50 Melby Macromolecules

References and Notes of fiber plus solvent.

(17) K. R. Porter and C. V. Z. Hawn, J. Erp. Med., 90,225 (1949).
(18) J. D. Ferry, S. Shulman, K. Gutfreund, and S. Katz, J . Am. Chem. Soc.,
(1) This work was supported by a research grant from the National Institutes 74,5709 (1952).
of Health (HL-20319). (19) P. Kaesberg and S. Shulman, J . Biol. Chem., 200,293 (1953).
(2) J. D. Ferry, Adu. Protein Chem., 4 , l (1948). (20) B. M. Siegel, J. P. Mernan, and H. A. Scheraga, Biochim. Biophys. Acta,
(3) H. A. Scheraga and M. Laskowski, Adu. Protein Chem., 1 2 , l (1957). 11,326 (1953).
(4) R. Doolittle, Adu. Protein Chem., 27,l-109 (1973). (21) L. Stryer, C. Cohen, and R. Langridge, Nature (London), 197, 793
(5) J. D. Ferry and P. R. Morrison, J . Am. Chem. Sac., 69,388 (1947). (1963).
(6) S. Shulman and J. D. Ferry, J . Phys. Colloid Chem., 54,66 (1950). (22) R. Gollwitzer, H. E. Karges, H. Hormann, and K. Kuhn, Biochim. Biophys.
(7) D. F. Waugh and M. J. Patch, J . Phys. Chem., 57,377 (1953). Acta, 207,445 (1970).
( 8 ) Z. S. Latallo, A. P. Fletcher, N. Alkjaersig, and S.Sherry, Am. J . Physiol., (23) W. Krakow, G. F. Endres, B. M. Siegel, and H. A. Scheraga, J. Mol. Biol.,
202,675 (1962). 71,95 (1972).
(9) M. E. Carr, L. L. Shen, and J. Hermans, Biopolymers, 1 6 , l (1977). (24) B. Mathews, Annu. Reu. Phys. Chem., 27,493 (1976).
(10) C. V. Z. Hawn and K. R. Porter, J. Erp. Med., 86,285 (1947). (25) N. M. Tooney and C. Cohen, J . Mol. Biol., 110,363 (1977).
(11) N. V. Bang, Thromb. Diath. Haemorr., Suppl., 13,73 (1964). (26) T. Neugebauer, Ann. Phys. (Leiprig), 42,509 (1943).
(12) D. Kay and B. J. Cuddigan, Br. J . Haematol., 13,341 (1967). (27) A. Guinier and G. Fournet, “Small Angle Scattering of X-Rays”, Wiley,
(13) G. J. Stewart, Scand. J . Haematol., Suppl., 13,165 (1970). New York, N.Y., 1955.
(14) J. M. Buchanan, L. B. Chen, T. Hamazaki, E. Lenk, and D. F. Waugh in (28) E. F. Casassa, J . Chem. Phys., 23,596 (1955).
“Chemistry and Biology of Thrombin”, R. L. Lundblad, J. W. Fenton, and (29) B. J. Berne, J . Mol. Biol., 89,755 (1974).
K. G. Mann, Ed., Ann Arbor Science Publication, 1977, p. 263. (30) 0. Kratky, Makromol. Chem., 35a, 12 (1960).
(15) G. E. Perlman and L. G. Longsworth, J . Am. Chem. Soc., 70, 2719 (31) V. Luzzati, Acta Crystallogr., 13,939 (1960).
(1948). (32) B. H. Zimm, J . Chem. Phys., !6,1099 (1948).
(16) Notice that the value of the mass-length ratio obtained by light scattering (33) R. G. Kirste, Z. Phys. Chem. (Frankfurtam Main),42,351 (1964).
refers to the mass of the scattering material, i.e., the protein. The density (34) A. Fedorov and V. G. Aleshin, Vysokomol. Soedin., 8,5016 (1966);Polym.
calculated here is therefore very much different from the buoyant density Sei. USSR (Engl. Transl.), 8,1657 (1967).

Nitric Acid Oxidation of High-Density Polyethylene.

Organic Chemical Aspects
L. Russell Melby
Central Research and Development Department, E. I . du Pont de Nemours and Co.,
Wilmington, Delaware 19898. Received J u n e 20, 1977

ABSTRACT: Preparative-scalenitric acid oxidation of bulk, high-density polyethylene has given high yields of ni-
trated apalkanedioic acids with molecular weights in the range 1300 to 3000 similar to products reported in the lit-
erature. The acids have polydispersity indexes of about 1.05-1.3 and thus are very nearly monodisperse. Carboxyl
to nitro group ratios vary between about 6:l and 1.6:l depending on preparation conditions. Concentrated sulfuric
acid converts the nitro groups to keto groups forming keto acids of unchanged molecular weight. Partial reduction
of the keto groups can be achieved by the Clemmensen or Wolff-Kishner procedures. In some of the oxidation reac-
tions, dinitroalkanedioic acids of molecular weight 400 to 500 were obtained;in these it is shown that the nitro groups
are situated near the chain ends in positions a , p, or y with respect to the carboxyl group.

T h e extensive work of Keller and his co-workers on the Results and Discussion
oxidation of polyethylene single crystals has done much to
unravel the problem of fold surface structure in crystalline Mechanics and Conditions of the Oxidation-Nitration
polyethylene.2 In such oxidations, the amorphous chain folds Reaction. The nitric acid oxidations described in the litera-
a t the periphery of the crystallites are severed and degraded ture were carried out on small amounts of carefully recrys-
back to the more resistant crystalline portions of the poly- tallized polyethylene (0.1-0.3 g) in heated, sealed glass tubes
mer. and reaction times of 1t o 3 week~.~-lO Since our preparative
Although some of the more recent work from Keller’s lab- work demanded large amounts of material, these limitations
oratory used ozone as the 0xidant,3>~ most of the earlier work were unacceptable.
was done with fuming nitric acid5q6following similar previous In preliminary experiments designed to permit larger scale
studies by Palmer and Cobbold.7 The products of the nitric development, several oxidations of high-density polyethylene
acid oxidation were identified as nitrated, long-chain alkanoic pelletsl1J2J3 were carried out on a 50-g scale in sealed rocker
acids,6,s and although some passing interest has been shown bombs. In general, these proceeded satisfactorily and gave
in chemical manipulation of such products as gel-permeation products very similar to those of Keller e t al. However, the
chromatography standardsQor polymer intermediates,1° most potential explosion hazard of such systems, particularly if
of the work in this area was done from a polymer physical accidentally heated above the polymer melt temperature,
structure viewpoint. encouraged us t o explore the reaction in open vessels a t at-
Thus, our attention was attracted to the organic chemistry mospheric-pressure reflux temperatures. In this way, we were
of the process, its preparative aspects, and further charac- able to operate on a 200-g scale.
terization of the products. In most of the published work referred to, 95% fuming nitric
Our characterization work in particular has clarified some acid was used although some was done with 82.5% a ~ i d . ~ , ~
inconsistencies in the literature and has led us to some further Constant reaction temperatures of 60 or 80 “C were used. We
speculation on the course of the oxidation-nitration reac- worked with three different strengths of acid, namely, ordi-
tion. nary concentrated (71% HN03), 9090 fuming, and red fuming

0024-9297/78/2211-0050$01 .OO/O 0 1978 American Chemical Society

Vol. 11, No. 1,January-February 1978 Nitric Acid Oxidation of High-Density Polyethylene 51

Table I
Nitric Acid Oxidation of High-Density Polyethylenes, a Nitrated Alkanedioic Acids
Preparation conditions Solids
Prepn Nitric Stirring Time, recovery, GPC data Rll IR ratiog
No. acid typeC mode h %e gw M, PIf titration COzH/NOz
l h RF 16 82 2400 1800 1.17 5.92
2 RF 96i 93 1700 1300 1.19 1350 5.19
3J RF 60 99 2900 2100 1.26 1900 5.35
4 RF 90 97 4400 3 100 1.27 2400 4.78
5 RF 24 92 7000 4100 1.27 3250 3.00
6 RF 24 33 2800 2400 1.06 1900 2.65
7 F 60 91 3400 2600 1.16 2300 4.15
8 F 60 95 3600 2800 1.14 2400 3.86
9 F 60 95 2800 2300 1.07 2200 3.77
10 F 40 92 3800 2800 1.21 2350 3.21
llJ F 40 94 2400 1800 1.17 1700 3.98
12 F 60 33' 2200 1900 1.01 1700 2.74
13 F 48 26 1900 1500 1.13 1800 2.56
14 F" F 24 3300 2600 1.17 2000 3.25
14 P" 3700 2800 1.19 2300 3.06
15 FnsO F 24 3400 2600 1.15 2150 3.40
15 Pnso 24} 89
65 3800 2900 1.17 2350 3.07
16 Conc 60 56 2800 2400 1.05 1900 2.63
17 Conc 20 43p 2800 2000 1.24 2150 1.63
Alathon 7040 polyethylene resin (see ref 12) except for preparation 15. Elemental analyses not given because they are dominated
by the hydrocarbon portion so that differences are not significant. RF is red fuming, F is 90% fuming, and Conc is ordinary concentrated
71% "03. Mag is magnetic stirring and Mech is mechanical paddlcstiEing. e These are not percent yields but rather the ratio
of weights of solid recovered to solid charged. f Polydispersity index, M,/M,, corrected for instrumental broadening. g 2% in KBr
wafers, ratio of absorbance at 1708 cm-l (COzH) to 1550 cm-' (NO2); see ref 21. Closed bomb, 80 "C, autogenous pressure. This
experiment by R. C. Ferguson. i Two stages; pellets oxidized 48 h, recrystallized from xylene, and oxidized 48 h more. I Recrystallized
from xylene before oxidation. Extra agitation caused by the thermometer inserted directly into the reaction mixture. Product ac-
companied by 31% nitro-carboxylic grease for which see Table 11. a,,
= 2000, ebullioscopic in xylene. Solid accompanied by low yield
of nitro-carboxylic grease. See Table 11. M , = 1950, ebullioscopic in xylene. Solid accompanied by 21% nitro-carboxylic grease. See
Table 11. " Two solid constituents, fines (F) and pellets (P). Starting material Alathon 7050 polyethylene resin. p Accompanied
by 11% nitro-carhoxylic grease.

nitric acid (>95% " 0 3 ) . Since these acids were used a t their carboxyl groups on treatment with hot concentrated sulfuric
reflux temperatures a t atmospheric pressures, reaction tem- acid. Indeed, such a transformation could occur by the well-
peratures were necessarily self-regulated within safe limits. known hydroxamic acid rearrangement and hydr01ysis.l~
However, since oxides of nitrogen are continuously lost under
these conditions, the concentration of acid changes as does the
reaction temperature. For example, with red fuming nitric
acid, the initial reflux temperature was about 55-60 "C but
increased to 90-100 "C after 24-48 h reaction time. These
RCH2N0, - [RCH
"O
'H
0
]- RC
//O

O
'H
variations caused some difficulty with respect to reproduc- On these bases, the original products would be a mixture
ibility. Nevertheless, these different conditions afforded of compositions such as 1,2, and 3. Because of the high car-
substantial variation in the products with respect to yield, boxyl to nitro group ratio, contributions by 2 and 3 would be
molecular weight, and carboxyl to nitro group content. This relatively small.
contrasts with the closed systems which gave products rela-
tively invariant in these respects.6,8-10 HO,C(CH,),CO,H 0 ,NCH,( CH, ), CO,H
Moreover, we found t h a t the degree of agitation of the re-
action mixtures is an important determinant of product 1 2
composition and yield. The importance of this factor is not O,NCH, (CH, ,) CH,NO,
evident from previouely published work. Although not spec-
ified, these were presumably static systems and one infers 3
from them t h a t the major determinant of product character I n a still later paper,s Keller and Udagawa modified their
is the traverse length of molecules in the crystallites. However, view and concluded t h a t the chains were terminated solely by
Keller and Udagawa8 have shown that state of subdivision and carboxyl groups b u t t h a t the "function and position of the
reaction time are also important factors. nitro group remains unknown", but they did not reconcile this
N a t u r e of t h e P r o d u c t s . Before discussing our products, revised view with their nitro group removal experiments. In
it would be well to summarize the conclusions of Keller and separate and still later work, Ballard and Dawkins retained
co-workers. the view t h a t the nitro groups were terminal and carried out
In a n early paper, they concluded t h a t each stem end of the nitro group removal procedure on the assumption that
their oxidized products possessed two oxygen atoms and they were generating uniform carboxylic acid bifunctional-
therefore each molecule was terminated by either a carboxyl ity.1° As we shall see from the work described below, there is
group or a nitro group.6 Moreover, the carboxyl to nitro group now no doubt t h a t the oxidation products are a t the outset
ratio was about 11:l.I h a later paper, these conclusions were terminated solely by carboxyl groups and that the nitro groups
sustained,g8 and the notion that the nitro groups were terminal are indeed secondary.
(primary) was reinforced by their presumed conversion to In Table I are summarized the properties of our primary
52 Melby Macromolecules

Table I1
Characterization of Nitro-Carboxylic Greases
Source
Table I Anal. Titration=
Prepn prepn Calcd Fouad neut equiv IR ratio
No. No. Formula FW %C %H %N %C %H %N 1st 2nd C02H/N0zb
1 C 12 CisH32N208 404.5 53.4 7.9 6.9 53.2 7.8 6.9
2d 13 C21H3sN208 446.5 56.5 8.6 6.3 56.2 8.6 6.3 402 190 1.06
3e 17 C ~ ~ H ~ S N ~446.5
OS 56.5 8.6 6.3 56.4 8.5 6.0 399 198 1.08
4f 6 C22H40NzOs 460.6 57.4 8.8 6.1 57.6 8.9 5.9 496 246 1.10

a Potentiometric in acetone with (n-Bu)dNOH titrant. Ratio of 1708 and 1550 cm-' absorbances. Run neat. Bn= 635, bp in
C6H6. lH NMR in CDC13 with Me4Si reference: 6 1.3-2.4 (m, CH2,32), 4.5 (broad s, HCN02, 2), 11.6 (s, COOH, 2). Reference mixture,
1:l mole ratio dodecanoic acid + 1-nitrododecane: 6 1.3-2.4 (m, CH2, 44), 4.39(t, CHzN02, 21, 11.5 (s, COOH, 1). M, = 780, bp in
C6H6. *H NMR: 6 1.3-2.4 (m, CH2,34),4.5 (5, HCN02,2), 11.4 (s, COOH, 2). e M, = 780, bp in C6H6. 'H NMR: 6 1.3-2.4 (m,CHz, 36),
4.5 (s, HCN02, 2), 11.2 (s, COOH, 2). f %, = 750, bp in C&. 'H NMR: 6 1.3-2.4 (m, CH2, 40), 4.5 (s, HCN02, 21, 11.3 (s, COOH, 2);
also 6 5.05 (8, vinyl CH, trace). l3C N M R 6 180.48,180.09, 179.70, 179.11,178.07 (COO), 123.03,120.10 (vinyl C), 88.91,88.71,88.52,
88.26 (CNOz), 34.05-20.86 (CH2).

products and the procedures used. The first notable feature original polyethylene or, conversely, the low amorphous
is the spread of number average molecular weights (R,) content.
which, by GPC, vary from 1300 for preparation 2 to 4100 for The matter of nitration will now be considered. As seen in
preparation 5. In these two extremes, the main determinant Table I, our products in general show a much higher degree
of the molecular weight difference is undoubtedly the state of nitration than those of Keller and co-workers who reported
of subdivision of the starting material. The effect of this pa- carboxyl to nitro group ratios of about ll:l.s79~20~21 Thus,
rameter is more clearly shown in preparations 10 and ll in preparation 17 with a COOH/N02 ratio of 1.63is very exten-
which the other factors, time, temperature, and stirring mode, sively nitrated. However, most of our examples cluster around
were constant; the a,of the product from pellets was 2800 3 or 4:l. A more pictorial way of expressing these ratios is in
whereas that from the finely divided, recrystallized material15 terms of percentage nitration.22 For example, a COOH/N02
was 1800, and yet the yields were essentially equivalent. A ratio of 4:l corresponds to 50% nitration which is to say that
similar order of change was shown in the work of Keller et aLgb statistically, 50% of the molecules are mononitrated and the
They observed peak molecular weights by GPC of 1240-1360 remainder are free of nitro groups. A ratio of 1.63(preparation
for products from recrystallized polyethylene and 1750-1860 17, Table I) corresponds to about 125% nitration; that is,
for that from bulk ~ e 1 l e t s .Ballard
l~ and Dawkin working with statistically 75% of all molecules are mononitrated and the
recrystallized starting material obtained product U,'s of remainder are dinitrated. There may, of course, be distribu-
1200-1550 by osmometry and viscometry. In neither of these tions among unnitrated and other polynitrated species, but
cases were polydispersity indexes given although calculations there is presently no way of distinguishing such detail.
in this laboratorylS from the data of Keller et al. suggest that The N i t r o Carboxylic Greases. The major difficulty in
their products had polydispersity indexes of about the same fully characterizing the high molecular weight acids is their
order as ours. relative intractability and the fact that the functional groups
The effect of agitation is dramatically illustrated by com- are so highly diluted by the hydrocarbon residue. But in sev-
paring preparations 7 and 12 where all factors are the same eral experiments in Table I, the solids were accompanied by
except for the method of stirring. In a magnetically stirred acetone-soluble greases (preparations 6, 12,13, and 17). We
mixture, the bed of pellets floats on the more dense acid and consider these to be low molecular weight homologues of the
rotates very slowly as a relatively undisturbed mass. Me- solids and present them as tractable models whose charac-
chanical paddle stirring constantly tumbles the pellets as the terization could reinforce otherwise questionable conclusions
bed rotates a t a rate very close to that of the stirrer rate. T h e about the solids.
yield of solid from the magnetically stirred reaction was 91% In Table I1 are summarized the data on characterization of
and the mn (GPC) was 2600. The mechanically stirred mixture these greases. The methods include elemental analyses, in-
product had a much lower an (1900) and the yield of solid was frared spectra, l H and 13C NMR, nonaqueous titration,
only 33%; however, it was accompanied by a substantial ebullioscopic molecular weights, and additional data on sev-
amount of low molecular weight grease. We shall discuss these eral derivatives. In appearance they are light-yellow, opales-
greases later. A similar agitation effect is shown in the results cent materials with the consistency of honey, and all have very
of preparations 5 and 6. similar compositions regardless of differences in the prepa-
As far as acid concentration effects are concerned, other ration conditions. In summary, they have compositions cor-
things being equal, there seems to be relatively little difference + + +
responding to 4,where x z = 0-6 and x y z = 14-17.
between the two kinds of fuming acid (preparations 3 and 7
or 4 and 14). But the more aqueous concentrated nitric acid HO,C(CH,),CH+CH,fy CH+CH,),CO,H
is considerably more destructive (preparations 16 and 17). I I
Polydispersity indexes (nw/an) for the products listed in NO, NO,
Table I vary between about 1.05 and 1.3, but there is no ap- 4
parent relationship between the breadth of dispersity and the
preparation conditions. The remarkable feature is that re- Again, it must be understood that this is a statistical compo-
gardless of conditions, the indexes are so very near unity or sition leaving open the possibility that some species are less
monodispersity. This in turn must reflect the integrity of the highly nitrated and others more highly nitrated.
crystallites in the starting polymer. In most of the examples, The infrared spectra clearly show almost exactly a 1:l mole
yields of high molecular weight products were high (90-95%) ratio of carboxyl to nitro groups and elemental analyses in-
which would be expected from the high crystallinity of the dicate compositions corresponding to about C10H19N04 or a
Vol. 11,No. 1, January-February 1978 Nitric Acid Oxidation of High-Density Polyethylene 53

multiple thereof. T h e simple formulation could correspond HO,C--R-CH-R-CO,H

to either 5 or 6. Structure 6 can be discarded out of hand be- 1
HO,C(CH,),CH,NO,

5
H0,C(CH,)xCH-(CH,)yCH3
I
NO2
- NO,

HJo+ 1
H0,C-R-C-R-CO,H
\T
II 1
6,x + y = 7
cause the survival of terminal methyl groups would be van- 10
ishingly small. Structure 5 with the terminal primary nitro
group is ruled out because the proton NMR spectra show that
the carboxyl proton ( 6 11.3-11.5) and protons (Y t o the nitro
group ( 6 4.5) are present in a ratio of 1:l not 1:2; the 1:2 ratio
[HOZC-R-C-R-CO,H

L
II
0
12
J
Zn
HO,C(CH,),CO,H
9

is clearly seen in an equimolar reference mixture of dodecanoic 11

acid and nitrododec,ane. Therefore, the structure must be a
dimer of a fragment such as 7. T h e yield of 9 was very low (<2%) presumably because of
poor solubility relationships in the Clemmensen reduction
system and because the simple aqueous acid catalyzed Nef
reaction of a free nitro compound is extremely slow; ordinarily
NO, the Nef reaction is carried out on an alkali metal salt of the
already enolized nitronic acid (cf. 10) from which the ketone
7 is generated very rapidly.25
T o the extent one believes the greases to be models of the
T h e mass spectrum of preparation 3, Table 11, shows a
long-chain acids, the implication is that the nitro groups in
major parent ion with mass number 398 which would corre-
+ +
spond approximately to 4 with x y z = 14, and potentio-
those acids are very near the carboxyl ends. I t follows that the
nitration reaction is not random. This near-terminal nitro
metric titration clearly distinguished two carboxyl groups a t
group disposition may simply be determined by the accessi-
breaks corresponding to the molecular weight (399) and the
bility of the polar-ended crystallites to the nitrating species.
neutral equivalent (I 99).23Thus, the titration data agree al-
T h a t is to say, after oxidation of crystallite chain folds, the
most exactly with the mass spectral data.
residual crystallites would have polar surfaces (COOH) a t the
From the 13C NMII, we conclude that the nitro groups are
longitudinal termini. One could conceive that the nitrating
near neighbors of the carboxyl groups. The 13C resonances of
species would penetrate a t least a short distance into such
t h e COOH a n d C-NO2 groups are each resolved into several
surfaces.
peaks. These different chemical shifts can be evident only if
Nitro Group Removal. Earlier we referred to the ambi-
t h e two groups are within a few carbon atoms distance of one
guity of published conclusions concerning the nitro groups in
another, otherwise the resonances would appear as two single
the high molecular weight a ~ i d ~ . ~We , ~now
, ~ introduce
~ J ~
peaks with chemical shifts somewhat lower than the minor
evidence which shows conclusively that the nitro groups are
peaks observed for each. T h e nitro groups must, therefore, be
not primary but secondary and that their removal by hot
in positions a, p, y, anid perhaps 6 with respect to the carboxyls
+
and hence the conclusion t h a t in 4 x z = 0-6.
sulfuric acid generates ketonic carbonyls not carboxyl
groups.
These low molecu!lar weight materials show a very strong
Procedures in the literature involve heating the nitro acids
tendency t o associate as reflected in the inordinately high
with concentrated sulfuric acid in sealed tubes for periods of
molecular weights determined ebullioscopically in benzene.
about 1to 2 days.26We found this unnecessarily cumbersome
T h e values shown in Table I1 (650-750) were, in fact, ob-
and instead achieved the same result by stirring recrystallized
tained by extrapolating to zero concentration. Nitroalkanes
nitro acid with 95-97% sulfuric acid in an open glass vessel a t
are known to associate more or less strongly, depending on the
85-95 "C for 6 h. This eliminates substantially all the nitro
solvent;24t h e presence of a multiplicity of nitro groups and
groups as judged by infrared spectra. As obtained, these
strongly associating carboxyl groups in the same molecule no
products are dark gray and dissolve in boiling xylene to form
doubt accounts for the high apparent molecular weights. Es-
terification of preparation 3 gave a dimethyl ester whose in- dark gray-brown solutions. This color obscures the phenol-
tegrated proton N M l i and elemental composition conformed phthalein end point in neutral equivalent determinations so
closely to the C23 product 8. The ebullioscopic molecular the products must be bleached. Treatment with 15%hydrogen
peroxide a t 60-70 "C served this purp0se.2~
CH30zC(CHzhCH+CH,)jj CH-(CH,hCO,CH, The data on these "denitro" acids are summarized in Table
I I 111. The salient point is that, considering the nitro group
NO, NO, contents of the precursor acids, the neutral equivalents of the
8, x + y + 2 = 17 denitrated products should be between 15 and 25% lower i f
t h e nitro groups are converted to carboxyls, but such is not
weight of the ester was 470 (theory 474) with no indication of so. In general, the neutral equivalents were within experi-
association.
mental error (f2-5%) of being identical for product and pre-
Final support in chmacterizing the nitro-carboxylic greases cursor, and where there is a significant difference, the devia-
was found in the Clernmensen reduction of preparation 4 to tion is toward larger neutral equivalents. Thus, the nitro
a semicrystalline dicarboxylic acid whose composition, inte- groups are not converted to titratable end groups. The only
grated proton NMR, molecular weight, infrared, and neutral
alternative is that they are secondary a t the outset, and the
equivalent showed an average composition corresponding to sulfuric acid converts them to ketonic carbonyls via the Nef
9. I t was presumably formed by an acid-catalyzed Nef reaction
reaction (see above). This should augment the carbonyl in-
HOzC(CHzh7CozH frared absorption band relative to an internal standard such
9 as a CH band. Unfortunately, this could not be established by
and sequential Clemmensen reduction of the derived ketone infrared peak heights because absorption by the putative keto
group(s); schematically carbonyls is not linearly additive to that of the carboxyl car-
54 Melby Macromolecules

Table I11
Keto Acids Derived from t h e Nitro Acids
Keto acid Nitro acid GPCdata a n by Sulfuric acid
prepn No." Table I prepn No w, Mn PIb titration treatment time, h
1 4 4500 3200 1.26 2450 6
2 4 4300 3200 1.20 2500 4
3' 7 3790 2900 1.16 2150 3
4 8 3200 2500 1.13 25
5 8 3000 2500 1.11 2500 25
6 9 2800 2300 1.08 2150 6
7 13 3600 3000 1.20 6
8 14 F 3400 2600 1.28 2000 5
9 17 2900 2500 1.17 6
a Bleached with H202 unless otherwise noted. C02H/N02 IR ratios varied between about 15:l and 30:1, NO2 absorbance being so
weak as to be near the noise level, Bleached with Clz in water. Product contained organic chlorine. Corrected for instrumental
broadening.

Table IV
Infrared Absorption Ratios a of Nitro Acids and DNP's of Keto and Reduced Keto Acids
Nitro acid Keto acid Hydrazone/COzH ratiod
Exptl Table I N02/COzH Table I11 Before After %
No. prepn No. ratioC prepn No. reduction reduction reduction
1 17 0.61 9 0.78 0.41e 47
0.27f 65
2 13 0.39 7 0.76 0.24f 68
3 6 0.38 g 0.71 0.35h 51
0.19; 73
4 9 0.27 6 0.43 0.301 30
0.22" 49
5 4 0.21 2 0.45 0.Mh 60
0.13j 71
a 2% in KBr wafers. 2,4-Dinitrophenylhydrazone.e 1550/1710 cm-I absorbance ratio. 1615/1710 cm-I absorbance ratio. e 72-h
Clemmensen, atmospheric pressure. f Wolff-Kishner. g Prepared but not further characterized. 24-h Clemmensen, atmospheric
pressure. I 96-h Clemmensen, atmospheric pressure. J 24-h Clemmensen, 145 "C, bomb. 36-h Clemmensen, 145 "C, bomb.

bonyls at the wavelength of maximum absorption. There was the insoluble zinc. Under these conditions, reduction is very
some broadening of t h e carbonyl band but t h e evidence was slow so that even after 96 h of reaction time it was only about
at best only suggestive. 75% complete as inferred from t h e infrared spectrum of t h e
However, condensation of t h e "denitro" products with derived 2,4-dinitrophenylhydrazone.
2,4-dinitrophenylhydrazinegave hydrazones whose infrared When the reaction is carried out in a closed bomb at 145 "C,
spectra showed strong, well-resolved bands a t 1615 and 1590 t h e keto acid is completely soluble in t h e xylene, but even
cm-l. Moreover, t h e peak height ratios of t h e 1710 (carboxyl under these conditions complete reduction was not achieved
carbonyl) to 1615 hydrazone bands were approximately pro- even after 36 h.
portional to the CO/N02 ratio in the parent nitro acids t h u s Modified Wolf-Kishner reduction29of t h e keto polymers
demonstrating the transformations in refluxing xylene with hydrazine and potassium tert- bu-
-CH- -C- -C- toxide also eliminated a substantial proportion of the keto

I
NO2 - II
0 - Lp0*
groups. In this case, the reaction should ostensibly be homo-
geneous since the keto acids are soluble in boiling xylene.
However, t h e potassium tert- butoxide converts them to po-
tassium salts which are insoluble gels. Thus, the reaction re-
NO2 mains heterogeneous and consequently very slow, being about
and establishing that the "denitro" acids are in fact keto acids. 65-70% complete in 18 h.
Representative infrared data are compiled in Table IV. T h e infrared data relating t o these experiments are given
Reduction of the Keto Acids. The keto groups in the keto in Table IV.
acid can be partially reduced t o methylene groups either by
Clemmensen reduction with amalgamated zinc and hydro- Experimental Section
chloric acid or by Wolff-Kishner reduction with hydrazine and Materials and Methods. The polyethylene resin used was high-
potassium tert- butoxide. density Alathon 7040,12 except one instance in which Alathon 7050
In the Martin modification of the Clemmensen reduction,2S polyethylene resin was used.13This is noted in Table I.
a water-insoluble organic phase is used t o dissolve t h e sub- The nitric acids were commercial c ~ n c e n t r a t e d9Wo
, ~ ~ fuming,30
strate. We used xylene. However, when t h e reaction is run at or red fuming.31
atmospheric pressure, the maximum temperature attainable Titration of the high molecular weight products was conducted as
is limited by the reflux temperature of t h e aqueous hydro- follows: the titrant was prepared by dissolving 1.4 g of KOH in 70 mL
of ethanol and then diluting with 375 mL of reagent-grade xylene. The
chloric acid (-105 "C). This is not high enough t o dissolve the titrant is nominally 0.5 N and was standardized against decanoic acid
keto acid. Thus, one is dealing with a four-phase system: t h e (0.2 g) in 25 mL of methanol with 3 drops of 5% methanolic phenol-
dispersed keto acid, the organic liquid, the aqueous acid, and phthalein as indicator. To titrate the oxidation products, 1 g was
Vol. 11, No. 1,January-February 1978 Nitric Acid Oxidation of High-Density Polyethylene 55

dissolved in 50 mL of xylene heated just below its boiling point (-125 sucked free of the bulk of water and was redispersed in 500 mL of
"C) on a magnetic stirring hotplate and 3 drops of pheGolphthalein water. To it was added 100 mL of 30% hydrogen peroxide and the
was added. The stirred solution was kept hot and titrated dropwise slurry was stirred and heated at 65-70 "C for 6 h. It is important not
to the pink end point. When stearic acid was run as a control, neutral to overheat the mixture because vigorous foaming may occur and eject
equivalents within fO.!j% of the theoretical value were obtained. material from the flask.
Infrared spectra were run on a Perkin-Elmer Model 621 grating The product was then collected on a coarse frit funnel and washed
spectrophotometer. as before until free of peroxide (negative test with starch-KI paper).
'H NMR spectra were run on a Varian A-60 spectrometer and 13C It was then washed with 1L of acetone in portions, sucked dry, and
spectra on a Bruker W13-90 (22.63 MHz) spectrometer. dried in a vacuum oven at 95-100 "C. The dry product was recrys-
Gel-permeation chromatography (GPC) was carried out on 0.125% tallized from xylene (20 mL/g) as previously described. Recovery was
solutions in 1,2,4-trichlorobenzene through three linear styragel essentially quantitative except for minor mechanical losses.
columns at 150 "C using National Bureau of Standards 1475 poly- Clemmensen Reduction at Atmospheric Pressure. One hundred
ethylene for calibration. In this way, weight average molecular weights grams of mossy zinc was amalgamated with 10 g of mercuric chloride,
(M,)and number average molecular weights (@,)_ere obtained and 150 mL of water, and 5 mL of concentrated hydrochloric acid by
from them the polydispersity indexes (PI = M,/M,). shaking for 5 min. The liquid was decanted and the zinc was washed
General Oxidation Procedures. Two setups were used which by decantation with two 100-mL portions of water.
differed mainly in the method of stirring, Le., either with magnetic The zinc, together with 75 mL of water, 175 mL of concentrated
stirring or mechanical paddle stirring. hydrochloric acid, 150 mL of xylene, and 5 g of keto acid was added
For magnetic stirring, a 2-L three-neck flask was fitted with an to a 1-L, three-neck flask equipped with a mechanical stirrer and an
Allihn condenser (30-cm jacket) whose upper outlet was vented into efficient reflux condenser. The mixture was vigorously stirred and
a cascading water trap by means of an adapter and Tygon tubing. The heated at reflux temperature for 24 h and then cooled to room tem-
flask was then equipped with an electric heating mantle and a mag- perature. It was diluted with 500 mL of water and the polymer-solvent
netic stirrer (1.5-in. bar coated with Teflon fluorocarbon resin). The slurry was decanted from the zinc residue. Any product adhering to
two remaining necks were glass stoppered, one being used for in- the zinc was dissolved by boiling briefly with 50 mL of xylene and the
tromission of a thermometer if necessary. hot xylene was decanted into the initial decantate which was then
For mechanical stirring, a similar setup was used but the center cooled to room temperature. The product was collected on a coarse
neck was equipped with an all-glass stirrer bearing a rod with a 10-cm frit funnel and washed with 1 L of acetone, sucked dry, and dried
paddle of Teflon fluorocarbon resin. The rod was attached to an further in a vacuum oven at 95-100 "C. It was dissolved in 200 mL of
electric motor adjusted to rotate the stirrer at a rate of 140-150 boiling xylene and the hot solution was suction filtered through a
rpm. preheated medium frit funnel to remove traces of zinc. The filtrate
The flask was charged with 900 mL of the appropriate acid and 200 was cooled to room temperature and the solid was collected, washed
g of polyethylene pellets, then stirring and heating were begun. with hexane, and again dried in a vacuum oven. About 3.5-4 g of
When concentrated nitric acid (71%) was used, the mantle tem- material was recovered.
perature was set to give a reaction mixture temperature of 110-115 Clemmensen Reduction under Pressure. The ingredients de-
"C. After about 15 min, brown fumes were evolved and refluxing scribed above were added to a l-L, glassed lined bomb and the bomb
began. The original mantle setting was maintained throughout the was sealed. It was then shaken and heated at 145 "C for 24 h. After
whole reaction time. With 90% fuming nitric acid, reflux began at a cooling, the mixture was worked up as described above.
nominal mixture temperature of 50-60 "C and the mantle setting Wolff-Kishner Reduction. A mixture of 5 g of keto acid, 5 g of
required to achieve this: was left constant. After about 24 h, the pot potassium tert-butoxide, 5 mL of anhydrous hydrazine, and 250 mL
temperature had risen to 80-90 "C and refluxing continued. Red of xylene was boiled under reflux in a nitrogen atmosphere for 18 h.
fuming nitric acid behaved very much like the 90% fuming acid. The mixture was cooled to room temperature and the product was
After the appropriate reaction time, the mixture was cooled to room collected on a coarse frit funnel. It was then washed with 500 mL of
temperature and poured into 2 L of crushed ice and was stirred oc- ethanol in portions and sucked dry. It was suspended in a mixture of
casionally until the ice imelted. 250 mL of xylene and 50 mL of 12 N sulfuric acid which was then
When short reaction times and/or magnetic stirring were used, the magnetically stirred and boiled gently for 15 min. After cooling, the
product was a mixture of dispersed fines and apparently intact pellets. product was again collected, washed with 500 mL of acetone in por-
These were separated by passing the aqueous suspension through a tions, sucked dry, and boiled with several 200-mL portions of water
14-mesh (ASTM) stainless-steel screen and the pellets were washed until the wash was neutral. After washing with acetone again and
free of fines by agitating:the pellets on the screen while washing with drying, it was recrystallized from xylene as previously described.
water. The fines and pellets were kept separate. The easily friable 2,4-DinitrophenylhydrazoneFormation. A mixture of 25 mL
pellets were dispersed with water in a high-speed blender and of xylene, 0.5 g of keto acid or reduced keto acid, and 0.1 g of 2,4-
thenceforth the fines and dispersed pellets were treated similarly. In dinitrophenylhydrazine was boiled under reflux for 2 h and cooled
some reactions the product was completely dispersed so the separation to room temperature. The product was collected and washed on the
was not necessary. funnel with acetone. The solid was suspended in 50 mL of acetone,
The suspended solid was collected on a coarse frit funnel and the mixture was boiled gently on a steam bath for 5 min, and the
redispersed in 2 L of water which was then heated to the boiling point product was collected again and sucked dry.
and filtered again. This was repeated several times until the filtrate Clemmensen Reduction of Nitro Carboxylic Grease. A 5-g
was neutral. The solid on the frit was sucked free of adherent water portion of the nitro carboxylic grease from preparation 6, Table I, was
and was dried to constant weight in a vacuum desiccator over phos- subjected to Clemmensen reduction at atmospheric pressure for 18
phorus pentoxide. Filtrates up to this point were discarded. h as described above for the high molecular weight products. The
The dried solid was boiled with three 1-L portions of acetone, the xylene layer was separated, washed with two 100-mL portions of
filtrates being combined and reserved for separate workup. water, and then dried over anhydrous magnesium sulfate. The xylene
The solid was sucked dry and recrystallized from xylene (20 mL/g). was distilled out under vacuum at 80-90 "C and the residual oil was
It was then collected, washed with hexane, and dried to constant freed of solvent by pumping at 0.1-0.3 Torr while heating on a steam
weight in a vacuum own at 95 "C (overnight) before titrating or bath. The crude wax was boiled with 100mL of hexane for 10 min then
otherwise manipulating. chilled briefly in ice to solidify the oil. The supernatant hexane was
The acetone filtrate was freed of bulk solvent on a rotary evaporator quickly poured off and chilled thoroughly to obtain 0.1 g of white solid,
and if any residual low molecular weight material remained, it was mp 73-77 "C. Neutral equivalent found 150; mol wt (freezing point
further freed of solvent. by heating at 80-90 "C while pumping at in benzene) 325; NMR (CDC13)6 1.28 (s, CH2, 15), 2.34 (t,J = 6 Hz,
0.1-0.3 Torr. a-CH2,2),10.9 (s, COOH, 1).The data correspond approximately to
Nitro Group Removal and Bleaching. A solution of 10 mL of the dicarboxylic acid HO~C(CHZ)I~COZH, mol w t calcd 328.5.
water and 500 mL of concentrated sulfuric acid in a 1-L Erlenmeyer Esterification of Nitro-Carboxylic Grease. A mixture of 8 g of
flask was heated to 80-8,5 "C while being stirred on a magnetic stirrer the grease from preparation 17,Table I, 100 mL of toluene, 10 mL of
hotplate. Fifty grams of finely divided, recrystallized nitroalkanoic methanol, and 0.5 g of p-toluenesulfonic acid was refluxed under a
acid was added in teaspoon portions allowing each to disperse before Dean-Stark head for 5 h. The solution was cooled and extracted with
adding the next. When all was added, stirring and heating were con- two 50-mL portions of saturated sodium bicarbonate and two 50-mL
tinued for 6 h. The gray-black mixture was then cooled to room portions of water. The organic layer was dried over anhydrous mag-
temperature and poured into 2 L of crushed ice. The product was nesium sulfate and freed of solvent by vacuum distillation on a steam
collected and washed by dispersing in several portions of boiling water bath. Residual solvent was removed by pumping at 0.14.3 Torr while
as described above. When the wash was neutral, the product was heating at 85-90 "C. Anal. Calcd for C23H42N208:C, 58.2; H, 8.9; N,
56 Mead, Porter, Reed Macromolecules

5.9; mol wt, 474.6. Found: C, 58.0; H, 9.1; N, 5.4; mol wt (bp in ben- Resins Department, E. I. du Pont de Nemours and Co., Wilmington,
zene), 470. Delaware. Alathon polyethylene resin is a registered trademark of the Du
The infrared spectrum showed a characteristic ester band at 1735 Pont Co. Alathon 7040 polyethylene resin hasii?, = 89 OOO, ii?, = 25 000
(by G P O .
cm-’ and additional bands at 1190,1160,1100,1010,and 850 cm-’ (13) Alathon 7050 polyethylene resin has ii?, = 68 000, ii?, = 11 000.
all absent in the staring material: NMR (CDC13)6 1.29 (s, CHz), 2.32 (14) R. B. Wagner and H. D. Zook, “Synthetic Organic Chemistry”, Wiley, New
( t , J = 6 Hz,a-CH2),3.69 (s,OCH3),4.50(s, HCN02), and no COOH York, N.Y., 1953, pp 418-419.
absorption at 6 10-11. (15) Pellets were recrystallized from xylene (50 mL/g) without any special care
to control cooling rate or optimize crystal growth. Boiling solutions were
Acknowledgments. I n addition to t h e acknowledgments merely allowed to cool spontaneously to room termperature.
in t h e footnotes, several others must be made. Thus, I thank (16) Corrected for instrument broadening.
(17) Marlex 6009.
Eleanor G. Applegate for her expert technical assistance. Dr. (18) Rigidex 2.
Raymond C. Ferguson of this laboratory contributed explor- (19) Private communication from Mr. C. R. Ginnard of this laboratory.
atory experiments which gave impetus t o this work. Discus- (20) For our estimations, we used the value 1.00 for the ratio of the extinction
coefficient of one COOH group to one NO2 group. For known mixtures
sions with Dr. William H. Sharkey were a valuable stimulus. of decanoic acid and 1-nitrododecane, the ratio varied between 0.82 and
Special thanks are due Mr. Charles R. Ginnard for t h e GPC 1.22 depending on whether the samples were run neat or in solution in
analyses,, and I also thank Naomi E. Schlichter and Ellen dichloromethane or mineral oil. The average was 1.03 which for practical
purposes we took to be unity. There is further uncertainty in this, however,
Wallace for determining t h e infrared spectra. Dr. J. Lazar did because although the ratio was linear with different proportions run neat,
t h e mass spectrum determinations, Dr. F. E. Weigert and Mr. they were significantly nonlinear in solution. Keller and Udagawa (ref 8)
Fred Davidson assisted with t h e 13C NMR, and Mrs. Flora used the ratio 0.71.
(21) Nitrogen elemental analyses are not useful in determining the degree of
Youngken did t h e ebullioscopic molecular weight determi- nitration because the low levels dealt with are within the error limit of
nations. combustion analysis.
(22) Percent nitration can be calculated from the relationship 200/(COOH/N02
References and Notes ratio) assuming only that there are two carboxyl groups per molecule.
(23) The solvent was acetone and the titrant was tetra(n-buty1)ammonium
(1) Contribution No. 2473; please include in reprint requests. hvdroxide. I thank Mr. Robert J. Berndt for the measurements.
(2) For leading references see G. N. Pate1 and A. Keller, J. Polym Sci , Polym. (24) P A . D. De Maine, M. M. De Maine, A. A. Briggs, and G. E. McAlonie,J
Phys. Ed., 13,2275 (1975). Mol Spectrosc , 4,398 (1960).
(3) D. J. Priest, J . Polym. Sci., Part A.2, 9,1777 (1971). (25) H. Feuer and A. T . Nielsen, J Am. Chem. Soc , 84,688 (1962). After 2
(4) A. Keller, E. Martuscelli, D. J. Priest, and Y. Udagawa, J. Polym. Scz , Part weeks in refluxing hydrochloric acid, 2-nitrooctane was only 65% converted
A-2, 9,1807 (1971). to 2-octanone.
(5) A. Keller and S. Sawada, Makromol. Chem., 74,190 (1964). (26) In the procedure of ref 10,0.3 g of nitro acid was heated with concentrated
(6) D. J. Blundell, A. Keller, and T . M. Connor, J . Pol.ym. Sci., Part A-2, 5 , sulfuric acid in a sealed tube at 70 “C for 20 h. In ref 9a, the treatment was
991 (1967). with 0.2 g a t 60 O C for 2 days.
(7) R. P. Palmer and A. J. Cobbold, Makromol. Chem., 74,174 (1964). (27) Keller and Udagawa (ref 9a) used activated charcoal in hot xylene solu-
(8) A. Keller and Y. Udagawa, J . Polym. Sci., Part A-2, 9,1793 (1971). tions, but this failed in our hands. Ballard and Dawkins (ref 10) used hot
(9) (a) A. Keller and Y. Udagawa, J . Polym. Sci., Part A-2, 8, 19 (1970); (b) concentrated nitric acid but in a t least one such trial we had evidence for
T . Williams, Y. Udagawa, A. Keller, and I. M. Ward, ibid., 8,35 (1970). the reintroduction of nitro groups.
(10) D. G. H. Ballard and J. V. Dawkins, Eur. Polym. J.,9,211 (1973). (28) “Organic Reactions”, Vol. I, Wiley, New York, N.Y., 1942, p 155.
(11) These preliminary experiments were done by Dr. R. C. Ferguson of this (29) M. G. Grundon, H. B. Henbest, and M. D. Scott, J . Chem. Soe., 1855
laboratory. (1963).
(12) The polymers were commercial resins in the form of extruded pellets about (30) Fisher Scientific Co., Fair Lawn, N.J.
4 mm X 7 mm in size. They were obtained from the Plastic Products and (31) J . T. Baker Chemical Co., Phillipsburg, N.J.

An ESR Study of Oxidation in Uniaxially Deformed

Polybutadiene and Polyisoprene
W. T. Mead, Roger S. Porter, and P. E. Reed*
Department of Materials, Queen Mary College, University of London,
London E l , England. Received J u n e 20,1977

ABSTRACT: The possible reactions of polybutadiene and other diene rubbers, with oxygen during initial processing,
low-temperature mechanical degradation, and subsequent handling, are reviewed. Radicals formed following tensile
testing of polybutadiene and polyisoprene at 83 K in predominantly nitrogen atmospheres, hut containing different
amounts of oxygen, are examined using electron spin resonance (ESR) techniques. It is concluded that two or more
radical species result from mechanical deformation at low temperature when oxygen is present. Radicals identifed
are a peroxy radical and most probably an allyl radical arising from main chain rupture. Peroxy radicals are shown
to arise mainly from combination of radical species with oxygen in the test environment, although a small proportion
is attributed to oxygen incorporated into the rubber network during processing. The relative stability with increasing
temperature of the peroxy to other radical species present is studied quantitatively.

I. Introduction nance (ESR) technique to monitor radical formation during

T h e effect of atmospheric oxygen on unsaturated rubbers mechanical deformation of diene rubbers in the glassy state,
at ambient temperature and above is pronounced and has close t o the liquefaction temperature of nitrogen, has also
been widely It is known t h a t chain rupture me- revealed oxygen-containing radicah4-8 The reactivity of
chanically induced by mastication produces radicals which oxygen with polymer radicals formed by mechanical degra-
are stabilized by reaction with oxygen.3 For natural rubber this dation has been recently reviewed by Sohma and Sakagu-
leads to a reduction in molecular weight into t h e processable ~ h i , while
~ J ~ DeVries” has reviewed radical formation in
range. More recently, the application of electron spin reso- rubbers. T h e presence of such oxygen-containing radicals is
0024-9297/78/2211-0056$01.00/0 0 1978 American Chemical Society