Oxidation of General-Purpose Polyethylene Resin


Oxidation of General-Purpose Polyethylene Resinhttps://pubs.acs.org/doi/pdf/10.1021/ie50509a047Similarby CS Myers - ‎1...

16 downloads 48 Views 506KB Size

INDUSTRIAL AND ENGINEERING CHEMISTRY

May 1952

(3) Clark, W.L.,and Leonard, J. M., Naval Research Laboratory, Rept. C-3114 (July 1947). (4) Frager, M., Philadelphia, Frankford Arsenal, R e p t . R-537 (July 1944). (5) Galloway, L. D., J. Testile Inst., 26, T123-9 (1935). (6) Hutchinson, W. G., Scientific Monthly, 43, 165-77 (1946). (7) Luckiesh, M.,“Applicationa of Germicidal, Erythemal, and Infrared Energy,” pp. 123-6, New York, D. Van Nostrand Co., Inc., 1946. ( 8 ) Ibid., pp. 250-3. (9) Magee, C. J., Hansen, C. T., and Grant, C. K., “Report on the Condition of Service Materiel under Tropical Conditions in

New Guinea,” Scientific Liaison Bureau, Melbourne,Australia, October 1943. (10)Sational Defense Research Committee, “Tropical Deterioration of Equipment and Materials,” Vol. 1, Summary Technical Rept. of Tropical Deterioration Administrative Committee, 1946.

1095

(11) Office of Scientific Research and Development, Washington, D. C., OSRD Rept. 1833 (July 1943). (12)Ibid., 3803 (June 25, 1944). (13)Ibid., 3952 (October 1945). (14) Ibid., 4118 (September 1944). (15) Ibid., 4371 (December 1944). (16) Ibid., 5684 (Oct. 31, 1945). (17) Ibid., 5767 (April 1945). (18)Ibid., 6055 (Oct. 15, 1945).

(19) Richards, 0.W., J. Bact., 58,453-5 (1949). (20) Scheffer, T. C.,and Duncan, C. G., IND.END.CHEM.,38,619-21 (1946). (21) Turner, J. S.,McLennan, E. I., Rogers, J. S., and Matthaei, E., Nature, 158, 469-72 (1946). (22) Vicklund, R. E., IND. ENG.CHEM., 38, 774-9 (1946). RECEIVEDE or review July 28, 1951.

ACCEPTED December 13, 1951.

Oxidation of General-Purpose Polvethvlene Resin J

J

C . S . MYERS Development Laboratories, Bakelite Co., Division of Union Carbide and Carbon Corp., Bound Brook, N . J .

0

XIDATIVE degradation of high polymers under the influence of light and heat is well known. Polyethylene resin, no exception in this respect, is a good subject for fundamental studies of oxidative degradation phenomena because of relative simplicity of its chemical structure. Considerable data have been published (6, 9,11, 14) on changes in electrical loss characteristics and alteration in chemical structure of polyethylene resins imparted by exposure to light and heat. Increases in carbonyl oxygen concentration and, to a lesser extent, increases in other oxygenated groups are the major chemical effects previously reported. This paper presents quantitative data on dielectric loss properties and corresponding chemical effects caused by thermal oxidation which permit a definition of the kinetics of the oxidation reaction. TIIEORETICA L CON SIDERATION S

Recent chemical researches ( 1 7 ) in the field of hydrocarbon oxidations have demonstrated t h a t hydroperoxides ( R O O H ) appear as one of the first products of reaction. Kinetic studies have shown that subsequent oxidation of the parent hydrocarbon is autocatalyzed by the decomposition of hydroperoxides, which produces free-radical initiators for the chain reaction. The decomposition of hydroperoxides also leads t o the formation of secondary oxidation products, such as ketones, aldehydes, acids, alcohols, water, and carbon dioxide. One of the simplest and best delineated autoxidation mechanisms is the one proposed by Bolland and Gee ( 1 ) for the oxidation of ethyl linoleate, C1,HalCOOCzHs. At sufficiently high pressures of oxygen and in the range where a sufficiently high concentration of hydroperoxide has been built up so t h a t hydroperoxide decomposition is the important chain initiation step, they proposed the mechanism ki BROOH-R,

+ ROO. + other decomposition product Oz

ROO.

+ R. + RH

kz +ROO. k3 +ROOH

(I)

(2)

+ Re

(3)

ROO.

lie + ROO ---+ROOX + Oz

(4)

where R H represents ethyl linoleate nnd the H in question is the one t h a t detaches most easily, because of the activating influence of the double bonds. It is commonly accepted t h a t antioxidants, although stable in atmospheric oxygen, operate by interrupting the above chain reaction whereby the inhibitor is oxidized t o stable compounds. Some theories on the mechanism of antioxidant action postulate t h a t the stabilizer is reformed, to some extent, during the series of reactions. However, experimental measurements show that the antioxidant is destroyed or becomes ineffective, ultimately. I n the absence of inhibitors, Equations 1t o 4,inclusive, demonstrate the formation and destruction of free radicals a t steady reaction rates, IC1, k,, 123, etc. Stable secondary oxidation products created by the decomposition of hydroperoxides result in a n increase in concentration of ketones, aldehydes, acids, alcohols, water, etc. Designating RO as stable oxygenated hydrocarbon, the secondary reaction is assumed schematically as kr ROOH+RO

+ H20 + other decomposition products

(5)

Since ROOH is being formed. continuously, increase in concentration of R O is expected t o be defined by dt

=

h7(RO)

(6)

Integration of Equation 6 and conversion t o the common log form gives log ( R O ) =

-+ B

k7t 2.303

(7)

Therefore, concentration of stable combined oxygen, (0), is exponential with time, t , and limited by the concentration of ROOH present which is decomposing at a rate, k,. On the basis of this hypothesis, determination of concentration of polar oxygen groups in polyethylene as a function of milling time and temperature should supply suitable data for defining the kinetics of the oxidation reaction. Power factor measurements at high fre-

I N D U S T R I A L A N D E N G I N E E R I N G CHEMISTRY

1096

quencies are convenient and commonly employed in aasessing the presence of polar oxygen groups in degraded polyethylene. However, no relations appear to be in existence for correlating power factor values with combined oxygen concentration independently established by other methods of analysis. This correlation was established, as described in the following, by the use of the infrared spectrophotometer shown t o be applicable b y Jackson (@, Thompson (16),Maibauer (9),and Prow (14), and their coworkers. EXPERIMENTAL RESULT§

Unstabilized DYNH resin was milled on a n 8 X 12 inch, %roll differential speed mill at various roll surface temperatures. Power factor measurements a t 50 Mc. versus milling time are shown in Figure 1. The electrical loss measurements were made

Vol. 44, No. 5

The relation of carbonyl oxygen concentration to 50-Mc. power factor, independent of milling time and temperature, is shown in Figure 4 along with corresponding mole ratios of O/CH2 equivalent to a given carbonyl oxygen concentration. Linearity of the plot in Figure 4 implies that 50-Mc. power factor is closely associated with carbonyl oxygen concentration and increases exponentially with time. Infrared absorption characteristic of ketonic oxygen was quite pronounced a t 5.81 microns and no aldehyde oxygen and carboxyl oxygen absorption was noted in the spectra No evidence of hydroxyl oxygen vias apparent in the spectra of the degraded specimens. Normally, polyethylene resin compounds employed in highfrequency dielectric applications contain aa antioxidant to prevent oxidative degradation during thermal processing. Figure 5 demonstrates the protective action of a commercial stabilize1, Akroflex C (65% phenyl-1-naphthylamine and 35% diphenyl-pphenylenediamine), on DYNH resin during milling in air a t a roll surface temperature of 150' C. It may be noted that 0.01% stabilizer is effective for a milling period of approximately 30 minutes. The effective life of higher concentrations of Akroflex C indicated is in excess of 60 minutes. Milling of the compound consisting of 99.8y0D Y N H resin and 0.2% Akroflex C, as noted, was extended to 7 hours at 150 O C. with no appreciable increase in 50-Mc. power factor. The unstabilized control, milled for the same period, increased in power factor from an original value of 30 X 1 0 6 to an ultimate value of 1225 X IOb5.

MILLING TIME, MINUTES

Figure 1. Degradation Rate us. Milling Temperature by the technique described by Wangsgard and Hazen ( 1 8 )using a Type 160-a &-meter. Slopes of the linear plots in Figure 1 were calculated and expressed as initial and minimum rates of power factor increase per minute versus milling temperature, which is appropriately described by the classical Arrhenius relation (8) defining chemical reaction rate and temperature - E k = 10'e --

RT

where k is reaction rate, 10' the frequency constant, e the natural logarithm base, E the energy of activation of the reaction, R the gas constant, and T the absolute temperature. A plot of log A power factor x 106 min.+ as log k in the logarithmic form of Equation 8 log k

=

- E 2.303 RT

_ _ I -

+

D

versus reciprocal of absolute temperature is shown in Figure 2. Activation energy, E, of the reaction was calculated as 15,960 calories mol.-'. This is based on initial and minimum values of reaction rates. A survey of infrared absorption spectra of the milled specimens indicated t h a t carbonyl oxygen was the predominant oxygenat& group present. Accordingly, carbonyl oxygen concentration was determined a t a wave length of 5.81 microns in films pressed from unstabilized DYNH, milled for periods up to 120 minutes a t 110 and 150" C. Absorption coefficients from spectra of methyl undecyl ketone in heptane were employed as standards. Plots of carbonyl oxygen concentration and 50-Mc. power factor values, versus milling time at 150" C., are shown in Figure 3. The relation of carbonyl oxygen concentration and lime, predicted by Equation 7, is confirmed by Figure 3.

Figure 2.

Oxidation of Polyethylene Resin u1SCTiS SlON

The rneulianism of oxidation of polyethylene is expected to be similar to that of paraffin lvax (10) and other saturated hydrocarbons, The specific points revealed in this study are quite consistent with many of the principles described by Stossel (16) in his review of the literature of the oxidation of hydrocarbons, and those presented in more recent publications ( 1 7 ) on the subject. One definite fact revealed by the many previous investigations in this field is the intermediate formation of unstable peroxides which decompose to oxygenated hydrocarbons. The parameter of power factor in Figure 3 is expected t o reflect electrical loss effects of intermediate polar peroxides in the case of polyethylene resins, as suggested by Jackson (6). However, peroxides and hydroperoxides, which are of limited stability a t the elevated temperatures of milling, are low in concentration a t any instant as shown by Equations 1 to 4, inclusive. Accordingly, the major

May 1952

INDUSTRIAL AND ENGINEERING CHEMISTRY

1097

TABLEI. DEGRADATION KINETICSO F SOLIDHYDROCARBONS (k

lOCe

- m) E

Natural Rubber Vuloanizate

Polyethylene Grade 2

Oxidation in air

Photocatalyzed oxidation

Depolymerized in vacuo

110-140

75-120

375-436

Fisher-Tropsch

Pol ethylene S;YNH 21,000 Oxidation in air 100-160

Property Relative av, mol. wt. Process Temperature, C.

-

Wax

...

350

00.110 1018.8e-T

16.000

21.100

l6,BBO

23,000

104.8e-RT

Rate constant, k

106.0'e-RT

Gm. carbonyl oxygen (100gm. resin) -1 aec. -1

Gm. oxygen (100 gm. wax) -1 am. - 1

Gm. oxygen (I00 gm. rubber)

Activation energy, E,cal. mol. -1 Average temperature coeffioient of reactionrate, k((t + 1 0 q / k t

15,960

21,700

16,000

Gm. produot (100 gm. resin) 66,110

1.74

2.0

1.87

2.0

Reference

This work

1 0 9 4 ~ - ~

-1

aec. -!

sec. - 1

Jellinok (7)

Tobolsky (17)

Pardun (13)

-1

between concentration of carbon monoxide added during polymerization and power factor of the resultant polymer. The expression illustrated by Equation 8 is equivalent in principle to that defined by Pardun (1.9)in describing oxidation rates of Fisher-Tropsch synthetic hydrocarbon wax-a homolog of polyethylene resin. Oxidation rates reported by Pardun were baaed on liters of air per gram of substrate required to produce maximum yields of organic acids over a temperature range from. 160

Figure 3.

Oxidation of Polyethylene Resin

140

electrical loss effects observed are attributed to stable oxygenated structures resulting from the autocatalytic decomposition of hydroperoxides (and peroxides) in accordance with Equation 5. Extreme sensitivity of 50-Mc. power factor measurements to polar oxygen groups is demonstrated by Figure 4. For example, an increase in power factor from an initial value of 35 X 10" to 350 X 10-6 for the oxidized resin corresponds to an increase in concentration of carbonyl oxygen of 0.05% by weight or 1 oxygen

%H, k 0 0

I

0 "3800

h400

0001

GOO2

0003 0005 0008

009

CARBONYL OXYGEN, WT X

004 006

01

02

Figure 4. Oxidation of Polyethylene Resin atom per 2280 methylene groups. Qualitatively, the observations are quite consistent with Pross and Black's data (14) on polyethylene resin oxidized by photocatalysis. Infrared absorption spectra of polyethylene resins involved in studies by these investigators were marked by pronounced carbonyl absorption and the presence of bonded hydroxyl and some carboxyl groups after irradiation. I n addition, a linear correlation of oxygen uptake and power factor after similar irradiation periods were found by them which is analogous to the relation shown in Figure 4. Jackson and Forsyth (4, 6) found a similar linear relation

I

10

I

20

I

30

I

40

I 50

I

60

MILLING TIME AT I S 0 DEG. C , MINUTES

Figure 5. Milling Stability us. Stabilizer Concentration

110"to 140' C. Studies by Tobolsky et al. ( 1 7 )on autoxidation ,of vulcanized natural rubber should not be ignored, because the rates of oxygen absorption by the vulcanizate during concurrent thermal and light irradiation could also be defined by Equation 8. Their measurements demonstrated the powerful effect of light a8 an oxidation catalyst which was more effective a t lower ambient temperature of irradiation. Equation 8 also applies to pyrolysis of polyethylene in a vacuum aa studied by Jellinck (7). This investigator found that a polyethylene msin similar in average molecular weight to DYNH resin was more resistant to depolymerization in a vacuum than polystyrene-an observation in accord with findings of Madorsky (8),who reported polyethylene to be more resistant to vacuum pyrolysis than polyisoprene, polyisobutylene, polystyrene, GR-S rubber, and polybutadiene over a temperature range of 300" to 475' C. Oakes and Richards (12) also considered thermal degradation of ethylene polymers in the absence of air a t temperatures above 300 ' C. where depolymerization begins. Whereas

1098

INDUSTRIAL AND ENGINEERING CHEMISTRY

rate constants were not presented, their data are of significant interest and demonstrated increased resistance of polyethylene to scission-type reactions in the absence of oxygen. An evaluation of initial and minimum rate constants in Equation 8 for D Y N H involved in this work and for constants calculated from data reported by Pardun (IS), Jellinck (79,and Tobolsky ( 1 7 ) is shown in Table I. Obviously, reaction rates, k , cannot be compared in Table I because of differences in units and differences in conditions of the degradative processes applied. However, activation energies, E, and temperature coefficients of reaction rate, k‘+loc./kt, derived from ratios of reaction rates, are independent of units and can be compared. Temperature coefficients of reaction rate and activation energies of oxidation of polyethylene DYNH, Fisher-Tropsch wax, and vulcanized natural rubber are roughly comparable. Also, activation energies and frequency factors found in these specific cases are in the same order of magnitude of the values (23,000 cal. mol.-1 and 105) reported by Eyring et al. ( 2 ) for oxygen-catalyzed isomerization and polymerization of substituted ethylenes. On the other hand, polyethylene resin depolymerized in the absence of oxygen at much higher temperatures displays a higher energy of activation of scission-type degradation which is characteristic of increased stability. ACKNOWLEDGMENT

Aid extended by I,. H. Wartman and F. M. Rugg in making the reported infrared spectrophotometric measurements, and by G. A. Crowe and other associates of the author is gratefully acknowledged.

Vol. 44, No, 5

LITERATURE CLTED

Bolland, J. L., and Gee, G., Trans. Faradag Soc., 42, 2 3 w 4 (1946). Eyring, H., Hulburt, €1. AI., and Harman, R. A . , ZND. ENQ, C F I E M . , 35,511-21 ( h k y 1943). Getman and Daniels “Outlinea of Physical Chemistry,” 7th ed., p. 363, New York, John Wiley & Sons, Inc., 1943. Jackson, W., and Forsyth, J. S. A., J. I n s t . Chem. EWTS.,Pt. 111, 91, 23 (1949). Jackson, TV., and Forsyth, J. S. A, J . Inat. EEec. Eiigrs. (London), Pt. I l l , 91 (June 1944). Ibid., 94, 55 (1947). Jellinck, H. H. G . , J . Polgmer Sci.,4, 1-36 (1949). Madorsky, S. L., Straus, S., Thompson, D., and Williams, L., Ibid., 4, 639-64 (1949). hlaibauer, A. E., and Myers, C. S.,TTans. Electrocha. SOC.,90. 449-67 (1946). Morawete, H., IND.ENG.CHEM.,41, 144247 (July 1949). Myers, C. S., and Rfaibauer, A. E., Elec. Eng., 64, 916-18 (December 1945j. Oakes, W. G., and Richards, R. G., J. Chem. Soc., 1949,2929-35. Pardun, H., Fette u. Seifen, 48,397-403 (1941);49, 441-6 (1942). Pross, A. W., and Black, R. A I . , J . Sop. Chem. I d . (London),69, 113-6 (April 1950). Stossel, E., Oil Gas J., 43, 130-9 (July 21, 1945), 145-51 (August 18, 1945), 69-74 (September 1945). Thompson, H. W., and Turkington, P., Trans. Faraday Soc., 41, 246-60 (1945). Tobolsky, A. V., Mete, D. J., and Mesrobian, R. B., J . Am. Chem. Soc., 72, 1942-52 (May 1950). Wangsgard, A. P., and Raeen, T.,Trans. Elactrochem. Swc., 901, 177-91 (1946). RECEIVED for review May 31, 1951.

ACCEPTEDSorexukr 13, 1961,

Alcohol as an Antiknock Agent in Automotive Engines JAMES C. PORTER AND RICHARD WIEBE Northern Regional Research Laboratory, Peoria, Ill.

A

LCOHOL-water injection in automotive engines for the purpose of improving the octane rating of gasoline a t high engine loads has received considerable attention in recent years (9, I S , 17). I n military aircraft alcohol-water injection was used extensively during the last war both for take-off and in combat operations (10). h very complete set of references will be found in “The Technical Literature of Agricultural Motor Fuels” ( 1 6 ) and subsequent papers of this laboratory ( 7 , 8, 1 7 ) . It is principally the high octane number of alcohol which determines the high antidetonant quality but the cooling effect of the high heats of vaporization of alcohol and of water are import a n t contributing factors. Table I shows the physical properties of the three lower alcohols, of water, and of a 50-50 mixture by volume of ethyl alcohol and water. If the heat of vaporization of gasoline is assumed t o be approximately 130 B.t.u. per pound, the heat of vaporization of a 50-50 mixture by volume of ethyl alcohol-water is almost six times as large. An additional cooling factor is the lower maximum combustion temperature when alcohol is used, as shown in Figure 1. The maximum temperatures, as well as thermal efficiency, were calculated from the thermodynamic data reported by Hershey, Eberhardt, and Hottel (6) and Hottel, Williams, and Satterfield ( 6 ) for iso-octane and Wiebe, Schulta, and Porter ( 1 8 ) for ethyl alcohol. See Table I of (18) for data on gasoline and alcohol at a compression ratio of 6 t o 1. The 25’% alcohol-gasoline blend data were obtained by interpolation. Results a t the other compression ratios up t o

12 to 1 were calculated in a similar manner. An initial temperature of 80” F. (540’ R.) and complete initial vaporization of the fuels were assumed. LABORATORY AND ROAD OCTANE NUMBERS W I T H ALCOHOL

I n order to obtain a better understanding of the effect of alcohol on hydrocarbon fuels, four base stocksstraight-run, catalytically cracked, thermally cracked, and polymer gasoline-were selected which were similar to the ones used by Bogen and Nichols ( 1 ) in their work “Calculating the Performance of Motor Fuel Blends.” The results for the four base stocks and their mixtures, given in Tables I1 and 111, show an excellent response to additions of ethyl alcohol except in the case of the straight polymer. Here the octane values approach the rating of pure ethyl alcohol (iso-octane and 0.33 ml. of tetraethyllead, for the Research and 92.0 for the Motor Method), and little or no further gain can be expected. As indicated in the tables, the octane number determinations were made on blends; however, the engine will not react differently if the alcohol is introduced by injection rather than as a blend. The actual potential performance gain is not given on the octane number scale but on the so-called performance number scale (4), to which further reference will be made. On this more realistic scale a n octane unit in the upper range assumes a much greater importance than one in the lower range. I n general, the Research Method octane number is characteris-