .
829
-
830 DE ROOVER ET AL.
Gaylord et a1.9J8,39-45 propose mechanisms of homopolymerization
of maleic anhydride and graft- ing of the PMA during the
functionalization of polypropylene, polyethylene, and
ethylene-propyl- ene rubber. On the other hand, Russell46 invokes a
thermodynamic argument based on ceiling temper- ature of PMA which
would preclude any homopoly- merization of maleic anhydride at
temperature higher than 160C. The formation and consequently the
grafting of PMA during maleation of polyolefins in the melt would
therefore be impossible. But in both cases no explicit nor obvious
experimental ob- servations are given which prove the presence or
the absence of PMA in the maleated polyolefins. This particular
point will be discussed in detail herein.
From the literature, it can be deduced that the grafting
mechanism, concentration, structure, and location of the grafted
anhydride are to a large extent influenced by the grafting method
(solution, melt state, or solid state) and the reaction conditions
(temperature, pressure, concentration, solvent, ad- ditive, etc.) .
In this work the organic peroxide struc- ture, the polypropylene
grade, the melt temperature, and the reaction time were kept
constant. Only con- centration of maleic anhydride and organic
peroxide were varied. In this study, oxygen was assumed not to play
an important role in the reactions because the experiments were
realized under nitrogen at- mosphere. Although the present study
will be strictly valid only for the experimental conditions used
for this work it should be generalized providing those conditions
are not modified too much.
General Mechanism
Figure 1 summarizes all the possible mechanisms reported in the
literature for the grafting of maleic anhydride onto polypropylene
in the presence of or- ganic peroxide. Different pathways can
correspond to different experimental conditions.
The first steps of grafting mechanism are rela- tively well
established and can be summarized as follows: Homolytic scission of
each organic peroxide produces two radicals. The decomposition rate
(or half-life time of the peroxide) depends only on the
temperature. Polypropylene pending hydrogen is abstracted by a
radical attack which results in a new radical onto polypropylene.
Hydrogen abstraction from one of the tertiary carbons of the
polypropylene is generally mentioned for radical stability consid-
e r a t i o n ~ . ~ ~
Different possibilities exist for the second steps. They are
described in the following: The radical onto a polypropylene chain
can lead to a @-scission (re- action A ) or to a maleic anhydride
grafting (reaction B ) . The &scission is a fast intramolecular
reaction
Hydrogen abstraction
+ Ro' -- + ROH * + AAA E l
pi + A E Depolymerization . Recombination
Or-OR
MA m
Figure 1. Possible reaction mechanisms for the grafting of
maleic anhydride onto polypropylene in the melt state in the
presence of organic peroxide. MA represents maleic anhydride.
and seems predominant in the melt state in the presence of
organic per~xide.~'-~l Nevertheless, re- action B, maleic anhydride
grafting on these pri- mary radicals, thus before @-scission, is
sometimes sugge~ted.~*'' However, it seems that grafting in so-
lution, or in solid state could favour this mecha- n i ~ m . ~ * "
- ' ~ Reaction C seems unlikely as a conse- quence of the stability
of the anhydride radical but sequence B + C is, in fact,
undistinguishable from sequence A + H (see the following). Reaction
D which could lead to the grafting of maleic anhydride by
ene-reaction was already studied.16 Very severe conditions are
needed to favour this reaction: very low polypropylene molecular
mass (high concentra- tion of double bond chain ends) very high
concen- tration of maleic anhydride, high temperature and pressure,
and long reaction time. Even maleic an- hydride grafting onto
polyisoprene by ene-reaction needs severe conditions and leads to
moderate yields2' Depolymerization (reaction E ) seems very
-
MALEIC ANHYDRIDE-FUNCTIONALIZED PP 83 1
unlikely at 190C. Indeed, depolymerization of polypropylene
becomes only significant at temper- atures above 300C.52 Moreover,
no gaseous products were observed during processing of
polypropylene at 190C even for long mixing time. Transfer and
recombination (reactions F and G ) cannot auto- matically be
discarded and need further discussions. Maleic anhydride end chain
grafting (reaction H ) is generally ~ u g g e s t e d . ~ * ~ , ~ ,
~ Recombinations ( reac- tions I and J ) and mainly grafting of
poly (maleic anhydride) (reaction K ) are subjects of contro-
This work intends to contribute to the elucidation of the most
probable mechanism for maleic anhy- dride grafting on polypropylene
in our experimental conditions within all the possibilities
described in Figure 1. This requires accurate molecular mass de-
termination and determination of nature, location, and
concentration of grafted anhydride.
versy.6.9,18,39-45
EXPERIMENTAL
Materials
Commercial maleic anhydride grafted polypropyl- ene, pure
polypropylene, maleic anhydride, and or- ganic peroxide (
1,3-di-tert-butylperoxyisopropyl benzene, Perkadox 14) were kindly
supplied by Sol- vay & SA. n -0ctadecylsuccinic anhydride came
from Alfa Product. It was purified by recrystallization from
acetone solution. Its final purity (97% ) was verified by
titration. Toluene (99% purity) and 1,2,4- trichlorobenzene
(vacuum-distilled) were purchased from Merck-Belgolabo. Benzyl
alcohol (99% purity) was provided from Janssen Chimica. Methanol
was distilled from a technical grade. Poly ( maleic an- hydride)
(Belclene 200) was supplied by Ciba Geigy and was purified by
reprecipitation. This product was analysed by NMR, SEC and MS.
Belclene 200 was identified as mainly constituted of maleic an-
hydride oligomers containing less than 10 units.53
PP-g-MA Synthesis
Brabender Plasticorder
The Brabender Plasticorder was equipped with an electrically
heated W50EH mixing device of 50 mL volume. Oxidation of mixed
polymers was largely reduced by a nitrogen flow of 5 L/min above
the mixing chamber. No mechanical pressure was ap- plied on the
polymer during mixing. The charge processed in the Brabender
Plasticorder was 40 g. All the reactants (maleic anhydride, organic
per- oxide, and polypropylene) were dry mixed together before their
fast (< 1 min) introduction in the pre-
heated Brabender Plasticorder. The mixing speed was fixed at 75
rpm, the reaction time duration at 4 min, and the temperature at
190C.
Powder Mixing
Some grafting experiments were performed on pow- ders obtained
by dry mixing polypropylene, organic peroxide, and maleic
anhydride. Then the heating was performed in glass sealed tubes in
an electrically heated furnace. Temperature was regulated within
1C. Thermal treatments of PP-g-MA were also performed in such glass
sealed tubes, the conditions (temperature, time) of which will be
mentioned in the following.
Analytical Characterization
Titration of Anhydride Content
The anhydride concentrations present in raw, heated, and washed
PP-g-MA samples were deter- mined by titration of the acid groups
derived from the anhydride functions. After dissolution of 1 g of
PP-g-MA in 100 mL of toluene at boiling temper- ature, 200 pL of
water were added to hydrolyze an- hydride functions into carboxylic
acid functions. The boiling temperature was maintained for 1 h.
Car- boxylic acid concentration was determined directly by alkali
titration using 0.025N potassium hydroxide in methanol/benzyl
alcohol 1/9 (v/v) . The indi- cator used was a solution of 1%
Phenolphthalein in methanol. The PP-g-MA was completely soluble at
the boiling temperature and did not precipitate dur- ing titration.
A blank was carried out by the same method.
FTIR Spectroscopy
FTIR spectra were recorded on a Perkin-Elmer FTIR Spectrometer
1760-X from 4000 to 400 cm- with a 0.5 cm- resolution. For some
samples, ab- sorption bands between 1880 and 1690 cm- were
mathematically analyzed by an iterative curve-fit- ting software
called IGOR provided from WaveMetrics and working on an Apple
Computer. Each absorption band was approximated by a Lo- rentzian
function.54 The iterative software defines location, amplitude and
half-band width of each band to restore the original spectrum by
the addition of all Lorentzian functions. Each Lorentzian func-
tion is considered to be the actual absorption band. Maximum values
of these Lorentzian curves are considered as absorbance values for
quantitative analyses.
Films of 50-100 pm thickness were obtained by
compression-molding 0.1-0.2 g sample between
-
832 DE ROOVER E T AL.
2000
1600 - - E 9 1200- 2 g 800- i-
400 -
0
PTFE recovered aluminium sheets under 1 MPa pressure at 190C for
30 s. These films were dried at 12OOC for 20 h to evaporate the
unreacted maleic anhydride as well as to perform complete
cyclization of any diacid into the cyclic anhydride form. For
accurate characterization of PP-g-MA it is of great importance to
insure total elimination of all the un- reacted maleic anhydride
and complete cyclisation of carboxylic acid into carboxylic
anhydride. Un- reacted maleic anhydride gives rise to absorption
bands in the same region than grafted anhydride ( 1785 cm-' ) which
prevents accurate determination of the grafted anhydride
concentration. Elimination of unreacted maleic anhydride can be
verified by the disappearance of a characteristic absorption band
at 720 cm-' associated to the carbon/carbon double bond of maleic
anhydride. Complete cyclization of carboxylic acid into a
carboxylic anhydride form is of great interest to simplify the FTIR
spectrum. This cyclization is assumed by the disappearance of the
absorption band at 1715 cm-' , assigned to the car- boxylic acid.
To check that anhydride are really grafted on the polypropylene
chains, some samples were purified by dissolution in boiling
toluene fol- lowed by precipitation in acetone.
PP-g-MA model compounds were rz -0ctadecyl- succinic anhydride
and poly (maleic anhydride). FTIR analysis of these model compounds
was per- formed after incorporating various concentrations in
melted polypropylene using the Brabender Plas- ticorder. This
strategy was adopted to take into ac- count the polypropylene
effect onto carboxylic an- hydride absorption bands. This procedure
enables the attribution of absorption bands and the deter- mination
of molar absorption coefficients.
All analyses are performed for
this mixing time
1 *.. a s . . . . . . . . . .
' I ' I ' I . 0 '
Size Exclusion Chromatography (SEC)
Molecular weight measurements of PP-g-MA were carried out by
high-temperature size exclusion chromatography. The chromatograph
was a Waters 150C equipped with 2 Shodex columns AT-80M/S an 1
Shodex column Styragel 300 A. A differential refractometer detector
coupled with a "microVAX Data Station" provided by Digital was used
for re- cording and analyzing the signal. Permeation solvent was
vacuum-distilled 1.2.4-trichlorobenzene stabi- lized with 2%
Irganox 1010. Samples dissolution was achieved at 160C during 1 h.
Concentrations were in the range of 6 to 8 mg/mL. Injection volume
was 120 pL. The system was maintained at 135C during the analyses.
Precautions similar to those described for FTIR were taken before
molecular mass analyses by high-temperature SEC, viz., elimination
of un- reacted maleic anhydride and cyclization of maleic
carboxylic acid. A polypropylene calibration was
used for SEC analyses. In the present case, the amounts of
grafted anhydride remained low and it was assumed that the
hydrodynamic volume in so- lution was not largely modified.
RESULTS AND DISCUSSION
Mixing Torque Determination
Figure 2 shows an example of the mixing torque ob- tained during
melt reaction of polypropylene with maleic anhydride in the
presence of organic peroxide. After a strong decrease of the torque
due to the plas- ticization of the polypropylene the torque becomes
nearly constant. In this work all the samples were characterized
after 4 min of mixing time.
Figure 3 presents the torque measured after 4 min of mixing as a
function of the concentration of mal- eic anhydride and organic
peroxide introduced in the samples. It appears that, for the
selected exper- imental conditions, the torque depends on organic
peroxide content but not on maleic anhydride con- centration.
Molecular Weight Determination
Figure 4 presents number-average molecular weights measured
after 4 min of mixing as a function of the concentration of maleic
anhydride and organic per- oxide for all the samples. Figure 4
shows that the molecular weight actually depends on organic per-
oxide content but does not depend on the maleic anhydride
concentration. Dispersion of the results corresponds to the
reproducibility of the SEC anal- ysis. The decrease of molecular
masses is attributed to the well known p-s~iss ion .~~-~ '
Figure 5 details the mechanism of this p-scission.
0 1 2 3 4 5
Mixing time (min)
Figure 2. Example of torque vs. mixing time during the grafting
of maleic anhydride onto polypropylene in the presence of organic
peroxide. Conditions: 190C, 75 rpm, maleic anhydride 1% by weight,
organic peroxide 1.2% by weight, 5 L/min Nz flow.
-
MALEIC ANHYDRIDE-FUNCTIONALIZED PP 833
1200
900 E cil v 2 600 + 4.80%MA g I- 300
0
0.0 0.5 1 .o 1.5 Peroxide weight percent
Figure 3. Torque after 4 min of mixing as a function of maleic
anhydride and organic peroxide concentrations introduced in
polypropylene. Conditions: 190C, 75 rpm, 5 L/min N2 flow. % MA: wt%
of maleic anhydride used during the functionalization.
As it can be observed on Figure 5, p-scission gives rise to one
radical chain end and one carbon-carbon double bond chain end. The
molecular mass deter- mination allows to calculate the efficiency
of the or- ganic peroxide to promote p-scission in polypropyl- ene.
Table I gives a comparison between the mea- sured molecular masses
of the samples and the calculated values considering that each
peroxide radical induces one 0-scission. Calculation is per- formed
using eq. ( 1 ) derived from demonstration in the following.
The difference between the number of polypro- pylene chains
after and before peroxide treatment is equal to the number of
polypropylene chains cre- ated by this treatment; thus:
1 / M, - 1 / M: = amount of polypropylene chains created by
peroxide treatment (in mol/g)
80000
- - m m 60000
0 0.25OhMA A 1 .m MA -
.- + 4.80 V ~ M A i a & 40000 I=
20000 0.0 0.5 1 .o 1.5
Peroxide weight percent
Figure 4. M,, as a function of maleic anhydride and organic
peroxide concentrations introduced in polypro- pylene. Conditions:
190C, 75 rpm, 5 L/min N2 flow. % MA: wt % of maleic anhydride used
during the function- alization.
Figure 5. Detailed mechanism of polypropylene @-scission.
If each radical issued from the peroxide splitting induces one
@-scission, then:
l / M n - 1/M: = [Rad]
with [ Rad] = the radical concentration (in mol/g) . Using [ PP
] = 1 /ME the initial polypropylene
concentration viz. the initial number of mol of poly- propylene/
g, this relation can be transformed in eq. (1):
where [ PP] = polypropylene concentration (in mol/ g), [ Rad] =
radical concentration (in mol/g), and ME = 60000 = initial
number-average molecular weight (in g/mol).
In eq. ( I ) , [ PP] is the actual polypropylene con- centration
(in mol/g) taking into account the num- ber of chains of
polypropylene and not the number of structural units. For low
concentrations of organic peroxide (0.01-0.05% ) the ratio between
the cal- culated and experimental M,, lies close to unity ( 1.2 and
1.1 ) , while for medium to high concentrations of organic peroxide
(0.25-1.25% ) , it decreases from 0.6 to 0.2. This behavior can be
explained by the stoechiometric attack of the peroxide leading to
p- scission for low peroxide (0.01-0.05% ) concentra- tions. For
higher peroxide concentrations ( 0.25- 1.25% ) , the difference
between calculated M,, and experimental M, may be due to radical
recombi- nation: more and more recombinations occur when organic
peroxide concentration increases.
Those results are very important because they can lead to
conclude that grafting mechanism is not occurring through path B of
the Figure 1. Indeed, grafting of maleic anhydride (path B) onto
the polypropylene chain competes with p-scission (path A ) and thus
increasing concentration of maleic an- hydride would progressively
reduce the molecular weight decrease through p-scission and that is
not the case.
Path F consisting of a transfer reaction seems not to occur to a
large extent: the experimental mo- lecular weight decrease is never
significantly larger than the calculated thus increasing
concentration of maleic anhydride would progressively reduce the
molecular weight decrease through p-scission and that is not the
case. Path F consisting of a transfer
-
834 DE ROOVER ET AL.
Table I. Comparison between Experimental M,, and Calculated M,,
of PP-g-MA.
Peroxide Experimental M,," Calculated M,,/ (wt %) (SEC)
Calculated M,b Experimental M,,
0.00 60,000 60,000 1 0.01 46,000 55,000 1.2 0.05 42,000 44,000
1.1 0.25 36,000 21,000 0.6 1.25 27,000 6000 0.2
a Average value for all maleic anhydride concentrations.
Obtained considering that one radical induces one 0-scission.
reaction seems not to occur to a large extent: the experimental
molecular weight decrease is never significantly larger than the
calculated values based on the steochiometric attack of the chain
by the peroxide leading to @-scission.
Molecular Weight-Mixing Torque Relationship
In Figure 6 a linear relationship can be observed between the
logarithm of the mixing torque and the logarithm of M,. Taking into
account that Braben- der Plasticorder mixing torque is a function
of melt viscosity, the slope of 3.2 does not disagree with
rheological theory55 nor with similar results on other
polymer^.^^-^^
Titration of Grafted Anhydride
Grafted Anhydride Concentration
Titrations were performed onto washed and dried PP-g-MA to
determine the amount of grafted an- hydride. Results are summarized
in Figure 7 which shows that grafted anhydride concentrations in-
crease with the concentration of maleic anhydride as well as with
the concentration of organic peroxide used for the reaction.
8.6 - y = -27.9 + 3.2X
F k 0.87769
E .-
5.0
-1
3.2 10.1 10.5 10.9 11.3
Logarithm of Mn
Figure 6. of M,.
Logarithm of the mixing torque vs. logarithm
Grafted Anhydride Concentration and Chain Ends Concentration of
the PP-g-MA
Following the reaction mechanism discussed before, it is
expected that polypropylene undergoes a @- scission after the
attack by organic peroxide and that grafting takes place a t the
new radical chain end (paths A and H of Figure 1 ) . In this case
it is of great interest to compare the grafted anhydride
concentration onto polypropylene and the polypro- pylene radical
chain ends concentration created by P-scission. Equation ( 2 )
allows the calculation of the concentration of radical chain ends
which were generated by @-scission during grafting. This equa- tion
is easily deduced thanks to the demonstration of eq. (1)
where N = concentration of radical chain ends cre- ated during
grafting (in peq/g), M; = number av- erage molecular weight before
grafting (in mol/g) , M, = number average molecular weight after
graft- ing (in mol/g), and 106 = scaling factor converting mol / g
into peq / g.
300
200
100
0 0.0 0.5 1 .o 1.5
Peroxide weight percent
Figure 7. Grafted anhydride concentrations determined by
titration vs. maleic anhydride and organic peroxide concentrations
introduced in polypropylene. Conditions: 190C, 75 rpm, 5 L/min N2
flow. % MA: wt % of maleic anhydride used during the
functionalization.
-
MALEIC ANHYDRIDE-FUNCTIONALIZED PP 835
Equation 2 takes into account the P-scission mechanism described
in Figure 5 in which one rad- ical chain end is always accompanied
by a carbon- carbon double bond chain end. Consequently, con-
centration of radical chain ends is just the half of the
concentration of chain ends concentration cre- ated during
grafting.
In Figure 8 concentrations of grafted anhydride are plotted as a
function of reciprocal M, for each sample. The straight line
represents the concentra- tion of radical chain ends generated
during grafting and calculated using eq. ( 2 ) .
Figure 8 shows that, for most samples, the con- centration of
grafted anhydride is higher than the concentration of new radical
chain ends created by the &scission mechanism. Referring to
reaction paths summarized in Figure 1, several mechanisms can
explain this behavior. The first one was already mentioned and is
described by path B in Figure 1. It was previously discarded
because no decrease of molecular weight during grafting of maleic
anhydride would be observed if the reaction stops after grafting.
No real variations of molecular weight versus grafted anhydride
concentration is observed in our experi- mental conditions.
Scission after grafting described in path C of Figure 1 could be
invoked. Effectively it is important to noticed that paths B and C
(graft- ing followed by scission ) give the same products and
cannot be distinguished from paths A and H (0- scission followed by
grafting). Nevertheless, the former sequence seems less probable
because this mechanism requires a hydrogen abstraction by the
anhydride radical. This radical should be stabilized by resonance
effect5' as described in Figure 9.
Another explanation consisting in radical recom- bination is
illustrated in reaction I of Figure 1. This
150 I I
100
50
0
0 0.050/oMA 0 0.25VoMA A l.ooo/oMA ' + 4.80 %MA
' A
1.5 3.0 4.5
llM, (rnol/g)
Figure 8. Grafted anhydride concentration determined by
titration and radical chain ends concentration calcu- lated from
eq. ( 2 ) vs. reciprocal M,. Conditions: 190"C, 75 rpm, organic
peroxide 0.01 to 1.2%, 5 L/min N2 flow. % MA: wt % of maleic
anhydride used during the func- tionalization.
r
Figure 9. Radical stabilization by resonance effect.
one cannot occur to a large extent in the experi- mental
conditions. Indeed, if this radical recombi- nation took place, the
molecular masses would be higher, the higher the amount of grafted
maleic an- hydride. As already pointed out, this behavior was not
observed. Finally, as suggested by Gaylord the formation of poly
(maleic anhydride) during func- tionalization of polypropylene
(path B + H + K ) should be envisaged. This polymerization of
maleic anhydride was actually observed. A future study will detail
the homopolymerization of maleic anhydride in our experimental
conditions53 and the presence of free poly (maleic anhydride) in
highly grafted PP- g-MA. Indeed, maleic anhydride polymerization
can explain that, for many samples of PP-g-MA, con- centration of
grafted anhydride is higher than the concentration of radical chain
ends.
Infrared Analysis of PP-g-MA
FTIR Spectrum of PP-g-MA
Figure 10 shows a FTIR spectrum of a polypropylene processed in
the presence of organic peroxide and
2000 1900 1800 1700 1600 ls00
Wave number (m-11
Figure 10. Infrared spectra of PP and PP-g-MA be- tween 2000 and
1500 cm-'. (A) Polypropylene processed with organic peroxide and
containing all processing ad- ditives except maleic anhydride. ( B
) PP-g-MA. Spectrum obtained after drying the samples at 120C
during 20 h under vacuum.
-
836 DE ROOVER ET AL.
processing additives but without maleic anhydride (blank) and a
FTIR spectrum of a PP-g-MA. Drying at 12OOC for 20 h under vacuum
was applied to elim- inate free maleic anhydride which could be
confused with grafted anhydride and to transform carboxylic acid
into carboxylic anhydride.
As it can be seen in Figure 10, the FTIR spectrum of PP-$-MA
shows two intense overlapping absorp- tion bands at 1784-1792 cm-'
and weak absorption bands around 1850 cm-'. No absorption bands are
observed at 1784-1792 cm-' for the polypropylene sample degraded in
the presence of the same per- oxide under the same experimental
conditions. Ab- sorption bands at 1784-1792 cm-' and around 1850
cm-' can be assigned to grafted anhydride because five members
cyclic anhydrides exhibit an intense absorption band near 1780 cm-'
and a weak ab- sorption band near 1850 cm-' due to symmetric and
asymmetric C = 0 stretching respectively.60,61 As it
1840 1800 1760 1720
Wave numbex (cm-1)
1840 1800 1760 1720
Wavenumba(cm-I)
ID
1840 1800 1760 1720
Wave numbs (cm-1) 1040 1800 1760 1720
Wave numbs (cm-1)
1840 1800 1760 1720 Wave numbs (cm-1)
Figure 11. Curve fittings of infrared spectra of four industrial
samples of PP-g-MA and of a polypropylene sample containing all
processing additives except maleic anhydride (blank). (A) PP-g-MA
001, ( B ) PP-g-MA 002, (C) PP-g-MA 003, ( D ) PP-g-MA 004. See
Table I11 for characteristics. (E) Polypropylene processed with
organic peroxide and all processing additives except maleic an-
hydride.
2000 1900 1800 1700 16M) 1500
Wave numbex (cm-1)
Figure 12. Infrared spectra of n-octadecylsuccinic an- hydride
dispersed in polypropylene. ( A ) Polypropylene containing
processing additives and n-octadecylsuccinic anhydride. ( B )
Polypropylene containing only processing additives.
can be seen on Figure 10, the absorption bands at 1784-1792 cm-'
seem to result from the overlapping of two different absorption
bands, which should therefore be assigned to two different
anhydride species. To identify the two bands more clearly,
curve-fitting of this spectrum was undertaken into a series of
Lorentzian curves54 defined by their po- sition, half bandwidth,
and intensity. A curve fitting procedure restores the actual
spectrum from these Lorentzian curves. Only these two intense
absorp- tion bands at 1784-1792 cm-' (symmetric stretch- ing) will
be taken into account in the present article for identification and
calibration.
Figure 11 shows curve fittings of four industrial samples ( A-D
) and a fitting realized on the poly- propylene sample containing
organic peroxide and processing additives but no maleic anhydride
(E) . As it can be seen in Figure 11, six absorption bands (or part
of absorption bands) are found by curve fitting. Several of those
absorption bands are due to polypropylene or processing additives
[Fig. 11 (E) 3 . Absorption bands I, 11, and VI are due to polypro-
pylene while absorption band V is due to a processing additive.
Only absorption band I11 at 1792 cm-' with a half bandwidth of 10
cm-' and absorption band IV at 1784 cm-' with a half bandwidth of
25 cm-' correspond to anhydride functions. For the four samples,
position and half bandwidth of those ab- sorption bands are
reproducible although concen- trations of grafted anhydride range
within one order of magnitude.
Absorption Band Assignment
With a view to identify the two new absorption bands detected in
the PP-g-MA spectrum, some an-
-
MALEIC ANHYDRIDE-FUNCTIONALIZED PP 837
Table II. Position and Half Bandwidth of the Anhydride C=O
Stretching of Model Compounds Blended in Molten Polypropylene
Position Half Bandwidth Model Compound Formula (cm-' )
(cm-')
Citraconic anhydride 1780 10
10
10
Poly(ma1eic anhydride)
Maleic anhydride 1780 O f y 0
10
hydride model compounds were selected. For ex- ample, Figure 12
shows the FTIR spectra of 0.5 wt % of n -octadecylsuccinic
anhydride dispersed in polypropylene and of the blank sample. Table
I1 gives position and half bandwidth of anhydride ab- sorption
bands of these model compounds blended in molten polypropylene at
low concentrations (0.5 wt % ) . As can be seen in Table 11,
different positions are observed for the same symmetric stretching
band of different anhydrides. Moreover, the half band- width of
poly (maleic anhydride) is clearly larger than the others. In
PP-g-MA, a band is observed at 1792 cm-' with a half bandwidth of
10 cm-' and a second band is observed at 1784 cm-' with a half
bandwidth of 25 cm-'. It is thus proposed to asso- ciate this later
one to poly ( maleic anhydride ) and the former one to succinic
anhydride end-groups. Indeed, upon reaction a single maleic
anhydride group undergoes a transformation into saturated anhydride
(succinic type) as described in Figure 1.
Influence of Model Compound Concentrations and of
Temperature
The influence of concentration and temperature on FTIR spectra
of polypropylene/model compound blends was also studied. At high
concentrations (e.g., 2 wt % of n-octadecylsuccinic anhydride) two
over- lapping absorption bands situated at about 1780 and 1792 cm-'
are observed when FTIR analysis is per- formed at room temperature.
When the temperature
is increased the intensity of the absorption band at 1780 cm-'
decreased while the intensity of absorp- tion band at 1792 cm-'
increased. When the tem- perature reached 60C the absorption band
at 1780 cm-' totally disappeared and the intensity of ab- sorption
band at 1792 cm-' reached its maximum. This behavior was attributed
to physical interactions between anhydride dipoles. Those are only
possible if the concentration of anhydride is high and if ther- mal
agitation does not suppress the interactions.
The same experiment was undertaken with poly (maleic anhydride).
For this model compound such a behavior was never observed
absorption band at 1784 cm-' is not modified with concentration or
temperature. Indeed, in poly (maleic anhydride ) the anhydride
groups are chemically linked together and physical interactions
between anhydride dipoles should be always possible even at high
temperature or at low concentration.
FTIR analyses of PP-g-MA were performed at high temperature. The
infrared spectra of PP-g-MA at 25C before heating ( A ) , at 180OC
(B) , and at 25C after cooling ( C ) are reported in Figure 13. In
Figure 13, it is shown that even at about 180C the two overlapping
absorption bands of the cyclic car- bony1 anhydride at 1784 and
1792 cm-' are not modified (spectrum B ) . At this temperature, the
FTIR polypropylene spectrum becomes similar to atactic
polypropylene due to the melting. Conse- quently, it can be assumed
that the two overlapping absorption bands at 1784 and 1792 cm-'
observed
-
838 DE ROOVER ET AL.
Figure 13. Infrared spectra of PP-g-MA at 25C before heating:
(A) , at 180C (B) , and at 25C after cooling (C).
in PP-g-MA correspond effectively to two anhydride species. When
the temperature decreases to 25C the usual aspect of
semi-crystalline PP-g-MA FTIR spectrum appears again ( spectrum C )
.
In the Figure 1 those results support paths A + H + J (with R =
hydrogen) leading to graft suc- cinic anhydride and paths A + H + K
leading to graft poly ( maleic anhydride). Further experiments will
be undertaken in the following section with a view to corroborate
the above assignments concern- ing PP-g-MA chemical structure.
Depolymerization of Crafted
Poly (maleic anhydride)
Poly(ma1eic anhydride) is reported to exhibit a ceiling
temperature of about 150"C.62,63 This tem- perature seems quite low
after examination of the results presented in this article and with
those deal- ing with the oligomerization of maleic anhydride at
190C reported elsewhere.53 This can arise from the fact that the
thermodynamic parameters used for the ceiling temperature
calculation were not adapted to the polypropylene melt
functionalization condi- tions. Indeed the AH, A S , and the
ceiling temper- ature values were obtained for maleic anhydride
homopolymerization in benzene solution.63 Extrap- olation for the
maleic anhydride melt homopoly- merization and for the
polypropylene melt grafting conditions at high temperature is
therefore certainly not obvious. Consequently, the ceiling
temperature argument which was mentioned to preclude any
homopolymerization of maleic anhydride may be inaccurate.
Prolonged heating of PP-g-MA at a temperature of 165C does not
lead to poly( maleic anhydride) depolymerization and no decrease of
the 1784 cm-' absorbance is observed. However, in another exper-
iment, PP-g-MA 001 was heated at 300C for 20 h in a glass tube,
continuously evacuated to a pressure
1900 1850 1800 1750 1700 1650
Wave number ( c m ~ l t
Figure 14. Infrared spectrum of PP-g-MA 001 ( A ) (see Table I11
for characteristics) and PP-g-MA 001 heated at 300C under high
vacuum during 20 h (B) .
of about 1 mm Hg. Figure 14 shows anhydride region of FTIR
spectra for the original PP-g-MA ( A ) and for the thermally
treated PP-g-MA ( B ) . In Figure 14, the intensity of the 1784
cm-' infrared band [as- signed to poly (maleic anhydride)]
decreases with regard to the intensity of the 1792 cm-' infrared
band. Figure 15 shows curve-fitting performed on these spectra.
Figure 15 shows that the ratio between poly ( maleic anhydride )
absorbance and succinic anhydride absorbance decreases from about 5
before heating to 2 after heating. The point quantitative analysis
of this article enables the concentra- tion calculation of single
succinic anhydride and poly ( maleic anhydride) from curve-fitted
spectra. Calculations performed following this calibration gave the
absolute values 34.9 and 10.3 peq/g re- spectively for poly (
maleic anhydride) concentra- tions before and after heating while
succinic anhy- dride concentration did not change significantly
(6.9 to 6.2 peq/g). This result conclusively supports the
assignment of the 1784 cm-' infrared band to poly (maleic
anhydride).
I* n
1840 1800 1760 1720 1840 1800 1760 1 7 2 0
Wavcnurnba (cm-1) Wave number (cm-1)
Figure 15. Curve fitting of infrared spectra of PP-g-MA 001 (see
Table I11 for characteristics) (A) and PP-g-MA 001 heated at 300C
under high vacuum during 20 h (B) .
-
MALEIC ANHYDRIDE-FUNCTIONALIZED PP 839
+
" 6 0
I
0
0
mechanism was performed during melt processing (Brabender
Plasticorder ) by adding maleic anhy- dride a sufficiently long
time after the organic per- oxide to prevent any radical grafting:
pure polypro- pylene was mixed in the Brabender Plasticorder (
19O"C, 75 rpm) and 1 wt % of organic peroxide was added to promote
@-scission and thus also unsatu- rated end groups formation as
shown in Figure 5. Three minutes after the addition of the organic
per- oxide, the mixing torque was constant and corre- sponded to a
M , of about 25,000 g/mol. One minute after the stabilization of
the mixing torque, 5 wt % of maleic anhydride were added and
allowed to react during 4 min. Infrared analysis were performed
after this reaction time and very low amounts of grafted anhydride
were detected. Those concentrations ( 1 wt percent of organic
peroxide and 5 wt percent of maleic anhydride) are in fact the most
favorable of all our experiments for the ene-reaction and even in
this case ene-reaction mechanism remains neg- ligible.
Consequently, the experimental conditions used for the melt
maleic anhydride grafting of polypro- pylene with organic peroxide
are, by far, too weak to promote the ene-reaction with high yield
and path D in Figure 1 does not occur significantly in our
experimental conditions.
Figure 16. Mechanism of the ene-reaction. Following this
mechanism, more than one anhydride per polypro- pylene chain could
be grafted.
ene-Reaction
Path D of the Figure 1 is called the ene-reaction. This path is
detailed in Figure 16 which shows that more than one maleic
anhydride for one polypro- pylene chain can be grafted by this
mechanism. Fol- lowing the literature, 16,29 this reaction is
possible but needs drastic conditions: very low molecular mass to
promote a high concentration of unsaturated chain ends (e.g., M , =
1000) , a high concentration of maleic anhydride (e.g., 20 wt % ) ,
a very high temperature, and a long reaction time (e.g., 225C and 4
h using a mechanical stirrer a t 300 rpm) .
At first, an experiment was attempted in order to check the
possibility of ene-reaction: a polypro- pylene with unsaturated end
groups was first syn- thesised by peroxide treatment only ( 1 wt %
) . After reprecipitation, unsaturated end-groups were ob- served
by their IR absorption band at 820 cm-'. This sample was heated at
190C in sealed glass tube in the presence of 25 wt % of maleic
anhydride for 18 h. However, even in those severe conditions only
limited amounts of grafted anhydride were detected.
Secondly, an attempt of grafting by ene-reaction
Quantitative Analysis of PP-g-MA
FTIR spectroscopy was used for quantitative anal- ysis of
PP-g-MA. Curve-fitting enables to separate the two overlapping
absorption bands at 1792 cm-' (assigned to succinic anhydride) and
at 1784 cm-' [assigned to poly (maleic anhydride ) ] . A
calibration was realized by using representative model com- pounds,
melt mixed in polypropylene: n -0ctadecyl- succinic anhydride and
poly (maleic anhydride) at three concentrations. This calibration
considers an- hydride as well as carboxylic acid forms. Results of
those calibrations are given by eqs. ( 3 ) and ( 4 ) .
For n -octadecylsuccinic anhydride:
[Anhydride] = 21.5 (Abs 1792 cm-'/
Abs 1100 cm-') + 24.5 (Abs 1715 cm-'/ Abs 1100 cm-l) ( 3 )
For poly ( maleic anhydride ) :
[Anhydride] = 51.3 (Abs 1784 cm-'/
Abs 1100 cm-') + 52.5 (Abs 1715 cm-'/ Abs 1100 cm-') (4)
where [anhydride] represent the anhydride concen- tration (in
peq/g) , Abs 1792 cm-' is the absorbance
-
840 DE ROOVER ET AL.
Table 111. Infrared and Titrations
PP-g-MA Characterization by SEC and Quantification of Grafted
Anhydride by
Radical Chain Poly(ma1eic Total Total Ends Succinic Type
anhydride) Anhydride Anhydride
MI Concentration Concentration Concentration Concentration"
Concentrationb PP-g-MA (g/mol) (wq/g) (wq/g) (CLeq/g) (jleq/g)
(Peq/g)
00 1 43,000 12.8 6.9 34.9 41.8 002 50,050 10.2 3.2 15.4 18.6 003
54,900 6.7 1.5 4.5 6.0 004 66,700 4.2 1.5 3.7 5.2
44 15 6 7
a Calculated by infrared spectroscopy. Calculated by
titration.
of the succinic anhydride symmetric C = 0 stretch, Abs 1784 cm-'
is the absorbance of the poly (maleic anhydride) symmetric C = 0
stretch, Abs 1715 cm-' is the absorbance of the carboxylic acid
symmetric C = 0 stretch, and Abs 1100 cm-' is the polypro- pylene
reference absorbance. Absorbance is the maximum values of the
Lorentzian curves obtained by curve fitting.
The anhydride concentrations resulting from curve fitting and
calibration are summarized in Ta- ble 111. Analyses were undertaken
onto dried PP-g- MA films (24 h at 100C under vacuum) to decrease
the carboxylic acid and to totally volatilize unreacted maleic
anhydride. Concentrations based upon chemical titration are also
reported.
The results reported in Table I11 show that the succinic
anhydride concentration is always signifi- cantly lower than the
poly (maleic anhydride) con- centration. Titration results are
coherent with the succinic and poly ( maleic anhydride )
concentrations determination by FTIR. Succinic and poly ( maleic
anhydride) concentrations enables to calculate the average length
of grafted poly (maleic anhydride) moieties. Indeed, radical chain
end concentration generated during melt reaction can be estimated
us- ing eq. ( 2 ) which compares initial (before grafting) and
final (after grafting) number-average molecular weight of
polypropylene. Grafting of anhydride only at those polypropylene
radical chain ends is dem- onstrated herebefore. Thus, both
succinic anhydride and poly ( maleic anhydride) are grafted on
polypro- pylene radical chain ends.
For each PP-g-MA the number of grafted succinic anhydrides can
be subtracted from the total number of radical chain ends.
Consequently, the average length of poly (maleic anhydride)
moieties can be calculated by dividing the poly (maleic anhydride)
concentration by the remaining amount of radical chain ends. For
example, the approximate length of poly (maleic anhydride) is 6 in
the case of PP-g-MA 001. Lower lengths are calculated for PP-g-MA
002,
003, and 004. This result suggests that all the radical chain
ends were not actually functionalized by any kind of anhydride.
Some of them should be elimi- nated by recombination (with low
molecular weight radicals), transfer or dismutation.
CONCLUSIONS
This work consists in the characterization of PP-g- MA realized
in the melt state in the presence of organic peroxide. Reaction
mechanism as well as position, concentration, and nature of grafted
an- hydride present in PP-g-MA were studied. Melt grafting of
maleic anhydride onto polypropylene in the presence of organic
peroxide is always associated with molecular weight decrease. This
molecular weight decrease depends mainly on organic peroxide
concentration and seems not affected by maleic an- hydride content.
Grafting yield depends on maleic anhydride concentration as well as
on organic per- oxide concentration. Reaction mechanisms able to
justify those experimental observations consist in grafting of
maleic anhydride onto radical chain ends arising from the
@-scission of polypropylene. However, concentrations of grafted
anhydride determined both by chemical titration and FTIR
spectroscopy exceed radical chain ends concentra- tions generated
in polypropylene. Consequently, poly (maleic anhydride) grafting
was taken into ac- count to justify this observation. This
assumption was confirmed by FTIR spectroscopy using model
compounds.
Figure 17 summarizes the proposed overall mechanism for maleic
anhydride grafting onto poly- propylene in the melt state using
organic peroxide. As a matter of fact, it consists only in the
reaction routes A + H + ( K or J ) of Figure 1. The grafting starts
with the homolytical scission of organic per- oxide. The radical
abstracts a tertiary hydrogen from the chain of polypropylene
forming a macroradical.
-
MALEIC ANHYDRIDE-FUNCTIONALIZED PP 841
Initiatipn reactions
Grafting reactions
ROOR - 2 RO' + RO'
H abstraction I I
+ ROH
0-scission
u. + dJ, End chain grafting and homopolymerization of MA
Termination reactions n = 0,1,2 ...
or qo +
or * Figure 17. Proposed PP-g-MA grafting mechanism.
This macroradical undergoes quickly a @-scission with the
simultaneous formation of a radical chain end and a vinylidene
chain end. Maleic anhydride grafting takes place on the radical
chain end and does not take place before @-scission. Ene-reaction
between vinylidene chain end of polypropylene and maleic anhydride
does not occur in the present ex- perimental conditions. After
grafting of one maleic anhydride on the radical ended polypropylene
chain, termination reaction or oligomerization of maleic anhydride
can take place. Termination of radical reactions call out
recombination with low molecular weight radicals or dismutation
reactions. Possible recombination can involve for example, peroxide
residues or process additives. Dismutation would lead to saturated
(succinic) and unsaturated ( mal- eic) anhydride.
Determination of termination of radical reactions is always
difficult and specially in this case owing to the very low
concentrations of species which are, moreover linked to
polypropylene chains. It can be noticed that the homopolymerization
of maleic an- hydride decreases the amount of termination reac-
tion per grafted anhydride. Grafting of poly ( maleic anhydride)
onto polypropylene as proposed by Gay- lord et al. is now
experimentally confirmed, both by direct IR determination and by
calculation based on chain ruptures. Homopolymerization of maleic
an- hydride without grafting is also possible. Grafting of
anhydride units along the chain seems to remain negligible in the
conditions of reactive processing. However, this grafting, without
decrease of molec- ular weight can be obtained with chemical
deriva- tives of MA, for instance imides used as grafted sta-
bilizing agents.64
The main difference between both case resides probably in the
insolubility of maleic anhydride in molten p~lypropylene.~~ This
insolubility could also explain the results reported by Lambla et
al.65 on the influence of the nature of the peroxide (polar- i ty)
on the grafted succinic anhydride/poly (maleic anhydride ) ratio.
Following their polarity, perox- ides would dissolve preferentially
into the maleic anhydride phase or into the polypropylene phase. A
direct determination of solubility diagrams in molten polypropylene
should be a definite confir- mation of these mechanisms but is
beyond the scope of this article. This article will be followed by
a second dealing with the occurrence of maleic an- hydride
homopolymerization in the conditions used for the polypropylene
grafting53 and by a third con- cerning the characterization of
several industrial PP-g-MA.66
REFERENCES AND NOTES
1. R. J. Ehrig, Plastics Recycling, Products and Processes,
2. A. Hogt, Proc. 46th Antec '88, 1478 (1988). 3. A. H. Hogt,
Proc. Compalloy '90, 179 (1990). 4. P. A. Callais, Proc. Compalloy
'90, 359 ( 1990). 5. G. Ruggeri, M. Aglietto, A. Petragnani, and F.
Ciar-
6. C. Macosko, Y. Trolez, and A. Bouilloux, Proc. 33rd
7. Y. Watanabe and T. Hatakeyama, J. Appl. Polym.
8. R. M. Ho, A. C. Su, C. H. Wu, andS. I. Chen, Polymer,
9. N. G. Gaylord and M. K. Mishra, J. Polym. Sci. Polym.
10. D. E. Reid and H. M. Spurlin. U.S. Pat. 3,414,551
11. R. Rengarajan, V. R. Parameswaran, S. Lee, and
12. R. P. Singh, Prog. Polym. Sci., 17, 251 (1992). 13. Y.
Minoura, M. Ueda, S. Mizunuma, and M. Oba, J.
14. U. G. Krefeld, J. M. Korchenbroich, and R. B. Krefeld,
Carl Hanser Verlag, 1992.
delli, Eur. Polym. J., 19, 863 (1983).
IUPAC Macro '90 Session 3.2.7., (1990).
Sci., 37, 1141 (1989).
34 ( 15), 3264 ( 1993).
Lett. Ed., 21, 23 (1983).
( 1968).
P. L. Rinaldi, Polymer, 3 1, 1703 (1990).
Appl. Polym. Sci., 13,1625 (1969).
U S . Pat. 4,370,450 ( 1983).
-
842 DE ROOVER ET AL.
15. R. Rengarajan, M. Vicic, and S. Lee, Polymer, 3 0 ,
16. R. Mulhault, T. Suschek, and B. Rieger, Makromol.
17. M. Lambla and J. J. Flat, Makromol. Kollo. Fribourg,
18. N. G. Gaylord, U.S. Pat. 4,506,056 (1985). 19. S.
Al-Malaika, ACS Symp. Ser., 364, 409 (1988). 20. T. C. Chung and D.
Rhubright, Macromolecules, 2 4 ,
970 ( 1991 ) . 21. D. E. Bergbreiter and J. Zhou, J. Polym. Sci.
Part A:
Polym. Chem., 30, 2049 (1992). 22. G. De Vitto, N. Lanzetta, G.
Maglio, P. Malinconico,
P. Musto, and R. J. Palumbo, J. Polym. Sci. Part A: Polym.
Chem., 22,1335 (1984).
23. B. Immirzi, N. Lanzetta, P. Laurienzo, G. Maglio, G. Maglio,
M. Malinconico, E. Martuscelli, and R. Pal- umbo, Makromol. Chem.,
1 8 , 951 (1987).
24. C. Caze, J. C. Decroix, C. Loucheux, and A. Nicco, Bull.
SOC. Chim. France, 1 1 , 2977 (1973).
25. C. H. Wu and A. C. Su, Polymer, 33,1987 (1992). 26. R. Greco
and P. Musto, J. Appl. Polym. Sci., 44,781
27. A. Crespy, B. Joncour, J. P. Prevost, J. P. Cavrot, and
28. A. J. Oostenbrink and R. J. Gaymans, Polymer,
29. D. Derouet, P. Phinyocheep, and J. C. Brosse, Eur.
30. J. Sheng, X. L. Lu, and K. D. Yao, J. Macromol. Sci.
31. A. Sipos, J. McCarthy, and K. E. Russell, J. Polym.
32. R. Greco, G. Maglio, and P. V. Musto, J. Appl. Polym.
33. R. Greco, P. V. Musto, F. Riva, and G. Maglio, J.
34. R. Greco, G. Maglio, and P. V. Musto, J. Appl. Polym.
35. S. Datta and J. Lohse, ACS Symp. Ser., 6 4 , 121
36. K. J. Ganzeveld and L. P. M. B. Janssen, Polym. Eng.
37. C. A. Strait, G. M. Lancaster, R. I. Tabor, U S . Pat.
38. S. W. Caywood, Jr., U S . Pat. 3,884,882 (1975). 39. N. G.
Gaylord, J. Macromol. Sci. Rev. Macromol.
40. N. G. Gaylord and J. A. Koo, J. Polym. Sci. Polym.
41. N. G. Gaylord and S. Maiti, J. Polym. Sci., Polym.
42. N. G. Gaylord and R. Mehta, J. Polym. Sci. Part A:
933 (1989).
Chem. Macromol. Symp., 4 8 / 4 9 , 317 (1991).
70 (1993).
( 1992).
C. Caze, Eur. Polym. J., 2 3 ( 4 ) , 279 (1987).
33 ( 14), 3086 ( 1992).
Polym. J., 2 6 ( 1 2 ) , 1301 (1990).
Chem., A 2 7 ( 2 ) , 167 ( 1990).
Sci. Part A: Polym. Chem., 27,3353 ( 1989).
Sci., 33,2513 ( 1987).
Appl. Polym. Sci., 3 7 , 789 (1989).
Sci., 3 7 , 777 (1989).
(1991).
Sci., 32 ( 7 ) , 467 ( 1992).
4,762,890 ( 1988).
Chem. Phys., C 1 3 ( 2 ) , 235 (1975).
Lett. Ed., 19 ,107 (1981).
Lett. Ed., 1 1 , 253 (1973).
Polym. Chem., 26, 1903 (1988).
43. N. G. Gaylord, M. Mehta, and R. Mehta, J. Appl. Polym. Sci.,
3 3 , 2549 (1987).
44. N. G. Gaylord, R. Mehta, V. Kumar, and M. Tazi, J. Appl.
Polym. Sci., 38, 359 ( 1989).
45. N. G. Gaylord and M. Mehta, J. Polym. Sci. Polym. Lett. Ed.,
20 ,481 (1982).
46. K. E. Russell and E. C. Kelusky, J. Polym. Sci. Part A:
Polym. Chem., 2 6 , 2273 (1988).
47. M. I. Hendewerk, U S . Pat. Int. Pub. No. WO 90/ 13582
(1990).
48. C. Tzoganakis, J. Vlachopoulos, and A. E. Hamielec, Polym.
Eng. Sci., 2 8 ( 3 ) , 170 (1988).
49. A. Pabedinskas, W. R. Cluett, and S. T. Balke, Polym. Eng.
Sci., 29( 15) , 993 (1989).
50. C. Tzoganakis, Y. Tang, and J. Vlachopoulos, Po1ym.- Plast.
Technol. Eng., 28 ( 3 ) , 319 (1989).
51. S. H. Ryu, Ph.D. thesis, Stevens Institute of Tech- nology,
1991.
52. F. M. Herman, N. M. Bikales, C. G. Overberger, and G.
Menges, Encyclopedia of Polymer Science and En- gineering, Wiley,
New York, 1987, Vol. 4, p. 719.
53. B. De Roover, J. Devaux, and R. Legras, New York, to
appear.
54. W. F. Maddams, Appl. Spec., 34( 3 ) , 245 (1980). 55. F. C.
Billmeyer, Textbook of of Polymer Science, fourth
ed., Wiley, New York, 1966. 56. Ph. Markhal, R. Legras, and J.
M. Dekoninck, J.
Polym. Sci. Part A: Polym. Chem., 31,2057 (1993). 57. B.
Jacques, E. Nield, J. Devaux, and R. Legras, to
appear. 58. B. De Roover, G. Coppens, J. Devaux, and R.
Legras,
to appear. 59. J. March, Advanced Organic Chemistry. third
ed.
Wiley, New York, 1985. 60. L. J. Bellamy, The Infrared Spectra
of Complex Mol-
ecules, Wiley, New York, 1964. 61. C. J. Pouched, The Aldrich
Library of Infrared Spectra,
Aldrich Chemical Company, 1970. 62. F. M. Herman, N. M. Bikales,
C. G. Overberger, and
G. Menges, Encyclopedia of Polymer Science and En- gineering,
Wiley, New York, 1987, Vol. 9, p. 225.
63. M. S. Kellou and G. Jenner, Eur. Polym. J., 1 2 , 883 (
1976).
64. G. Scott, Atmospheric Oxidation and Antioxidants, Vol 11,
Elsevier, Amsterdam, ( 1993).
65. M. Lambla and C. Macosko, Reactive Extrusion of Polymers,
DSM Chair Course, Leuven, 1991.
66. M. Sclavons, V. Carlier, P. Franquinet, B. De Roover, J.
Devaux, and R. Legras, to appear.
Received April 10, 1994 Accepted September 28, 1994