DEGRADATION AND E.s.R. SPECTRA OF IRRADIATED METHACRYLATE POLYMERS
189
On the Degradation and Electron Spin Resonance Spectra of Irradiated Methacrylate Polymers1
by J. F. Kircher, F. A. Sliemers, R. A. Markle, W. B. Gager, and R. I. Leininger Battelle Memo~ialInstitute, Columbus 1 , Ohio
(Received July 10, 1964)
A study is reported of the degradation of irradiated methacrylate polymers, the formation and trapping of free radicals, and their electron spin resonance (e.s.r.) spectrum. Several different radicals are apparently formed and it seems necessary to assume that they are formed by different mechanisms. This is also suggested by a number of observations of main chain degradation and the ratio of chain scission to ester removal for several methacrylate polymers. It is probable that several mechanisms of degradation and radical formation are operative and the one which predominates is dependent on the pendant ester structure.
Introduction At first glance, the problem of the origin of the 5-4 line e.s.r. spectrum in methacrylate polymers irradiated in vacuo appears rather straightforward. It has been shown that radicals trapped in polymerizing methyl methacrylate give a nine-line e.s.r. spectrum like that observed in irradiated polymethylmethacrylate (PMMA) - 2 The propagating radical (I) then appears H CH,
I 1 I 1
-c-c* H COOR I to be the one responsible in both cases. It has been shown that radical I could produce the observed e.s.r. spectrum through the interaction of the electron with the three methyl protons and either none or one of the methylene protons. This produces either a five- or four-line spectrum, depending on methylene group orientation, and the two components appear superimposed to provide the usual nine-line spectrum. 3a,b However, it has not been conclusively demonstrated that this interpretation is correct. Further, it has been known for some time that, in the case of PMMA, approximately one ester scission occurs for each mainchain scission.4,6 Moreover, the same general type of nine-line spectrum was observed for a number of other
methacrylate polymers in addition to P34r\.IA.6-8 Therefore, most attempts to explain the origin of the spectrum invoke a mechanism that resulted in the loss of one ester group per main chain scission and left the propagating radical (I) on one end of the broken chain.8 However, as more data became available from various laboratories it became increasingly apparent that this view was oversimplified. Charlesby and Ormerodgand Campbell and Looney,lo observing the decay of the usual nine-line spectrum, found that the five-line component appeared to decrease more rapidly than the four-line component. This has been confirmed in our own laboratories and is direct (1) This work was supported by the Division of Isotope Development under AEC Contract W-7405-eng-92. ( 2 ) G. K. Fraenkel, J. M .Hirshon, and C. Walling, J . Am. Chem. Soe., 7 6 , 3606 (1954). (3) (a) R. J. Abraham, H. W. Melville, D. W. Ovenall, and D. H. Whiffen, Trans. Faraday SOC.,54, 1133 (1958); (b) D. J. E. Ingram, M . C. R. Symons, and M. G . Townsend., ihid., 54, 409 (1958). (4) P. Alexander, A. Charlesby, and M. Ross, Proe. Roy. 9oc. (London), A223, 392 (1954). (5) A. Todd, J . Polymer Sci., 42, 223 (1960). (6) D. W. Ovenall, ibid., 41, 199 (1959). (7) I. S. Ungar, W. B. Gager, and R. I. Leininger, ibid., 44, 295 (1960). (8) I. S. Ungar, J. F. Kircher, W. B. Gager, F. A. Sliemers. and R. I. Leininger, ibid., A I , 277 (1963). (9) A. Charlesby and ,M.G. Ormerod, Fifth International Symposium on Free Radicals, Uppsala, Sweden, 1961. (10) I. D. Campbell and F. D. Looney. Australian J . Chem., 15, 642 (1962).
Volume 69, Number 1
January 1966
J. KIRCHER, F. SLIEMERS, R. MARKLE,W. GAGER,AND R. LEININGER
190
evidence for the presence of more than one radical. We had reached a similar conclusion earlier from somewhat different considerations of radical decay data.8 Lenk,” on the other hand, came to the same conclusions by observing that the five- and four-line portions of the e.s.r. spectrum did not grow at the same rate during irradiation. There can no longer be any doubt that more than one trapped radical is contributing to the normal nine-line c.s.r. spectrum from irradiated PMMA observed a t room temperature or that these radicals have sufficient lifetimes to be readily detected many hours after irradiation. I t is apparent that the formation and trapping of free radicals in irradiated PMMA is more complicated than first believed and it is probable that more than one mechanism is contributing.
Experimental All experimental techniques have been described previously in some detail and will not be repeated here. The polymers were prepared by irradiation polymerieation of carefully purified monomers. E.s.r. spectra were obtained with a Varian-4500 spectrometer from samples irradiated in vacuo by Co60y-rays a t a dose rate of about 2 X lo5 rads hr.-l. Mass spectrometry and vapor phase chromatography were employed to determine the types and concentrations of the low-molecular weight radiolysis fragments by previously described techniques.’,* l 2
Results and Discussion I t has been established in several investigations that approximately one ester group is removed for each chain scission in irradiated PMMA. However, our work with the isomeric butyl esters shows that this ratio is definitely influenced by the structure of the butyl ester group.8 The ratio of ester removed to chain scission is shown in Table I. The concentration of ester removed was calculated from data on the products Table I : Events per Thousand Monomer Units a t a Dose of 3.0 X 107 Rads
Ester removal Chain scission Ratio
PMMA
P-n-BMA
P-iBMA
BMA
P-tBMA
2.5 3.5 0.71
2.7 10 0.27
3.9 11 0.35
5.3 7.1 0.75
9.6 13.6 0.71
P-sec-
of radiation-induced decomposition and the chain scissions from the decrease in molecular weight for Samples irradiated to 3 X lo7 rads. The product distribution from the irradiated polybutylmethacrylates has been previously described.8 The value of 0.7 for PMMA The Journal of Physical Chemiatry
compares to about 0.8 obtained by Charlesby, et u Z . , ~ and Todd.5 Among the butyl esters there was not a large change in the extent of chain scission but there was a large change in the amount of ester removal. This suggests that the mechanism of scission may change if the structure of the ester group is sufficiently altered or that it is not a t all influenced by ester removal. If the pendant ester group is long enough, cross linking through the pendant ester predominates rather than the usual chain degradation. Graham showed that the n-heptyl methacrylate polymer forms a gel.13 The effect of branching in the pendant alkyl was striking, however. Poly-sec-nonyl methacrylate did not gel even when irradiated to a dose almost three times that which produced gel in the n-heptyl polymer. These results are consistent with the frequent observation that branched hydrocarbons are more susceptible to radiation degradation than unbranched. Degradation in the pendant ester apparently tends to promote main chain scission in the methacrylates. It has been shown previously that very pure PMMA and polyethylmethacrylate (PEMA) do not give the usual nine-line spectrum.7 However, when small amounts of monomer were present, the spectrum was always the characteristic nine-line e.s.r. signal. It was also observed with these and other methacrylate polymers that continued irradiation caused the initially diffused spectrum to sharpen to the normal nine-line pattern.12 This has been interpreted as being due to the initial formation of a radical formed by the loss of an ester group.
Depending on the orientation of the methylene groups, radical I1 would probably give rise to either a four-, five-, or six-line spectrum, the five-line spectrum being more probable than the other two. If all P-hydrogens interact equally, an eight-line spectrum would be expected. Such contributions could lead to the diffuse spectra observed with the very pure polymers. If monomer is present, radical I1 is of course converted to the propagating radical I. It was further suggested
*’’
(‘I’ R. Lenk, J . Phus.9B121 (lg61). (12) F. A. Sliemers, E. Gulbaran, W. B. Gager,
J. F. Kircher, and R. I. Leininger, Second International Radiation Chemical Symposium, Harrogate, England, 1962; ”Radiation Effects,” M. Ebert and A. Howard, North Holland Publishing Company, Amsterdam, 1963. (13) R. K. Graham, J . Polymer sei., 5 8 , 2 0 9 (1959).
DEGRADATION ,4ND E.s.R.SPECTRA O F
IRRADIATED
IfETHACRYLATE POLYMERS
that radical I1 leads to main chain scission by rupture of the C-C bond ,B to the radical site.
H CH3
I /
radical I1 -+ -C-C=CH2
i H
CH3
H
I
I
I
I
+ sC--C-
COOR H
I A reaction of this type produces free radical I again and accounts for the loss of an ester group. Because resolution of the e.s.r. spectrum from irradiated polymers very often is not extremely sharp, it would be difficult to discern the six- or eight-line component from radical I1 in the presence of a four-line component from two different radicals (I and 11). However, it probably would not be lost from the mixture a t the same rate as I and could explain in part some of the effects observed during postirradiation radical decay. With the methacrylate polymers and polymethacrylic acid (PXAA), whenever the initial radiation event is on the main chain, it must be in the vicinity of a quaternary carbon which is well known to represent a weak point on the chain.14 Moreover, when the “R” group is small (H or CH3, for instance) and assuming the initial attack is random, then the vicinity of the quaternary carbon is the most likely place for the initial reaction. It is also known from studies of aliphatic acids that the carbonyl group is particularly susceptible to radiation decomposition, and, as has been pointed out by Todd, if the “R” group is removed by a radiationinduced reaction, then it is very likely that the small stable COz would also be ejected. The net result is radical I1 either way and probably accounts for the fact that in P A N A main-chain scission is accompanied by ester removal a t least 7040% of the time. As the alkyl portion of the ester becomes larger, the probability of having the initial act there increases. Since these events will not always lead to ester scission, the probability of ester removal should decrease with respect to the number of initial acts. If ester removal is necessary for main chain degradation, then degradation should decrease proportionally and the ratio of main chain scission to ester removal should remain about the same. In the case of the butyl methacrylate polymers, it is seen that this is not the situation even when the increased energy absorbed per monomer unit is taken into account (see Table I). With these polymers there are four carbons and nine hydrogens on the ester and only four carbons and five hydrogens in the rest of the monomer unit, so one might - exDect a t least the acts to Occur On groups‘ However, the rate of chain scission does not decrease in
191
proportion to the increase in ester size; it actually increases. The two facts seem to indicate that another mechanism is leading to chain degradation. If the initial act on the pendant alkyl group results in formation of a radical on that group, the radical is in a good position to abstract a hydrogen from a neighboring polymer chain and perhaps thereby initiate chain scission. Designating the alkyl radical R . , we envision a sequence of reactions such as CH3 mCH2-C-
I
CH3
I
+wCH2-Cm
I
hvdroaen abstraction from neiehborine chain
I
COOR
COOR. CH3
CHI
H CHI
I I / c-c-----c-cI I I 1
I I
-CHz-C-
f
a
II
COOR H
COOR H COOR
I11 or CH3
(%I2
I
43HZ-C-CHz-Cm
I
COOR IV
I
I
COOR
followed by chain degradation. CH3
I I
CH3
111 -+ -CHz-C-CH=C
H CH3
I I 1 + .C-CI I /
COORCOOR
H COOR
v CH3
CHz
I V +NCH~-C-CH~-C
I
COOR
/I
I
H CH3
I 1 + ,C-Cw
COOR
I / v
H COOR
Hydrogen atoms or alkyl radicals which are always present during irradiation would also be able to start the same sequence of events. These may account in part for the 20-30y0 of the chain scissions which apparently occur without ester removal even in the case of PMMA. Radicals formed on the pendant alkyl group (14) A. N. Pravednikov, E. N. Teleshov, I. M ,Shen-Kan, and S. J. Medvedev, J . Polymer sei., 58, io39 (1962).
Volume 69,Number 1
January 1966
J. KIRCHER, F. SLIEMERS, R. MARKLE, W. GAGER, AND R. LEININGER
192
probably also account for the results obtained by Graham.I3 Todd5 has suggested that radical V ejects an ester to form a stable chain end. H CH3
CH3
1 1
Tr-C=C-CH
+
C-
I
I
2-
H
*COOR
COOR
However, it does not seem that it can make a major contribution, at least in the case of the butyl methacrylate polymers since this reaction would always lead to an ester removal for each chain scission. However, radical V could possibly abstract a hydrogen from a neighboring butyl group to initiate further chain scission as described above. This could account in part for the increase in scission without corresponding loss of ester as shown in Table I. It must also be remembered that radicals I and V can be formed by direct chain cleavage and initiate the same reactions. Another, but perhaps less likely, way in which scission could occur without loss of an ester group is by the formation of radical VI, which can then decompose to give radicd I and chain scission.
CH3
CH3
I I
42H2-C-CH2-Cw
i
C'OOR
I
+
I
COOR
volatility, e . g . , the methyl butyrates, were formed but not detected in the analyses. Also, it seems quite likely that as the pendant ester group is made larger, reaction a t the a-methyl group becomes less probable. Radical I11 would be expected to produce a two-line e.s.r. spectrum, radicals IV and V three-line spectra, and radical VI one, two, or three lines. How easy it would be to see these in the presence of other radicals giving four- and five-line e.s.r. spectra is open to question. Radicals IV and V would be expected to be very reactive and might not remain trapped in the solid as long as I11 or some other radical somewhat sterically protected. It must also be remembered that many of the polymers used in studies reported in the literature contain some unreacted monomer. Radicals such as IV or V would then be converted to I, increasing the contribution of the latter radical to the five-four-line spectrum. From an analysis of the volatile products arising from the radiation-induced reactions, it is possible to make several statements regarding the positions on the polymer chain where radicals are likely to be formed. Some data from previous work are given in Table I1 for convenience.8 l 2 It is seen that, when R is small (methyl) or branched adjacent to the carbonyl position (sec or tert), half or more of the events apparently lead to the removal of an ester group. These are the same polymers where the ratio of ester removal to chain scission is high. One would expect that the influence of the ester group structure would become more significant as the size of the group was increased. In the case of the n-
CH3 wCHr-c-cHa-dl-
I
I
COOR
+
.CHs
Table I1 : Estimated Concentrations of Free Radical Sites in Irradiated Methacrylate Polymers"
COOR --Site6
VI
VI
-
I
COOR
+ d-CHaI
COOR I
This requires the formation of CH3. which should be accounted for in the volatile products measured after irradiation. In the case of PMMA, it was found that there are more methyl radical-derived products than can be accounted for on the basis of methyl radicals from ester group scission, which suggests such a mechanism is occurring to some extent.'2 This is not true in the case of the isomeric butyl esters. However, in those cases, it is possible that some products of very low The Journal of Phyaical Chemielry
P-tBMA
Radical
PMAA
H CH,Hb
5.4
2.5
2.7
3.9
5.3
9.6
. .
2.7
4.7
5.0
5.4
3.6
5.6
7.7
8.9
11.0
13.2
CH3 -CH2-C=CH2
per thoueand and monomer units-P-nP-iP-seePMMA BMB BMA BMA
I 1 I wc-c-cI I H
H
Formed b y H atom extraction from backbone of ester' Total
,
5.4
a After exposure to a dose of 3.0 X lo7 rads. Calculated from sum of concentrations of CO, CO,, and ester. Calculated from sum of concentrations of hydrocarbon (excluding isobutene in the case of the &butyl polymer), hydrogen, alcohol, and formate.
DEGRADATION AND E.s.R. SPECTRA OF IRRADIATED METHACRYLATE POLYMERS
butyl and isobutyl polymers, less than half of the products arise from the removal of an ester group. One might reasonably expect that some radicals could be trapped on the butyl group and thereby further confuse the-e.s.r.pattern. Several different structures are possible which could lead to four-, five-, or six-line spectra, for instance. Table I11 : Concentration of Free Radical Sites in Irradiated Methacrylate .Polymers as Determined by Different Methods -Sites
per thousand monomer unitsaFrom From total molecular weight From product change 8.s.r. signal analysis
PMAA PMMA P-n-BiLIA P-i-B-MA P-sec-BMA P-kBMA
5.4 5.6
7.7 8.9 11. 0 13.2
n.d. 3.5 10.0 11.0
7.1 13.6
1.3 0.8 0.06 0.2 0.6 0.6
Following irradiation to a dose of 3 X 10' rads. from e.s.r. data a t low doses.
Initial G-value*
2.4 2.4 1.6 3.4 2.9 2.5
* Calculated
From an analysis of the products of radiolysis, it is possible to estimate the number of bonds which must have been broken and hence the number of radicals which could be trapped in the polymer.* Another estimate of maximum trapped radicals can be obtained from the decrease in molecular weight by assuming a certain fraction of chain ends are trapped as free radicals. E.s.r. provides a more direct measure of these trapped radicals a t any given dose. However, the trapping efficiency is not 100% in any case and, consequently, somewhat lower free-radical concentrations are to be expected from the e.s.r. data. The e.s.r. data, a t low doses, can also be used to determine rate of formation or trapping of the radicals. Such data are briefly summarized in Table 111. Generally, it can be seen that the apparent rate of radical formation and trapping does not vary among the different polymers to any great extent, and this is to be expected. The most ob-
193
vious variation attributable to methacrylate structure is in the free radical concentration as observed by e.s.r. after an exposure of 3 X lo7 rads. However, it should be pointed out that this may be primarily due to the influence of ester structure on the physical properties of the polymer (particularly glass-transition temperature) rather than the chemical reactions undergone during irradiation. From the data of Graham13 and Shetter,I5 it appears that all of our experiments were performed below T , for all materials except possibly the P-n-BMA. T , for P-n-BMA is about 19-20', and the experiments were conducted a t 21-22", The trapping efficiencyis expected to be less above T,; therefore, the results shown in Table I11 are not unreasonable, since a lower trapping efficiency means the concentration of radicals seen by e.p.r. at any time after irradiationwillbelow. However, T, would not be expected to have this effect on radical concentrations calculated from molecular weight changes or product formation.
Summary From the evidence accumulated in this and other studies, it appears that a number of radical species are being formed and trapped when polymethacrylates are irradiated. It seems necessary to assume that more than one mechanism is producing main chain degradation and radicals that can be trapped. In the case of PMMA, 70-80% of the degradation is accompanied by ester group loss. The five-four-line e.s.r. spectrum has been seen in almost every study and one of the major reasons for this may be unreacted monomer present in the methacrylate polymer sample. In those cases where the polymer was stringently purified, some change was usually observed in the e.s.r. pattern. It is also possible that the propagating radical is the most efficiently trapped and stable radical so that its relative concentration tends to increase with exposure time. The concentration of radicals seen by e.s.r. a t a given dose is influenced by ester group structure in the polymethacrylates. However, this influence is predominantly through the effect on T , and hence the efficiency of radical trapping rather than radical formation. (15) J. A. Shetter, Polymer Letters, 1, 209 (1963).
Volume 69,Number 1
January 1966
