Low pressure RF plasma reactions in light hydrocarbons: Methane

Radiat. Phys. Chem. Vol. 15, pp. 485-495 Pergamon Press Ltd., 1980. Printed in Great Britain

LOW P R E S S U R E RF P L A S M A REACTIONS IN LIGHT HYDROCARBONS: M E T H A N E PIETRO CANEPA, GIANRICO CASTELLO, STELIO MUNARI and MARIO NICCHIA lstituto di Chimica Industriale, Universitfi di Genova, Corso Europa 30, 16132 Genova, Italy (Received 19 March 1979; in revised form 2 July 1979) Abstract--The plasma decomposition of methane in an inductively coupled radiofrequency glow discharge was studied by using a static system at pressures of 0.25, 0.5 and I torr. The effect of input power (between 25 and 100W) and specific energy E~ (up to 3x 106j/g) was investigated. The gaseous products were identified and measured by gas chromatography. The total pressure of the system increased with delivered energy, Es. The main products were H2, C2H6, C2H2, C3H8, C2H4 (in decreasing order of importance). The concentration of H2 increased with Es and reached a constant value depending on initial pressure and input power. During the initial stage of the discharge, the concentration of other gaseous products reached a maximum, and a regular decrease took place then, asimptotically approaching a small constant value, different from zero. The formation of unsaturated compoufids increases with input power and by decreasing the initial pressure. The experimental results were compared with previously reported literature data obtained by M. S. analysis. The yield of polymer and its hydrogefi/carbon ratio were calculated and correlated with the variation of discharge parameter.

INTRODUCTION PLASMA PRODUCED by radiofrequency (RF) glow discharges has been extensively used for organic synthesys "-~°~ and for polymerization of conventional and unconventional monomers t"-39) by using a wide range of experimental conditions (power and frequency of the discharge, pressure and flow of the reactants) and different discharge types (inductive and capacitive coupling). In our laboratory, a research program has been developed, concerned with the study of the behaviour of saturated and unsaturated hydrocarbons in R F plasma, in a range of experimental conditions that allow the synthesys of thin polymeric films. In fact, while R F plasma has been used for the preparation of polymer from unconventional monomers, and the properties of the solid products were widely investigated, few researchers have studied the gaseous products formed during the plasma decomposition. The knowledge of these species is important because it can permit the study of the reaction mechanism responsible for the polymerization, and the control of the characteristics of the final products as a function of the discharge parameters. Studies on the plasma decomposition of simple hydrocarbons have been, therefore, carried out in closed systems and under flow conditions, by carefully controlling the experimental conditions and by analyzing the

gaseous product at various stages of the reaction. In previously published papers the method used for the gas chromatographic analysis of the products formed in the plasma decomposition of hydrocarbons was described, t40-42> In the present paper the experimental results of studies on the decomposition of methane by inductively coupled R F discharges in a static low pressure system are described. Similar studies are in progress with ethane in order to permit a comparison of the behaviour of the two compounds and a kinetic study of the process, taking into account the available informations on pyrolysis, radiolysis and photolysis of these hydrocarbons. Plasma discharges in methane have been previously studied t22'24'43-5°)the gaseous products were examined by measuring the hydrogen yield in a closed system after freezing the condensable gases or by sampling the plasma gas at different points of the discharge and analyzing the ionic and neutral species by mass-spectrometry (MS).

EXPERIMENTAL The schematic drawing of the apparatus used for closed-system discharge is shown in Fig. 1. It is well known from literature~27~ that the shape and surface/volume ratio of the reactor play an important and not completely known role in the discharge behaviour. A simple cylindrical symmetry of constant diameter was 485

486

P. CANEPA

................

et al.

~:a_xi_s___r

1

~

mY, y

¢apacitance L. ,.Jstripchart manometer recorder@

|'"]

torr, P

®

Pmax

(~)

1.; 1,1 1 ON

OFF

X

I i

mV

ON

OFF

time sec

FIG. 1. Schematic diagram of the RF circuit and of the discharge unit. (a) Plot of the x - y recorder. Ea: delivered energy; Er: reflected energy. (b) Plot of the pressure recorder (see text).

chosen, in order to avoid the effect of the section variation that strongly influences the physical parameters of the discharge.(SLm Pyrex glass vials, having an internal diameter of 40 mm, a wall thickness of 1.5 mm and a length of 420 mm, with a total volume of 500 cm 3, necessary in order to permit a good sensitivity of the analysis, were used. Further experiments are in progress in order to elucidate the influence of the diameter and length of the reactor. The RF generator, delivering an output power from 0 to 100W, was operated at 13.5MHz and connected to the coil, wound around the center of the Pyrex vessels, through a network of tunable capacitors. The coil, having a length of 160 mm, an internal diameter of 50 mm, was made of 13 turns of silver plated copper tubing (6mmo.d.). The delivered and reflected power was monitored by a wattmeter (mod. 43 "Bird Electronic Co." (Cleveland, Ohio) 0-100W-+5% F.S.) connected with an integrating circuit and x - y potentiometric recorder. The delivered energy was displayed on the x-axis and the reflected energy, measured by an SWR meter (Fig. la) on the y-axis. The recorder trace monitors the plasma discharge indicating delayed starts, imperfect coil matching, etc. and allows for the calculation of delivered and reflected energy. The discharge time was controlled by an electronic timer (range from 0.1 to 99 sec). The pressure in the reactor was measured with a capacitance manometer (Baratron MKS type 220; l0 - 2 10 torr _ 1%). As the plasma gas methane was used, having a certified purity of 99.995%; Oz and H20 5 ppm, N2 and

C2H6 15 ppm (type N.45 obtained from S.I.O., Milano, Italy). The gas was used without further purification. The vials were baked at 400°C under air flow in order to remove the polymer produced in previous discharges and were connected to the vacuum system, carefully degassed and washed several times with pure methane. The desired initial pressure (PJ was set at 0.25, 0.5 or 1 torr. Taking into account the vial volume, 1 torr (133.32 Pa) of methane corresponds to 2.93 × 10-5 moles or to 0.047 mg. By using the SWR meter, the wattmeter and the x - y recorder, a careful tuning of the matching circuit was accomplished; the reactor was then replaced with a clean one, which was washed and filled with fresh methane, and the discharge was switched on and off by the electronic timer. Discharge times ranging from 0.5 to 20 sec were used. The discharge, grey-pink in colour, completely filled the reactor volume and did not show any structure visible to the naked eye in the considered pressure and power range. A dark sheath of about 2 mm was only observed in contact with the glass walls. In a series of experiments, the final pressure (PI) obtained with different delivered energies in various experimental conditions was measured. In order to avoid the "reservoir effect" due to diffusion of methane from the tubing connecting the reactor to the manometer into the discharge region, the discharges were normally carried out in the closed reactor. P / w a s measured when the reactor was returned at the room temperature, by opening the connecting stopcocks and taking into account the dead volume between the reactor and the manometer.

487

Low pressure RF plasma reactions in light hydrocarbons: Methane In order to follow the pressure variation during the discharges, some experiments were carried out with the reactor continuously connected to the capacitance manometer, During these runs, an increase of pressure was observed, that reached a maximum value (P,=), followed by a decrease when the discharge was switched off (Fig. lb). The final pressure measured in these runs, again taking into account the dead volume, was very close to the PI obtained during the discharge carried out in the closed reactor. The pressure behaviour shown in Fig. l(b) can be mainly attributed to the temperature change of the gas during the discharge, that can be approximately measured by using the reactor-manometer system as a gas thermometer, after appropriate calibration obtained by heating the reactor in an air thermostat and by monitoring the pressure of the gas at various temperatures. The difference Pmax--Pl was, therefore, used as an approximate measure of the mean temperature of the gas in the reactor during the discharge, Tin, obviously not taking into account the true kinetic temperature of the active species, ions and electrons. A T,, ranging between 70 and 120°C was measured during the discharges in methane showing that a relatively "cold" plasma is formed, and that the heat effects on the deposited polymer and on the gaseous products would be negligible. A contribution to the variation of the pressure after the switch-off may be due to recombination of the active species but this effect is probably neglibible. Runs carried out with monoatomic gases (He, Ar), where molecular dissociation is not possible, and with pure H2, that does not yield other products, showed a similar behaviour with the pressure increasing during the discharge and returning to the initial value when the discharge was switched off. The gaseous products of the discharge were analyzed by gas chromatography. After the discharges carried out in the closed reactor, this was connected to a manifold that permits the compression of the products, their dilution with helium and injection in a gas chromatograph equipped with a sensitive thermal conductivity detector (TCD). The separation was carried out at room temperature by using coupled columns of Porapak Q + S and of tetraisobuthylene + dimethylsulpholane, alternatively connected to the gas sampling valve and to the TCD. The sampling technique and the analytical system were described previouslyfl°,4~ The method permitted to measure the mean composition of the products formed in the whole system, for specific energies higher than 100 Jig -I . The evolution of the system as a function of increasing specific energy was followed by analyzing the products of various discharges at the same power input for increasing times. This procedure could give unaccurate results if the recombination of active species, when the discharge is switched off, plays an important role with respect to the reactions taking place during the discharge. In order to ascertain that the results of the used method were comparable to those obtained by sampling the products at various times during an unique discharge, some experiments were carried out by switching on and off the discharge several times, without changing the content of the reactor and delivering the same total energy of a continuous run. The composition of the products was the same, within the limits of the experimental error, thus confirming the validity of the analytical method.

RESULTS AND DISCUSSION The pressure variation in the s y s t e m during the discharge at various initial pressures (P~) and input p o w e r s ( W ) is s h o w n in Fig. 2. On the x axis the values o f specific energy given to the system (Es) and calculated f r o m the difference b e t w e e n delivered and reflected energy are r e p o r t e d . O f course only a small fraction of this energy, unk n o w n at this stage of the research, is used in chemical reactions, due t o energy dispersion in r e a c t o r walls, in the surrounding air and in light and heat production in the plasma. The values of Es are r e p o r t e d as J g-~ or multiples ( J g - ~ = l(P gray (Gy) = 10~ rad) b e c a u s e this unit takes into a c c o u n t the total a m o u n t of material in the reactor, which is constant throughout the whole reaction i n d e p e n d e n t on the composition and physical state of the reaction mixture, and permits an easy comparison with results of e x p e r i m e n t s carried out at various initial pressures and with m o n o m e r s having different molecular weights. S o m e typical analysis of the gaseous products are s h o w n in Table 1; the partial pressures and p o w e r inputs are reported. The characteristic b e h a v i o u r of the a m o u n t of the main products as a

P, torr 12!

_

.

~

.

_

~

100 W

50W

1.0(

0.75

100 W

"

25W

0.50, 50W

Q25'

1

2

3 Es, MTg "1

FZG. 2. Final pressure Pf at various power inputs at a various initial pressure, as a function of the specific energy Es.

P. CANEPAet al.

488

TABLE I. PARTIAL PRESSURES (torr × 10 -3) OF PRODUCTS AS A FUNCTION OF INITIAL PRESSURE P~ POWER INPUT W AND SPECIFIC ENERGY g~ Pi (torr)

0.5

input power (W)

Es

(MJg -I)

1.0

50

i00

50

0.5

1.5

0.5

1.5

hydrogen

330.3

510.3

387.1

methane

170.1

65.2

143.4

acetylene

15.6

11.5

ethylene

4.3 51.2

ethane

I00

0.5

1.5

580.1

425.3

825.3

590.2

936.3

50.3

420.2

180. I

415.7

154.2

17.8

8.5

24.5

22.0

30.3

27.1

2.0

3.9

1.2

6.1

5.1

8.2

3.6

36.1

47.7

26.5

109.0

82.3

86.3

79.8

0.5

1.5

propylene

1.5

1.0

I.I

0.6

2.9

1.9

2.4

1.3

propane

6.5

4.5

5.8

1.5

19.0

19.5

14.5

14.4

n-butane

0.8

0.7

0.4

0.I

4.0

3.7

2.1

1.7

i-butane

1.4

1.6

i.I

0.5

5.4

5.8

3.6

3.2

0.i

0.I

0.i

0.4

0.2

0.2

l-butene isobutene trans-Z-butene

0.2 0.2

0.2

0.4

0.4

0.6

0.3

1.4

0.3

0.3

i.I

0.3

cis-2-butene

function of the energy is shown in Fig. 3. The influence of the change of the input power W is shown in Fig. 4. The partial pressure of the formed H2 clearly increase with W while in the considered W range and with the used sensitivity of the analytical system no appreciable difference can be seen in the decrease of the CH4 partial pressure with Es at different W. On the contrary, a clear dependence of the decomposition speed of CI-L on P~ can be seen in Fig. 5, where the ratios PH2/P~ and PcnJP~ are reported instead of the partial pressures, in order to compare runs at the same W and different P~. The formation of H2 depends more on the variation of P~ than on the change of W, almost in the considered Pg and W ranges. The values of the ratio between the pressure of the formed hydrogen and the decrease of the pressure of methane,ZH2, are shown in Fig. 6(a and b) where the effects of P~ and W can be seen. The formation of C2H6 from methane yields a ZH2 = 0.5, while an "ideal" polymerization which leads to the formation of saturated and non-crosslinked polymethylene of infinite chain length would have a ZH2 = 1. At low Es the formation of gaseous products is predominant and the main product is ethane. The formation of unsaturated species increases with increasing W and decreasing P , as a consequence of the higher electron energy, and

I P,torr

,

l

i

--

(~)

1 . 0 ~

H2

0.5

O.5

1

1.5

i

i

i

0.1

C2H6

0.05

--

05

1.0

¶

C2H4

15

Es, M 1g-1

FIG. 3. Pressure of residual methane and of the main gaseous products as a function of the specific energy/7,. (Runs at Pi = 1 torr; W = 50 W.)

Low pressure RF plasma reactions,in light hydrocarbons: Methane P, torr

'

'

'

0.50

w

489

,

x--

0.25

+ f-H4

x H2

o CH4

• H2

'~ CH4 • H2

0.5

1.0

1.5

2.0

2.5

Es,MTg -I

FIG. 4. Hydrogen production and methane consumption as a function of E, at various W at P~ = 0.5 torr. this fact influences the behaviour of ZH2. At higher Pi the ratio between unsaturated and saturated products is lower. As an example, in the range 0.2-1.5 x 106jg -1 this ratio at P~ = 1 torr is between 0.25 and 0.33 while at P~ = 0.25 torr is between 0.45 and 0.7. During the initial stage of the discharge, the concentration of gaseous products reaches a maximum as shown in Fig. 3(b), and a regular decrease then takes place, asymptotically approaching a constant value, small but different from zero. The pressure of CH4 decreases to a near constant value, that changes slightly with Es and W and mainly depends on P~ (Figs. 4 and 5). The dependence of the yield of light hydrocarbons on Es is shown by the values of Zx (number of molecules of the product x divided by the molecules of CI-h consumed). Figure 7 shows that the values of Zx

decrease by increasing Es for all the gaseous hydrocarbons. The comparison of the yield of gaseous products with previously published data t47-5°) can be made on the basis of the MS analysis of the neutral species produced in a 0.45 or 0.5 torr, 150 V peakto-peak methane discharge, by sampling axially through the R F electrode or through the ground electrode, or radially through an electrically floating orifice in a system with capacitive coupling. Results of experiments accomplished by inductive coupling were not published, but seem to be similar to these obtained by axial sampling of capacitively coupled s y s t e m s : ") Not taking into account the different R F coupling system, the main differences between the technique of Smolinsky and Vasile t47-~°) and our method is that they operated in a flow system, with a concentration of

Px/Pi

1.12

I

I

I

0.5

1.0

1.5

DI

2.0 Es, M I g "I

I

2.5

FIG. 5. Ratio Px/P~ between the partial pressures of H2 and CH4 and the initial pressure of CH4 as a function of Es at various Pi and at a constant input power W = 50 W.

P. CANEPAet al.

490 T

i

i

1.4 ZH2

'Z

'

0.8

(~)

ZH2

~'

~

0.25 tort

L ~

05 torr

1

o8

@ Es, M Ig -1

FIG. 6. Ratio between the pressure of H2 and the decrease of the pressure of CH4 (ZH2)as a function of E,. (a) Pi = 0.5 torr. (b) W = 50 W. products stationary as a function of time, and applied MS analysis to samples taken locally in different points of the discharge, near to the walls or to the electrodes, measuring ionic and neutral species. Our gaschromatographic analysis measured the mean composition of the neutral species in the whole reaction volume, following their concentration in a closed system as a function of time, of the initial pressure and of the power input. In a closed system a mass balance is possible,

and the composition of the gas phase can be correlated with the composition of the polymer at a given Es. An attempt was made in order to compare our results with the literature data t'm-5°) by choosing among our experimental data these runs at P i - 0.5 torr that yielded amounts of formed HE and of unreacted CH4 close to Smolinsky and Vasile's data (see Tables 2 and 3) and comparing the yield of C2 hydrocarbons. Notwithstanding the differences of experimental conditions both during the discharge and in the analysis, and the fact that the species sampled locally near to the walls may be different from the species sampled in the whole system at the end of the reaction, the results fairly agree for radial sampling, that reflects the chemistry occurring essentially in the bulk of the plasma, (49) except for the yield of ethylene that seems to be much lower in our system. The agreement decreases when more energetic regions of the discharge are analyzed. The difference in the ethylene yield may be correlated to the observation that at very short discharge times in our system the mole fraction of C2H4 is high and decrease rapidly by increasing the specific energy. As the MS data were obtained in a flow system, the residence time of the species in the various regions of the discharge could play an important role and explain the difference observed with respect of our static system. A possible source of the higher amount of C2H4 in the flow discharge-MS system could also be the decomposition to CH3 ÷ and C2H4 of C3H7 + ions having the structure of a protonated cyclopropane, c54"5~ Table 3 shows that the sum of propane and ethylene yields in our system is very

r Zx 03-

i

!

°2FI 01!

05

1.0

1.'5

20

2.5

Es, M Tg-1

FIG. 7. Ratio between the pressures of the main products and the decrease of the pressure of CH4 (Zx) as a function of E,. A, C2H6; II, C2H4; El, C2H2; •, C3H8. Discharge parameters: P~ = 0.5 torr; W=50W.

491

L o w pressure R F plasma reactions in light hydrocarbons: Methane TABLE 2. MOLE FRACTIONOF GASEOUS HYDROCARBONSPRODUCEDBY CAPACITIVELYCOUPLED METHANE DISCHARGE(LITERATUREDATA), AT 0.45-0.5 torr, 150 V PEAK-TO--PEAK Data from ref. 48 RAD

Product

Data from ref.50

RFG

RFA

PAD

RFG

RFA

H2

O.15

0.23

O.31

0.22

O.41

0.63

CB 4

0.78

0.70

0.60

0.69

0.49

0.30

0.023

O.O18

C2B 2

O.O12

O.O13

O.O19

O.O17

C2H 4

O.O11

O.O11

0.013

0.O17

O.O12

O.013

C2H 6

0.045

0.O41

0.055

0.058

0.O61

0.040

C2H2/C2H6

0.27

0.32

0.35

0.29

0.38

O.45

RAD : radial sampling RFG : axial sampling through the RF electrode RFA : axial sampling through the ground electrode

T A B L E 3. M O L E FRACTION OF GASEOUS HYDROCARBONS PRODUCED BY INDUCTIVELY COUPLED METHANE DISCHARGE AT A Pi OF 0 . 5 t o r r . R U N S GIVING YIELDS OF n2. SIMILAR TO THE LITERATURE VALUES OF T A B L E 2 ARE REPORTED FOR COMPARISON (SEE TEXT) Energy (J/g) input power (w)

54.103 25

108.103 25

129.103 50

214.103 25

255.103 50

540.103 IO0

730-103 50

806.103 25

Product

H2

O.14

0.23

0.23

0.33

0.40

0.63

0.67

CH 4

0.79

0.68

0.68

0.56

0.49

0.23

0.22

0.20

C2H 2

O.O10

O.014

O.O19

O.017

0.027

0.029

O.O18

O.O16

C2H 4

0.006

0.005

0.005

O. 005

0.005

0.006

0.003

0.006

C2H 6

O.041

0.063

0.050

0.074

0.072

0.078

0.078

0.078

C3H 8

0.003

0.003

0.006

0.008

0.O10

0.009

0.008

0.009

close to the yield of C2H4 reported in the cited literature. The differences between the other products can be correlated with the different energy of the electrons due to the discharge parameters. A higher value of the C2H2/C2H6 ratio probably shows a more energetic situation of the plasma correlated to a higher power input or to a lower pressure. Table 4 shows the mean values of the C2H2/C2H6 ratio at various W and Pi. For a given experimental condition (constant Pi and W) this ratio is constant in the whole Es range investigated. Comparison with the data of Table 2 shows that this ratio increases with the electron energy also in the capacitive coupling-MS system, where the energy of the processes sampled axially through a ground electrode (RFG) from either a capacitive or inductive coupling discharge is intermediate between the high energy of processes

0.68

sampled axially through the RF electrode (RFA) and the low energy of processes sampled radially through an electrically floating orifice (RAD)} ~) Other literature data t'~) report the results of MS analysis of the gases formed during the plasma polymerization of methane at energies that produced a hydrocarbon decomposition greater than 98%. In these conditions the following gaseous products were detected: H296.7%, CH41.3%, C2H60.4%, C2H21.1%, C2H40.5%. Our experiment at high Es (Es - (4.2 × 106 J g-l) yielded H294.8%, CH44.1%, C2H60.7%, C2H20.3%, C2H4 less than 0.05%.

492

P. CANEPA et al. TABLE 4. RATIO BETWEEN THE MOLE FRACTIONS OF ACETYLENE AND ETHANE AT VARIOUS AND INITIAL PRESSURES Initial

Pressure

Input

Power

{Wa~

1

(Torr)

25

0.25

-

0.63

0.5

0.23

0.32

0.38

1.0

-

0.24

O.31

The results of other experiments on the plasma decomposition of methane alone or in mixture with other compounds t6'7"9) cannot be compared at this stage of the research with our data, due to the wide difference of experimental conditions. Experimental evidences show that at very high Es an equilibrium concentration is reached between the products formed by decomposition of CFL and by cracking of the polymer, depending on the conditions of the discharge. While the yield of gaseous products decreases with E,, the amount of polymer increases, and at high Es the system is formed by H2, polymer and some light hydrocarbons in nearly constant proportions, depending on the discharge parameters. The polymer yield Yp (weight of polymer divided by the initial weight of CH4) has been calculated by measuring the difference between the initial weight of methane in the reactor and the total weight of the gaseous products (H2, CH4, C2 to C4 hydrocarbons) deduced from the gaschromatographic analysis. As shown by some analysis by flame ionization detector ~4,.42>the amount of products having 5-8 carbon atoms was negligible and was not taken into account in this calculation. The behaviour of Yp is reported in Fig. 8 (a and b) as a function of W and Pi. Yp increases with increasing W, but the influence of this parameter is relatively small in the considered range, and three separate curves cannot be correctly drawn in Fig. 8(a). The effect of Pi on Yp is much greater and

50

iOO

1.0

®

v 05

f

i

J

1.0

2.0

r

I

Yp

1.0

@

10

20 Es,M Tg -1

FIG. 8. Polymer yield Yp as a function of Es. (a) At constant Pi=0.5torr: ©,25W; Fq, 50W; ~7,100W. (b) At constant W = 50 W; 0, 0.25 torr; IS],0.5 torr, A, 1 torr. sigmoid curves are observed (Fig. 8(b)) due to a change in the polymerization rate. The polymer formation rate is shown in Fig. 9, where the ratio AYo/AE~ (with AEs = 105 dJ g-J) is reported.

TABLE 5. HYDROGEN/CARBON RATIO 1N THE HYDROCARBON FRACTION OF THE GAS (R~) AND IN THE POLYMER (Rp). CONSTANT VALUES AT SPECIFIC ENERGIES Es < 1 0 6 j g - ' AS A FUNCTION OF INITIAL PRESSURE Pi AND INPUT POWER W Input

Power 25

R' g 50

0.25

-

2.62

-

-

1.30

0.5

3.0

3.06

3.1

1.56

1.44

1.26

i.O

-

3.14

3.1

-

1.50

1.4C

(Watt)

R iO0

25

P 50

iOO

Initial Press u r e (torr) -

Low pressure RF plasma reactions in light hydrocarbons: Methane 0.8L

AYp/AEs

'

'

r

~.'o

i5

,

"

493

I

/

o.61'

o.25 torr

0

.

~

4

torr

o15

2'.o

215

Es,MTg"1

FIG. 9. Polymer formation rate as a function of E~ at W = 50 W and various Pi. By taking into account the hydrogen-carbon balance in the system and the composition of the gas phase, the ratio H/C in the whole gas phase RB (including H2), in the hydrocarbon fraction of the gas phase R~ (excluding H2), and in the polymer, Rp, can be calculated. Obviously R~ = R~ = 4 when E~ =0, and the whole system is formed by methane. As soon as higher hydrocarbons are formed R~ values decrease and Rp values increase approaching constant values for Es > 106Jg-L A typical trend of the plot is shown in Fig. 10. Table 5 shows the constant "final" values (at Es > 106 J g-'), of R~ and Rp, that depend on the discharge parameters. The change of Pi has a greater effect on the "final" R~ values than the change of W, while the change of W has a more enhanced effect on the "final" Rp values. Below E, = 3 x 10~J g-~ the trend of Rp cannot be evaluated as the values are rather scattered, probably because they are obtained by carbonhydrogen balance on gaseous products at a very low concentration and by difference from the initial amount of CH4. The analytical technique must therefore be improved and more experiments at short discharge times have to be accomplished in order to clarify the behaviour of Rp at low Es. Furthermore, elemental analysis must be used to confirm the calculated Rp values and this requires a flow discharge system in order to produce larger amounts of polymer. Data that can be compared with our Rg and R~ values are not available in the literature while some values of the H/C ratio in the polymer were determined. Values of 1.47, 1.49, 1.52, and 2.4 were found °2'4~'m'56) respectively). IR analysis

i

i

10 HIC

/

s

4 .

_-

.

=

_-

-

•

R'Q

2 •

°

Rp

Es, M l g -I

FIG. 10. H/C ratio in the gas phase including H2 (R~);-in the gas phase excluding H2 (R~) and in the polymer (Rp) as a function of Es-Pi = 1 torr, W = 100 W. showed that the solid product from RF methane decomposition is a highly crosslinked-highly branched saturated polymer. (45'56) One may suppose that for high Es values the reaction proceeds to the formation of a solid phase consisting of pseudo graphitic carbon and of a gas

494

P. CANEPAet al.

phase formed by hydrogen due to increasing hydrogen abstraction from the formed polymer. This fact was not observed in the considered Es range while a valuable amount of C atoms were still present in the gas phase as light hydrocarbons. Some experiments showed that light hydrocarbons are formed by reaction of an hydrogen plasma on the polymer previously produced by methane plasmolysis. Furthermore, literature data show that methane, ethane and unsaturated hydrocarbons are formed by reaction of hydrogen and solid carbon.~ 57.5s) Therefore, light hydrocarbons are continuously formed by reaction of H2 with the solid phase, and can subsequently polymerize thus continuously producing fresh polymer. As a consequence, the distribution of H and C atoms between the gas and solid phases tends to an equilibrium depending on the experimental conditions, but an appreciable amount of hydrogen is present in the solid phase even at very high Es. CONCLUSIONS The above reported experiments permitted to follow the evolution as a function of time and energy of a methane plasma in a closed system, at various initial pressure and input power. The pressure variation and the concentration of the majority of the products were measured with a satisfactory accuracy. The behaviour of methane in R F plasma discharge at low pressure in a closed system can be summarized as follows: the pressure of the system increases with the energy at a rate depending on the power and on the initial pressure, that also influence the final pressure of the system. The mass balance of the system, shown in Fig. I1 for a typical run at 0.5tort, 100W, has the following behaviour: the methane consumption follows a nearly exponential decrease, and a con1

2

3

4

5

6

7

8

t, sec 9

10

CH4¼

I lo~ wo,ght / / / ~

stant concentration, small but different from zero, is reached at high energies. The final main products of the plasmolysis are hydrogen and a crosslinked polymethylene polymer. The constant concentration of residual methane at high energies is probably due to an equilibrium between polymerization and decomposition of the polymer. Low molecular weight products, mainly ethane, ethylene and propane, are formed, and their amount is greater in the initial stage of the plasmolysis, when the polymer formation is relatively small. The concentration of these gaseous products as a function of the energy shows maxima whose abscissas and heights depend on the experimental conditions and differ for the various compounds. The behaviour shown in Fig. 11 and the previously discussed results seem to confirm that only when an appreciable amount of these intermediate products is formed the polymer formation begins to become important. These products play therefore an important role in the polymerization mechanism. At high energies, the amount of these products decreases approaching a constant value, very slightly influenced by further increase of the energy. As in the case of residual methane, this steady concentration is probably due to the equilibrium between the formation and the decomposition of the polymer. Some aspects of the plasmolysis of methane have to be further investigated, as the influence of the composition of the polymer on the equilibrium concentration of the gas phase at high energies, the kinetic of the decomposition of methane and of formation and decomposition of the other gaseous products, the composition of the polymer formed in the first stage of the reaction and the influence of the formed gases (mainly H2) on the kinetic of the process and on the composition of the polymer. The geometry of the reactor (volume, diameter, length) and of the R F system (frequency., coupling, dimensions of the coil, etc.) must be changed in order to study the influence of these parameters. Research is in progress in order to clarify these aspects of the decomposition in a R F plasma of methane and other light hydrocarbons.

Othergaseous proc1ucts*/o

1.0

20

3.0

4.0

Es, iVl Tg-I

FIG. 11. Mass balance of the system in a run at 0.5 torr, 100 W. Weight percentages of the various components in the reactor are reported as a function of E,.

REFERENCES 1. R. F. GOULD (editor), in: Advances in Chem. Series No. 80, Am. Chem. Soc. Washington, 1969. 2. M. VENUGOPALAN(editor), Reactions under Plasma Conditions. Wiley-Interscience, New York, London, 1971. 3. J. R. HOLLAHANand A. T. BELL (editors), Techniques and Applications o[ Plasma Chemistry. Wiley Interscience, New York, London, 1974.

Low pressure RF plasma reactions in light hydrocarbons: Methane 4. S. W. LIU and J. P. WIGHTMAN, J. Appl. Chem. Biotechnol. 1971, V21,168. 5. H. SUHR (editor), Angew. Chem. Int. 1972, II, 781. 6. P. CAPEZZUTO, F. CRAMAROSSA, G. FERRARO, P. MAIONE and E. MOLINARE, Gazz. Chim. It. 1973, 103, 1153. 7. E. MOLINARI, Pure Appl. Chem. 1974, 39, 343. 8. J. P. WIGNTMAN, Proc. IEEE, 1974, 62, No. I. 9. J. AMOUROUX and D. RAPAKOULIAS, Ann. Chim. 1977, 2, 257. 10. M. TEZUKA and L. L. MILLER, J'. Am. Chem. Soc. 1978, 100, 4201. l l. H. YASUDA and C. E. LAMAZE, J. Appl. Polym. Sci. 1971, VI5, 2277. 12. A. R. WESTWOOD, Europ. Polym. J. 1971, 7, 363. 13. L. F. THOMPSON and K. G. HAYAN, J. Appl. Polym. VI7, 1519. 14. L. F. THOMPSON and K. G. MAYHAN, J. App. Polym. Sci. 1972, 16, 2317. 15. A. BRADLEY, 3". Electrochem. Soc. 1972, 119, 1153. 16. K. C. BROWN, Europ. Polym. J. 1972, 8, 117. 17. K. C. BROWN and M. J. COPSEY, Europ. Polym. J. 1972, 8, 129. 18. H. YASUDA and C. E. LAMAZE, J. Appl. Polym. Sci. 1973, V17, 1519. 19. H. YASUDA and C. E. LAMAZE, J. Appl. Polym. Sci. 1973, Vl7, 1533. 20. M. DUVAL and A. THI~OR/~T, J. Appl. Polym. Sci. 1973, 17, 527. 21. H. CARCHANO, J. chem. Phys. 1974, 61, 3634. 22. H. KOSAYASHI, A. T. BELL and M. SHEN, Macromolecules 1974, 7, 277. 23. H. KOBAYASHI, M. SHEN and A. T. BELL, J. Macr. Sci. 1974, A8, 373. 24. H. YASUDA, M. O. BUMGARNERand J. J. HILLMAN, J. Appl. Polym. Sci. 1975, VI9, 531. 25. H. YASUDA, m. O. BUMGARNERand J. J. HILLMAN, J. Appl. Polym. Sci. 1975, VI9, 1403. 26. H. YASUDA, i . O. BUMGARNERand J. J. HILLMAN, .]'. Appl. Polym. Sci. 1975, VI9, 2845. 27. H. YASUDA, T. HSU, J. AppL Polym. Sci. 1976, V20, 1769. 28. J. M. TIBBITT, M. SHEN and A. T. BELL, J. Macr. Sci. Chem. 1976, AI0, 1623. 29. H. U. POLL, i . ARTZ and K. A. WICKLEDER, Europ. Polym. J. 1976, 12, 505. 30. H. YASUDAand T. HIROTSU, J. Polym. Sci. 1977, Vl5, 2749. 31. H. YASUDA and T. HSU, J. Polym. Sci. Polym. Chem. Ed. 1977, VlS, 81. 32. J. M. TIBBITT, A. T. BELL and M. SHEN, J. Macr. Sci. Chem. 1977, I l l , 139. 33. H. J. TILLER, D. BERG, D. HERZBERG and K. DUMKE, J. Macr. Sci. Chem. 1977, A l l , 547.

495

34. J. M. Tissrrr, R. JENSEN, A. T. BELL and M. SHEN, Macromolecules 1977, 10, 647. 35. H. HIRATSUKA, G. AKOVALI, M. SHEN and A. T. BELL, .I. Appl. Polym. Sci. 1978, 22, 917. 36. S. I. IVANOV, S. H. FAKIROV and D. M. SVlRACHEV, Europ. Polym. J. 1978, 24, 611. 37. A. K. SHARMA, F. MILLICH and E. W. HELLUTH, Polym. Preprints, 1978, Vl9, No. 2, 435. 38. I. TANIGUCHI, K. TSUNETO, Polym. Preprints, 1978, Vl9, No. 2, 447. 39. H. YASUDA, T. HIROTSU, J. Appl. Polym. Sci. 1978, V22, 1195. 40. G. GASTELLO, P. CANEPA and M. NICCHIA, Atti

Convegno Sulle Methodologie Analitiche Avanzate, 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58.

p. 108. Rums, 1976. G. CASTELLO, P. CANEPA and M. NICCHIA, Ann. Chim. 1977, 67, 149. P. CANEPA, G. CASTELLO and M. NICCHIA, Ann. Chim. 1978, 68, 543. R. L. MCCARTY, J. chem. Phys. 1954, 22, 1360. A. S. STUDNARZand J. L. FRANKLIN, J. chem. Phys. 1968, 49, 2652. J J. P. WIGHTMAN and N. J. JOHNSTON, in: Am. Chem. Soc. 1969, 80, 322. F. J. VASTOLA and J. P. WIGHTMAN, J. Appl. Chem. 1964, 14, 69. G. SMOLINSKY and M. J. VASILE, Int. J. Mass. Spectrum. Ion Phys. 1975, 16, 137. M. J. VASILE and G. SMOLINSKY, Int. J. Mass. Spectrum. Ion Phys. 1975, 18. 179. M. J. VASILE and G. SMOLINSKY, Int. J. Mass. Spectrum. Ion Phys. 1976, 21,263. G. SMOLINSKY and M. J. VASILE, J. Macromol. Sci. Chem. 1976, Al0, 473. J. G. ANDREWS and R. H. VAREY, Nature 1970, 225, 270. J. G. ANDREWS and R. H. VAREY, Proc. R. Soc. London, 1971, 320, 459. M. J. VASILE, private communication. H. M. GRUBS and S. MEYERSON (Edited by F. W. McLafferty), Mass Spectroscopy of Organic Ions. Academic Press, New York, 1963. H. BUDZlKIEWICZ,C. DJERASSI and D. H. WILLIAMS, Mass Spectrometry of Organic Compounds. HoldenDay, San Francisco, 1967. C. T. WENDEL and H. H. WILEY, J. Polym. Sci. 1972, 10, 1069. J. D. BLACKWOODand F. K. MCTAGGART, Australian J. Chem. 1959, 12(4), 533. A. B. KING and H. WISE, J. phys. Chem. 1963, 67, 1163.