Journal of Experimental and Theoretical Physics, Vol. 92, No. 6, 2001, pp. 986–990. Translated from Zhurnal Éksperimental’noœ i Teoreticheskoœ Fiziki, Vol. 119, No. 6, 2001, pp. 1137–1142. Original Russian Text Copyright © 2001 by Brizhinev, Golubev, Dorozhkina, Eremin, Zorin, Litvak, Plotnikov, Razin, Semenov, Strikovskii, Tolkacheva.

PLASMA, GASES

Microwave Discharge on a Dielectric Surface in Vacuum M. P. Brizhinev, S. V. Golubev, D. S. Dorozhkina, B. G. Eremin, V. G. Zorin*, A. G. Litvak, I. V. Plotnikov, S. V. Razin, V. E. Semenov, A. V. Strikovskii, and O. N. Tolkacheva Institute of Applied Physics, Russian Academy of Sciences, Nizhni Novgorod, 603155 Russia *e-mail: [email protected] Received December 29, 2000

Abstract—The paper describes the results of investigation of a discharge arising in vacuum on the surface of solid dielectric materials when irradiated by intense (up to 25 MW/cm2) electromagnetic centimeter wave radiation. When the density of the microwave energy flux exceeds some threshold value depending on the target material, a discharge emerges in the vicinity of the surface. Its emergence is associated with the evaporation of the target material and the breakdown of evaporated matter. The thus forming plasma initially has the form of a thin (on the wavelength scale) layer with the electron density of the order of 1016 cm–3. It is demonstrated experimentally that effective generation of multiply charged ions occurs in the plasma. The measured energy distribution of ions in expanding plasma agrees with the predicted distribution obtained in solving the problem on quasineutral expansion into vacuum of a localized bunch of collisionless plasma with cold ions. © 2001 MAIK “Nauka/Interperiodica”.

1. INTRODUCTION Recently, a marked increase of interest has been observed in the investigations of a discharge arising in vacuum in the vicinity of the surface of solids when irradiated by intense electromagnetic radiation. This interest is due, on the one hand, to progress reached in the development of high-power microwave oscillators, which made it possible to investigate the discharge at high values of microwave radiation heretofore inaccessible (of the order of tens of megawatts per square centimeter) and, on the other hand, to the possible practical application of such discharge for modifying the surface of solids [1, 2] and for developing ion sources. It appears of interest to investigate this discharge from the standpoint of high-power electronics, because the development of a discharge at outlet windows and insulators of REB oscillators may restrict the power and duration of microwave pulse. This paper gives the results of investigation of a discharge arising on the surface of dielectric materials when irradiated by powerful quasioptical beams of electromagnetic waves; in particular, the paper contains the first experimental data pertaining to the behavior of the expansion of a multicomponent plasma and to the efficiency of generation of multiply charged ions. 2. EXPERIMENTAL RESULTS AND THEIR DISCUSSION The experiments were performed using high-power short-pulse microwave carsinotron radiation. Radiation with the frequency of 10 GHz and pulse duration of 40 ns was formed into a quasioptical beam of linearly

polarized electromagnetic waves and focused to a vacuum chamber (the intensity in the focal region reached 25 MW/cm2). The cross-sectional area of the focal spot was 10 cm2. The pressure in the chamber was maintained at a level of p ≈ 10–3 to 10–5 torr. When dielectric materials were brought into the focal region of the microwave beam, a discharge occurred on their surface (the discharge was registered by a flash of light), with the radiation intensity at the moment of emergence of the discharge exceeding some threshold value dependent on the target material and independent of the residual gas pressure in the range employed in the experiment (p < 10–3 torr). For example, the threshold value of the intensity during the emergence of the discharge was 20 MW/cm2 on Teflon, 8 MW/cm2 on glass, and 2 MW/cm2 on Plexiglas. A photograph of the discharge is given in Fig. 1. The plasma glow was a plurality of filaments extended in the direction of the electric field of the wave. The characteristic transverse dimension of filament was 0.1–0.2 cm, the mean distance between filaments was 0.2–0.4 cm, and the length of filaments was defined by the transverse dimensions of the microwave beam and reached several centimeters (up to 10 cm). The space-time characteristics of discharge luminescence were investigated using a high-speed electron-optical streak camera. The velocity of discharge propagation on the dielectric surface from the focal spot center in the direction of the electric field of the wave reached 108 cm/s, and the velocity of motion of the ionization front toward the incident microwave was Vz ≈ 3 × 107 cm/s. Figure 2 gives a characteristic optical

1063-7761/01/9206-0986$21.00 © 2001 MAIK “Nauka/Interperiodica”

MICROWAVE DISCHARGE ON A DIELECTRIC SURFACE IN VACUUM

( γ – 1 )x N ( x, t ) = N 0  1 – -------------------------  ( γ + 1 )V s t

Fig. 1. A photograph of a discharge on Plexiglas, taken along the axis of a microwave beam. The dark strip at the frame center is the shadow of a Langmuir probe, and the arrow on the left indicates the direction of the microwave electric field.

z

2/ ( γ – 1 )

,

(1)

where N(x, t) describes the spatial distribution of vapor density, x is the distance from the target, t is the time from the beginning of the evaporation process, N0 is the density of vapors at the target surface, Vs is the velocity of sound in vapors at the target surface, and γ is the adiabatic exponent in vapor treated as ideal gas. The density of the vapor flux from the target is in this case equal to the product N0Vs. For estimation, we assume that Vs ≈ 105 cm/s to find that the vapor density at the target JOURNAL OF EXPERIMENTAL AND THEORETICAL PHYSICS

t Fig. 2. An optical scan of a discharge. The z axis corresponds to the coordinate perpendicular to the target plane, and the t axis indicates time.

10 Intensity, arb.units

scan of a discharge that illustrates its propagation toward microwave radiation. The development of a discharge at a fairly high power of radiation was accompanied by complete shielding of microwave radiation. Characteristic oscillograms of a transmitted microwave signal are given in Fig. 3. These measurements were performed with the aid of a cooled, fast-neutron, germanium microwave detector and special calorimeters. Given a high oscillator power, the transmission factor through a plasma layer decreased more than 50 times after approximately ten nanoseconds. The observed development of a discharge may be associated only with the breakdown of evaporating matter of the dielectric target. The intrinsic absorption of microwave radiation by the dielectric is too low even for its appreciable heating. Batanov et al. [3] have assumed that, in the case of high intensity of microwave radiation, a secondary-emission discharge arises on the dielectric surface in vacuum, whose electrons bombard the surface to cause an appreciable increase in the electrical conductivity in the thin surface layer (the socalled induced conductivity). It was the absorption of microwave energy in this layer that apparently resulted in its heating and evaporation with subsequent breakdown of the vapors. The amount of evaporated matter in our experiments was estimated by the variation of pressure in the chamber after each discharge and by the recoil momentum acquired by the target during evaporation of matter. The pressure was measured with the aid of an open ionization lamp at several distances from the discharge at the moments of time 2, 3, and 30 ms after the termination of the microwave pulse, when the plasma no longer affected the accuracy of measurement. The pressure increment was (4–8) × 10–5 torr at a background pressure of 4 × 10–4 torr. The estimate of evaporated target matter for the experimental conditions varies from 6 × 1017 to 1.5 × 1018 particles per shot. The mass of evaporated matter was estimated by the recoil momentum (acquired by the target after the microwave shot and measured using a pendulum sensor) and found to agree with the estimate made by the pressure increment. The expansion of evaporating matter into vacuum proceeds at a speed of the order of the sound velocity corresponding to the evaporation temperature. For a constant evaporation rate, this process is described by self-similar expansion wave [4],

987

a

8 6 b

4 2 0

c 10

20 30 Time, ns

40

Fig. 3. Oscillograms of (a) incident and transmitted microwave radiation for different values of microwave radiation power; (b) 4 MW/cm2, (c) 12 MW/cm2. Vol. 92

No. 6

2001

BRIZHINEV et al.

Uplas, kV

4

8

‡

3

6 b

2

4

1 0

2 10

20

30 40 50 Time, ns

60

70

to demonstrate that the desired decrease in the field amplitude occurs if the total number of electrons per unit area of the layer reaches a value of the order of

Intensity, arb.units

988

50cN c /ω ≈ 2 × 10 cm . 13

80

Fig. 4. Oscillograms of (a) microwave radiation transmitted through plasma and (b) plasma potential.

reached values of (1–5) × 1019 cm–3 characteristic of air at atmospheric pressure. The vapors expanded to take up, during the characteristic time of discharge development τi ≈ 10–8 s, a region of thickness lg ≈ 5 × 10–3 cm. Within this region, the vapor density decreases rapidly away from the target, so that, even at a distance of lg/2, the effective frequency ν of collisions between electrons and vapor molecules comes to be of the order of the cyclic field frequency ω. Estimates indicate that, for the intensity of microwave radiation characteristic of the experiment, the frequency νi of ionization by electron impact in an undisturbed field reached values of the order of 1010 s–1 in dense gas at the dielectric surface and of 1011 s–1 in the region of its maximum, where ν ≈ ω . The breakdown of vapors in these conditions occurs very rapidly (during a time of less than 1 ns), and the plasma being thus formed proves to be localized initially in a thin layer with a thickness of the order of 10–3 cm, in which the collision frequency of electrons is of the order of the field frequency. The plasma density in this layer continues to increase until, because of its screening effect, the microwave field intensity in the vicinity of the target decreases to a level at which the ionization frequency turns out to be of the order of the characteristic frequency νloss of loss of plasma from the breakdown region. Estimates indicate that the main loss of plasma in the region of maximum of ionization is due to its ambipolar diffusion, and νloss ≈ 109 s–1. This value of the ionization frequency in the region being treated is attained with a microwave electric field amplitude of the order of several kV/cm, while the undisturbed value of amplitude is approximately 100 kV/cm. The thin (on the wavelength scale) plasma layer may provide for a corresponding decrease in the field amplitude owing to reflection of incident radiation. One can use the known formulas for the reflection factor Γ of the thin plasma layer, J Γ ≈ – ------------ , 1+J

N e dx ic -, J = ---- ------------------------2 N c ( ω – iν )

∫

E = E 0 ( 1 + Γ ),

(2)

–2

Here, c is the velocity of light in vacuum, Ne is the electron density, and Nc is the critical plasma density. Integration is performed over the entire thickness of the plasma layer; E and E0 denote the complex amplitude of electric field in plasma and the amplitude of incident plane wave, respectively. Therefore, the breakdown of a thin layer of vapor in the vicinity of the target must bring about, during a time of the order of 1 ns, the formation of a plasma layer 10–3 cm thick with an electron density of the order of 1016 cm–3, which shields the target from incident microwave radiation. Because of shielding, the ionization processes must decelerate considerably, but the plasma layer will expand first because of diffusion and then, when the plasma extends outside of the vapor cloud, its free expansion will begin at the ionic-sound velocity, which is apparently registered by the streak camera as the ionization front motion. Movable electric probes were used to investigate the plasma potential and the characteristics of plasma expansion. The plasma potential Uplas was measured by a solitary probe of high load resistance (several megaohms) placed in the vicinity of the dielectric (a typical oscillogram of the potential is given in Fig. 4). The measurement results demonstrated that the plasma potential increased rapidly after a dense plasma was formed, reached a value of several kilovolts, and was maintained at this level for a long time. The high potential of the plasma points to a high electron temperature. In all probability, this high potential is developed by the plasma on the periphery of the discharge, where the plasma density is low, and the microwave field amplitude at a distance of quarter the wavelength from the dense plasma layer (i.e., at the antinode of the standing wave being formed) may even exceed the incident wave amplitude. Therefore, the electron temperature here is maintained at a high level (of the order of oscillatory energy of electrons, which amounts to several keV). In addition, the ions escaping from the discharge were subjected to time-of-flight and energy analysis. The measurements were performed using a five-channel ion analyzer enabling one to determine the time that ions with different energies arrived at the analyzer. The analyzer was located at a distance of approximately 3 m from the discharge. The collimator axis of this instrument coincided with the direction of the electric vector in the wave. The analyzing element of the instrument was provided by a capacitor which deflected the ions to an angle defined by their energy. The particles were further delivered to five cylindrical capacitors separating ions of certain energy and were registered by a secondary-emission multiplier (SEM). A typical oscillogram

JOURNAL OF EXPERIMENTAL AND THEORETICAL PHYSICS

Vol. 92

No. 6

2001

MICROWAVE DISCHARGE ON A DIELECTRIC SURFACE IN VACUUM

2 –1/2

f (V ) ∝ [V0 – V ] 2

,

(3)

where V0 is the velocity of electrons of energy m V 0 /2 ≈ 3 keV. It is interesting to note that, in accordance with the results of Ignat’ev and Rukhadze [6], it is the forming of just such a distribution function that one should 2

JOURNAL OF EXPERIMENTAL AND THEORETICAL PHYSICS

C+++ C++

C+

+ H+ H 2

Fig. 5. An oscillogram of the SEM current of ion analyzer. The oscillograph scan, 10 µs per division.

Energy of carbon ions, eV

6000

4000

2000

0

20

40 Time, µs

60

80

Fig. 6. The time of arrival of C+ ions at the analyzer as a function of their energy.

dIv /dW, rel. units

of the SEM current IV (t) is given in Fig. 5. The peaks in the oscillogram correspond to ions whose energy per unit elementary charge was defined by channel tuning (the oscillogram in Fig. 5 corresponds to tuning the channel to the energy of singly ionized ions of 200 eV). Assuming that the ions fly from the target to analyzer uniformly for a large part of the track [5], one can further determine their velocity by the time of arrival of ions at the analyzer. Therefore, a quite certain ratio between the ion charge and its mass may be assigned to each peak in the SEM current oscillogram, which enables one to identify ions. The results of such identification are given in Fig. 5. The first (by the time of arrival) two peaks correspond here to ions of atomic and molecular hydrogen, and the next three peaks are associated with carbon ions with the charge numbers +3, +2, and +1, respectively. The results of these measurements lead one to conclude that multiply charged ions are effectively formed in the discharge plasma; the values of density of carbon ions with charges of 1, 2, and 3 are comparable. By varying the channel tuning, one could determine the dependence of the time of arrival of ions of each type on their energy W. The measured respective dependence for singly ionized carbon ions is given in Fig. 6 (points). Also given in Fig. 6 for comparison is a prediction curve obtained assuming that ions expand with a constant velocity (W ∝ t –2). Such a dependence may also be obtained using the solution of the problem on the expansion of a localized plasmoid to vacuum [5]. The agreement between the predicted and experimentally obtained results points to the validity of the initial assumption of the inertial behavior of ion expansion that was used in the identification of ions. Figure 7 gives the energy distribution (points) of C+ ions registered by the analyzer and a curve corresponding to the dependence proportional to 1/W. In the energy range W < 3 keV, the experimental points coincide well with this curve; in the case of high values of energy, the points lie much lower than the curve. This means that, in the above-identified range, the energy spectrum of ions arriving at the analyzer is inversely proportional to energy; at high values of energy, this spectrum decreases much more abruptly. Such an energy distribution of ions, experimentally recorded away from the plasma source, may be obtained in solving the problem on expansion of a quasineutral bunch of collisionless plasma with cold ions, if the initial velocity distribution of electrons along the collimator axis (i.e., in the direction of the electric vector in incident electromagnetic wave) is defined by the expression

989

3 2 1

0

1000

2000 3000 Ion energy, eV

4000

Fig. 7. The energy distribution of C+ ions. Vol. 92

No. 6

2001

990

BRIZHINEV et al.

expect in the case of gas ionization in superstrong microwave fields in which the free oscillation energy of electrons is much higher than 100 eV. In this case, the value of V0 is defined by the amplitude of the oscillation velocity of electrons. Under experimental conditions, the direction from the discharge plasma to the ion analyzer coincided with the direction of polarization of incident microwave radiation, and the characteristic values of the oscillatory energy of electrons reached several kiloelectron-volts. Therefore, one can assume that the ions registered by the analyzer accelerated as a result of expansion of the plasma formed on the discharge periphery where the collision frequency of electrons is low compared with the field frequency, and the amplitude of electric field is of the order of its amplitude in an incident electromagnetic wave. 3. CONCLUSION The investigation results have demonstrated that the plasma arising in the vicinity of a dielectric target in vacuum when irradiated by intense electromagnetic radiation is characterized by a number of unique properties that may define the future uses of this plasma. The plasma turns out to be substantially nonequilibrium, with its density reaching the value of 1016 cm–3. The volume and shape taken up by the plasma may vary depending on the shape and size of the microwave

beam and the target. In the experiments described in this paper, the plasma had the form of a thin disk approximately 10 cm in diameter. ACKNOWLEDGMENTS This study received support from the Russian Foundation for Basic Research (project no. 98-02-17052). REFERENCES 1. G. M. Batanov, V. A. Ivanov, M. E. Konyzhev, and A. A. Letunov, Pis’ma Zh. Éksp. Teor. Fiz. 66, 163 (1997) [JETP Lett. 66, 170 (1997)]. 2. L. V. Grishin, A. A. Dorofeyuk, I. A. Kossyœ, et al., Tr. Fiz. Inst. Akad. Nauk SSSR 92, 82 (1977). 3. G. M. Batanov, V. A. Ivanov, and M. E. Konyzhev, Pis’ma Zh. Éksp. Teor. Fiz. 59, 655 (1994) [JETP Lett. 59, 690 (1994)]. 4. L. D. Landau and E. M. Lifshitz, Course of Theoretical Physics, Vol. 6: Fluid Mechanics (Nauka, Moscow, 1986; Pergamon, New York, 1987). 5. D. S. Dorozhkina and V. E. Semenov, Pis’ma Zh. Éksp. Teor. Fiz. 67, 543 (1998) [JETP Lett. 67, 573 (1998)]. 6. A. V. Ignat’ev and A. A. Rukhadze, Fiz. Plazmy 9, 1317 (1983) [Sov. J. Plasma Phys. 9, 760 (1983)].

JOURNAL OF EXPERIMENTAL AND THEORETICAL PHYSICS

Translated by H. Bronstein

Vol. 92

No. 6

2001