The Polymerization of Benzene in a Radiofrequency Discharge

29 The Polymerization of Benzene in a Radiofrequency Discharge D A V I D D. N E I S W E N D E R

Downloaded by CORNELL UNIV on July 23, 2016 | http://pubs.acs.org Publication Date: June 1, 1969 | doi: 10.1021/ba-1969-0080.ch029

Central Research Division Laboratory, Mobil Research and Development Corporation, Princeton, N . J.

Benzene vapor has been converted to polymeric materials in a radiofrequency discharge. Depending on conditions employed, the reaction yielded either total conversion to a solid polymer or a low conversion to a liquid polymer and diphenyl. High energy dose and/or low benzene partial pressure favored the former while low dose and high partial pressure favored the latter. The solid is a very intractable material with interesting properties. Infrared, ultraviolet and nuclear magnetic resonance data suggest that both polymer products are similar to polystyrene in structure, the liquid being low molecular weight; the solid being high molecular weight and highly cross-linked. A synthetic route from benzene to polystyrene type polymers is suggested. The effect of several reaction variables is discussed.

' T ^ h e reactions of benzene in various types of electrical discharges have been studied by a number of researchers over a span of nearly 70

A

years. Reported results vary widely. Several early workers (3, 4, 5, 9), using ozonizer tubes, obtained a gummy wax-like substance, along with hydrogen, acetylene, and other light hydrocarbon gases.

One of these

researchers later obtained a liquid and a solid, both analyzing as compounds (10).

In 1930, Austin and Black (1)

C24H05

found diphenyl and a

solid which they suspected to be a polyphenylene containing phenol groups.

Harkins and Gans (6),

using an electrodeless discharge,

got

complete conversion to a red-brown insoluble solid analyzing for ( C H ) . X

Davis (2),

on the other hand, obtained diphenyl, p-terphenyl, a resin

( 0 Η ) ^ hydrogen, acetylene, ethylene, and light paraffins. A 1935 U . S. 6

4

patent (8)

describes a discharge apparatus for preparing diphenyl from

benzene. More recently, Streitwieser and W a r d (13, 14) obtained a 5% 338 Blaustein; Chemical Reactions in Electrical Discharges Advances in Chemistry; American Chemical Society: Washington, DC, 1969.

29.

Polymerization of Benzene

NEiswENDER

339

conversion of benzene in microwave discharge, the products being low molecular weight gases, toluene, ethylbenzene and phenylacetylene. Stille and co-workers

(12),

using a radiofrequency discharge, got a

10%

conversion to poly(p-phenylenes), diphenyl, fulvene, acetylene, aliène, and methylacetylene. Vastola and Wightman (15)

obtained a solid film

and concluded from the infrared spectrum of the film that no aromaticity remained in the polymer. Jesch et al. (7),

on the other hand, also ob-

tained a solid film, but interpreted its infrared spectrum as suggesting the presence of aromatic groups as well as olefinic and acetylenic unsaturation. The wide disparity of the results certainly indicates there is still much to be learned about the chemistry of benzene in electrical discharges.


This disparity is probably largely because of widely varying reaction conditions.

Especially important are considerations such as the power

dissipated in the discharge, the pressure, etc. The work described here is an attempt to systematize the study of benzene reactions

in radiofrequency discharges.

The only products

isolated in these reactions were diphenyl, a liquid polymer and a solid polymer. The polymers appear to be polystyrenes.

Experimental Apparatus and Procedure. The apparatus consisted of a 3.69 M H z . radiofrequency generator capacitively coupled to a cylindrical borosilicate glass flow reactor by means of two external copper electrodes. A n inductively coupled tank circuit was used to match the impedances of the generator and the reactor. A n approximate measure of the power dissipated in the discharge was determined by measuring the voltage and current during the experiments and making a phase correction for the capacitive component of the reactor current. A simple schematic of the apparatus is shown in Figure 1. The hydrocarbon or helium-hydrocarbon mixture was metered into the reactor at various rates. Reactor pressures from 1-20 mm. were employed. In most cases, the discharge established itself as soon as the R F . signal was applied. If it did not, it was triggered with a Tesla coil. Products were collected in a series of traps, one at room temperature, one at - 7 8 ° C . and one at - 1 9 5 ° C . Chemicals. Reagent grade, thiophene free benzene (Baker), and vacuum distilled styrene (Matheson, Coleman and Bell) were used in the discharge experiments. The helium (Matheson) had a reported minimum purity of 99.995%. The acetylene (Matheson) was at least 99.6% pure.

Results Depending on the conditions employed, the reaction of benzene in the R F . discharge resulted in either a complete conversion to a solid polymer or a lower conversion to a liquid polymer and diphenyl.

Blaustein; Chemical Reactions in Electrical Discharges Advances in Chemistry; American Chemical Society: Washington, DC, 1969.

The

340

C H E M I C A L REACTIONS IN E L E C T R I C A L

DISCHARGES

solid tends to form under conditions of high power dissipation and/or low partial pressures of benzene in the reactor. Conversely, the liquid polymer and diphenyl result under low power dissipation and/or high partial pressures of benzene.

J)


5 0 0 watt radio frequency generator

iïïïïil nnnnnn tuning circuit

borosilicate glass reactor

Figure 1. The

Radiofrequency discharge reactor

solid polymer, which can sometimes be observed leaving the

discharge zone as a fine smoke, deposits throughout the trap system, but principally in the dry ice trap. W h e n collected, it is a very light, fluffy, nearly white powder which picks up a considerable static charge upon handling. Some of the substance's physical and chemical properties are as follows: 1. It is completely insoluble in water and in all organic solvents which were tested. 2. It does not melt up to 435 ° C . and shows no thermosetting or thermoplastic properties at even higher temperatures. 3. Thermogravimetric analysis shows that the polymer undergoes a stepwise loss in weight, with the loss being complete at 5 8 0 ° C . 4. X-ray diffraction studies show it to be completely amorphous. 5. The density of a pressed pellet is 1.10 gram/cc. 6. A freshly prepared sample was found to have an electron spin density of 2 Χ 10 spins/cc. 17


29.

NEiswENDER


341

7. Its surface area (nitrogen adsorption) is 42 meter /gram, a rather high value for an organic substance. 2

8. It chemisorbs oxygen from the air at room temperature, adsorption continuing for extended periods of time.

the

9. The experimentally determined carbon-hydrogen ratios varied from 1.02 to 1.05 for the many polymer samples studied. Thus, the elemental composition is the same as in the starting benzene (ignoring the chemisorbed oxygen in some samples ). In those experiments which did not yield the solid polymer, the conversion of the benzene was of the order of 30%. benzene was converted to diphenyl; the other 25%

About 5%

of the

was converted to a

liquid polymer with an average molecular weight of 617.

This polymer,


a viscous amber liquid, was not characterized as completely as the solid, but infrared spectral data indicate it to be very similar structurally to the solid. Diphenyl is a commonly reported product of the reaction of benzene in electrical discharges and will not be considered further in this paper. Most authorities suggest that it arises via combination of phenyl free radicals.

Discussion Polymer Structure.

The physical properties of the solid

suggest

that it is a high molecular weight, highly cross-linked polymer with an irregular structure.

The extreme insolubility precludes spectral studies

requiring solutions. It was possible to obtain an infrared spectrum by preparing a K B r pellet containing 2% of the solid. The liquid polymer, which was soluble in organic solvents had an infrared spectrum very similar to the solid. Of several structural possibilities considered, the one which agrees best with the infrared data and seems the most likely from a chemical viewpoint is a polystyrene type structure.

In Figure 2, the infrared spec

trum of a reference polystyrene film ( A ) is compared with the spectra of the solid ( B ) , the liquid ( C ) , and a solid polymer obtained when a styrene-helium mixture was passed through the discharge ( D ). A n equimolar mixture of benzene and acetylene in helium also produced a solid polymer whose infrared spectrum was indistinguishable from spectra Β and D . A l l the spectra are similar, the only significant differences being the O H absorptions in the 3300-3400 cm."

1

range and the carbonyl

absorptions at about 1700 cm." . These bands in all of the spectra from 1

the discharge-derived polymers are caused by rapid oxidation by molecu lar oxygen. These bands become quite pronounced if the polymers are allowed to stand in air for a few hours.


342

CHEMICAL

REACTIONS

IN ELECTRICAL

CM"


3000

2000

1600 1200

DISCHARGES

1

900

MICRONS

Figure 2. A= Β= C= D=

Infrared spectra

Polystyrene film Solid polymer from benzene Liquid polymer from benzene Solid polymer from styrene

It is believed that the principal difference between the solid and liquid products is that the liquid is essentially a linear polymer, while the solid is highly cross-linked. Such a highly cross-linked polystyrene would be expected to have a lower ratio of aromatic C - H to aliphatic C - H bonds than would a linear polymer. Comparison of spectrum Β or D with C in Figure 2 shows that the intensity of the aromatic and aliphatic C—Η stretching vibrations agrees with this expectation. Unfortunately, the N M R spectrum of the solid polymer could not be obtained owing to its extreme insolubility.

However, because

the

solid and liquid seem to be structurally similar, the N M R spectrum of the later was examined and compared with that of an authentic poly styrene sample (see

Figure 3).

As is generally true with polymeric


29.

NEiswENDER


343

materials, the resolution was quite poor and only broad bands were observed.

The polystyrene (10%

solution in C C 1 ) spectrum simply 4

showed two broad peaks—one centered at δ = = = = = 1.50 owing to aliphatic protons and one at δ == 7.08 (with a small companion peak at δ == 6.58) owing to aromatic protons.

The spectrum of the liquid polymer

solution in C C 1 ) was quite similar to the peaks appearing at δ = 4

(10% 1.60

and 7.04 (no peak at δ == 6.58). In addition, a very small peak at δ = 5.70 was observed. This is probably because of protons on olefinic double bonds, and suggests either that some unsaturation is present in the poly mer backbone, or that some unpolymerized vinyl groups are present. A n attempt was made to increase the resolution of the N M R spectra by


using a time averaging computer on very dilute solutions of the polymers, but the resolution was unchanged.

Figure 3.

NMR spectra (60MC)

A = Polystyrene (10% solution in CC10 Β = Liquid polymer (10% solution in CCI,,) Since the insolubility of the solid polymer precluded any ultraviolet studies in solution, attempts were made to obtain a diffuse reflectance spectrum of the solid.

Meaningful data could not be obtained. Again,

however, the liquid polymer could be compared with authentic poly styrene.

Figure 4 shows the ultraviolet spectra of the two materials.

Although the resolution is poor for the liquid polymer, its spectrum generally agrees with that of polystyrene. Thus, although they are of limited value, the N M R and ultraviolet data do support a polystyrene type structure and, when coupled with the infrared data, strongly suggest that the polymers are both polystyrenes.



344


320

300

280

260

240

DISCHARGES

200

Wavelength (millimicrons)

Figure 4.

Ultraviolet spectra

: Polystyrene in cyclohexane = Liquid polymer in cyclohexane

Mode of Formation of Polymer.

The most likely route from ben

zene to polystyrene involves several steps, the first of which is the estab lishment of an equilibrium between benzene and acetylene:

ο

3 HC^CH

This interconversion has been observed in many high energy systems including electrical discharges.

In the next step, the benzene and acetyl

ene, one or both of which may be in a reactive state, combine to give styrene which then polymerizes :

The simple linear polymer thus formed, limited to small chains such as pentamers, hexamers, etc., would explain the liquid polymer product.


29.

NEiswENDER

The


345

formation of the highly cross-linked solid polymer can be ex-

plained by postulating the formation of benzenes substituted with more than one vinyl group which, when polymerized, would yield an extensive,


irregular structure.

Both ionic and free radical mechanisms can be written to explain the foregoing reactions in more detail, but these would be strictly speculative, since no definitive experimental evidence has been obtained. Some sort of benzene ion must form as a result of inelastic collisions between the electrons and benzene molecules in the plasma, but whether these ions or some derivative species are the reactive intermediates is not known. Recently, Potter et al. (11)

have shown that styrene, when perfectly dry,

does polymerize via an ionic mechanism when irradiated with gamma rays.

One can picture a similar mechanism occurring in the electrical

discharge. On

the other hand, the fact that the polymer has a high electron

spin density suggests free radical involvement. However, it can be questioned whether the unpaired electrons arose during or after the polymerization reaction. A n attempt was made to induce a spin signal in a finely divided polystyrene sample by passing the solid through a helium discharge. No signal was detected after this treatment. The

question of mechanism must await the results of more funda-

mental studies of the phenomena occurring within the discharge. Effect of Reaction Variables. Correlations between the types of product obtained and several reaction variables were studied briefly, and are discussed in the following paragraphs. A.

ENERGY DOSE.

A n obvious variable to examine for a correlation

is the amount of energy absorbed per mole of benzene.

This quantity,


346


DISCHARGES

calculated from experimental data for more than fifty experiments, was obtained by using the equation,

where Ε =

energy dose (joules/mole or watt seconds/mole)

F === power (watts) Τ =

residence time in the reaction zone ( seconds )

Β == quantity of benzene treated in time Τ ( moles )


Table I summarizes the data obtained from these calculations. Table I.

Effect of Energy Dose

Conversion and Product

Lowest Dose

(joules/mole)

Highest Dose

Complete conversion to solid polymer

8.9 X 10

Low conversion to liquid polymer and diphenyl

1.5 X 10°

e

Average Dose

5.2 Χ 10

7

1.9 Χ 10

1.3 Χ 10

7

7.0 Χ 10

7

6

Although there is an overlapping dose range (8.9 Χ 10 -1.3 Χ 10 ) 6

7

where either type product can be obtained, there was an indication that high doses favored solid formation and low doses favored liquid.

This

seemed reasonable since more energy would be required to furnish the additional acetylene necessary for cross-linking and to gain the complete conversion. This suggested correlation was tested by performing several experi ments under conditions that had always produced solid (10% benzene in helium; 10 mm. pressure; 2.6 second residence time), except the energy dose was kept low by detuning the tank circuit used to couple the R F . generator to the reactor. Doses as low as 4.9 Χ 10 joules/mole produced 6

the solid polymer, but the conversions dropped from 100% to about 30%. Therefore, while the energy dose appears to correlate well with the con version obtained, it alone does not seem to determine the type product. Since the energy dose and the dissipated power are directly pro portional, a similar correlation arises when the variation of products with power is studied. B.

RESIDENCE T I M E .

Very limited data are available to study the

effect of residence time because nearly all the experiments were per formed using 2.6-3.0 second times. Table I I shows two experiments using pure benzene vapor at 2 mm. pressure.


29.


NEiswENDER

Table II. Ε

Effect of Residence Time

Ρ

Τ

(joules/mole) (watts) 4.3 Χ 10 1.2 Χ 10

Results

(seconds)

63 40

6

7

347

Low conversion to liquid High conversion to liquid

0.6 2.6

This limited data suggests that decreasing the residence time, even at high power levels, causes low conversion to liquid polymer. Such a correlation would make sense since the solid polymer, being more highly cross-linked, would probably require more time for formation. C.

B E N Z E N E

CONCENTRATION.

The quantity

Β in Ε = PT/B can


best be studied by varying the partial pressure of the benzene while holding the power and residence time constant.

T h e partial pressure

can be varied either by adjustment of the total pressure of pure benzene vapor or by dilution with a rare gas. Table I I I shows some data obtained at various total pressures and benzene partial pressures. Table III.

Partial Press. φΗ (mm.) 1.0 1.8 2.0 2.0 3.0 3.6 4.0 5.0 a b

Total Press. (mm.) 10 18 2 10 10 18 10 10

Effect of Benzene Partial Pressure (Res. Time = 3.0 sec.)

Power (watts)

Energy Dose (joules/mole) 4.5 1.2 1.2 9.6 1.0 6.2 5.5 5.8

70 34 40 30 45 35 32 41

Χ Χ Χ X X X X X

10 10 10 10° 10 10« 10° 10° 7

7

7

7

Reactant

Result

10% φΗ in He 10% φΗ in He φΗ 20% φΗ in He 30% φΗ in He 20% φΗ in He 40% φΗ in He 50% φΗ in He

High conversion to solid polymer. Low conversion to liquid polymer. Similar series of experiments with benzene in argon and benzene in

neon gave the same results. mm.

L o w partial pressures of benzene—ca. 3.0

or lower—always seem to give the solid polymer while higher

partial pressures yield liquid polymer. One possible explanation of this observation lies in the fact that the energy distribution of the electrons changes with benzene partial pressure. ( The rare gas probably has little effect compared with benzene because of the much lower ionization potential of the latter. ) At higher partial pres sures of benzene, the electron-benzene collision rate increases and the mean free path of the electrons decreases. These changes would cause a limitation of the energy which electrons could gain in the alternating


C H E M I C A L REACTIONS

348

field.

Less energetic reactions

IN E L E C T R I C A L

DISCHARGES

(such as linear polystyrene formation)

would be expected to predominate over more energetic reactions (such as formation of cross-linked polystyrene). It must also be pointed out that Table III shows a general decrease in energy dose as the partial pressure of benzene increases. This suggests a cause and effect relationship between the benzene partial pressure and the energy capable of being absorbed by the system.

Because of this

possible relationship, it is difficult to say whether the energy dose or partial pressure is more important in determining the type of product. D.

INVOLVEMENT

O F RARE

GAS.

Although the solid polymer can

form in the absence of rare gas, some experiments suggest that the rare Downloaded by CORNELL UNIV on July 23, 2016 | http://pubs.acs.org Publication Date: June 1, 1969 | doi: 10.1021/ba-1969-0080.ch029

gas may be involved mechanistically, perhaps as an energy transfer agent. Experiments have not supplied any definitive answers. It is obvious from the above considerations, that it is difficult to study one variable at a time, since they interplay with each other. As in the question of mechanism, a better understanding of the effect of reaction variables will have to await more fundamental studies involving plasma probing, etc. Particle Size Studies.

Electron microscopy studies of the

solid

polymer show it to be composed of spherical particles which tend to agglomerate in chains and clumps. The average sphere diameter of one sample was 0.12 microns. Using this value and a density of 1.10

gram/cc,

the surface area of the sample was calculated to be 45 meter /gram. 2

The

experimentally determined area of the sample ( nitrogen adsorption ) was 42 meterVgram. Thus, the available surface is entirely on the outside of the spheres. Apparently, the cross-linking is so extensive that even nitro gen gas cannot enter into the polymer matrix. A few experiments were performed to see whether the particle size could be regulated by changing the residence time. Table IV summarizes the data obtained. Table I V . Res. Time (sec.)

Particle Size and Residence Time

Range of Particle Sizes (microns)

Average Particle Size (microns)

Refotive Particle Volume

1

0.1-0.5

3

0.09-0.6

0.12

3.00

0.06-0.5

0.14

4.76

—5

a

—

Not enough good particles for a meaningful average. Most of the exposures showed only cloudy blotches rather than discrete particles. α

Although the particle size range shows no correlation with residence time, the volumes of the average particles in the 3 and 5 sec. experiments


29.


NEiswENDER

correlate well.

349

T h e 5 sec. experiment was designed to give a 10 sec.

residence time, but the discharge filled only about one-half of the reactor volume (thus, the residence time of ca. 5 sec.), suggesting that the particle size may have been limited by the exhaustion of the reactants. If fresh reactants could be continually supplied, the particles might grow larger. Of course, there may also be a natural limit caused by a precipitation phenomenon.


Literature Cited (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14)

Austin, J. B., Black, I. Α., J. Am. Chem. Soc. 52, 4552 (1930). Davis, A. P. J. Phys. Chem. 35, 3330 (1931). DeHemptinne, Α., Bull. Sci. acad.roy.Belg. (3) 34, 269 (1897). DeHemptinne, Α., Ζ. physik. Chem. 22, 358 (1897). Ibid., 25, 284 (1898). Harkins, W . D . , Gans, D . M . , J. Am. Chem. Soc. 52, 5165 (1930). Jesch, K., Bloor, J. E., Kronick, P. L., J. Polymer Sci. 4, 1487 (1966). Kleinschmidt, R . V., U. S. Patent 2,023,637 (1930). Losanitsch, S. M . , Jovitschitsch, M . Z., Ber. 30, 135 (1897). Losanitsch, S. M . , Ber. 41, 2683 (1908). Potter, R . C., Bretton, R. H . , Metz, D . J., J. Polymer Sci. 4, 2295 (1966). Stille, J. K., Sung, R. L., Vander Kooi, J., J. Org. Chem. 30, 3116 (1965). Streitwieser, Α., Ward, H . R., J. Am. Chem. Soc. 84, 1065 (1962). Ibid., 85, 539 (1963).

(15) Vastola, F. J., Wightman, J. P., J. Appl. Chem. 14, 69 (1964). RECEIVED

May

2,

1967.


The Polymerization of Benzene in a Radiofrequency Discharge

Recommend Documents