Experimental Evidence of Charge Separation (Double Layer) in Laser-Produced Plasmas

132

IEEE TRANSACTIONS ON PLASMA SCIENCE, VOL. PS-13, NO. 3, JUNE 1985

Experimental Evidence of Charge Separation (Double Layer) in Laser-Produced Plasmas A. LUDMIRSKY, S. ELIEZER, B. ARAD, A. BOROWITZ, Y. GAZIT, S. JACKEL, D. SALZMANN, AND H. SZICHMAN

Abstract-Simultaneous measurements of the plasma target potential and plasma charged particle currents have been made for Nd laser irradiances between 4 X 1012 and 1015 W/cm2. The results appear to give the first direct indication of double layers in laser-produced plasmas.

DOUBLE LAYERS, which consist of two thin adjacent regions of opposite excess charges, have been observed in a variety of laboratory plasmas [11-[4]. These regions of positive and negative charges are, in general, separated by distances of the order of several Debye lengths and induce a large potential drop which can accelerate particles to high energies. Double layers have been thought to exist in laserproduced plasmas as well [5], [6]. However, no experimental evidence of their existence in such plasmas has been obtained to date, mainly because of diagnostic difficulties stemming from the very rapid time variation of the plasma and its small size. Probes cannot, in general, be used, though current has been measured in a laser-produced plasma using a probe embedded in the target [7]. In the present paper we describe a new technique for simultaneously measuring the target potential and the net current outflow from a laser-produced plasma. We believe the results indicate evidence of the existence of double layers in such plasmas. Our Nd: glass laser-plasma facility [8] is capable of delivering Gaussian-shaped 3.5-ns full width at half-maximum (FWHM) pulses of up to 10 J. Power densities of 4 X 1012_1015 W/cm2 were obtained by changing the laser energy and keeping the spot diameter, focused by an f/10 aspherical lens, constant at 20 pm (FWHM). The focal-intensity distribution was measured in the far field (focus) using a 15-m-length lens. About 150 experimental shots were analyzed. The experimental setup, which is shown in Fig. 1, includes an aluminum target which is used as one electrode of the capacitor C1. The dimensions of the target were varied from 2.5 to 10 mm in diameter and from 1 to 3 mm in thickness. The potential of the target V was measured by a capacitor voltage divider (C1 : C2 = 1: 16 000 with C2 = 500 pF) which had a high-frequency response of 1 GHz. The calibration was performed in situ using a nanosecond pulser and a 7904 Tektronix oscilloscope. Potentials of up to 100 kV were measured. The current probe (Fig. 2) is effectively a pulse self-inte-

AARON D.

KRUMBEIN,

i(t) Dielectric

Current / probe

Ci \

Laser beam

Electron flow direction for positive output signal i(t) Fig. 1. The experimental confilguration. The target T, the current probe, and the voltage divider (C1 and C2) are shown.

25

-Bross shield

a50Q BNC output

Fig. 2. Planar view of the current probe. All dimensions are in milimeters.

grating current transfonner [9] - [11] which measures the charged outflow from the plasma toward the laser beam. It consists of a coil 2.5 mm in thickness with an opening 1 cm in diameter and is placed at distances of 1.0 or 2.3 cm from the target. The coil is connected to a 50-2 output resistor and is surrounded by a brass shield 7.0 mm thick. The shield has a circumferential gap in order to eliminate the possibility Manuscript received April 2, 1984; revised November 19, 1984. The authors are with the Plasma Physics Department, Soreq Nuclear of undesirable currents flowing in it. It has a high-frequency response of -0.5 GHz and is shielded well enough to reduce Research Center, Yavne 70600, Israel.

0093-3813/85/0600-0132$01.00 © 1985 IEEE

LUDMIRSKY et al.: CHARGE SEPARATION IN LASER-PRODUCED PLASMAS

133

V

Fig. 3. Drawing of an actual calibration trace of the current probe. The upper trace is the input current flow from the pulse generator. The ordinate here is 40 mA/div. The lower trace is the output voltage from the current probe. The ordinate here is 0.1 V/div. The time scale (abscissa) is 1 ns/div.

the pickup signals to an undetectably low level. Currents of up to 5.5 A were measured. The current probe is shielded against all currents except those passing through the coil opening. Therefore the only contribution from a self-generated magnetic field produced by (Vn X VT) would occur if both Vn and VT were in the plane of the coil. Here Vn and VT are the plasma density and temperature gradients, respectively. However, from symmetry, Vn in that plane averages to zero so that there should be no contribution from that source. The calibration of the current probe (Fig. 3) was performed with a resistor of special design which was used as a load for a nanosecond pulser. Different target designs and dielectrics were used in order to exclude the possible influence of X-rays, secondary electrons, and polarization of the dielectric material. In addition, the measurements were performed not only in vacuum but also with varying values of air pressure in the experimental chamber. The voltage and current signals were displayed on a Tektronix 7844 dual-beam osciloscope and recorded on videotape by means of an ISIT-TV camera system [121. Fig. 4(a) shows a typical example of the temporal correlation between the incident laser pulse and the potential signal V on the target. The laser pulse displays the Gaussian shape with FWHM = 3.5 ns as mentioned previously. The laser intensity IL was 9.5 X 1014 W/cm2. The correlation between the time traces of the laser and voltage signals was made by first splitting the incident laser beam by means of a beam splitter and recording the two pulses with photodiodes. One photodiode was positioned in the target chamber in place of the target assembly and the second, or reference photodiode, was set in a fixed position outside the chamber. The two pulses were then matched so that their time of origin was the same. For the actual measurements, the diode in the target chamber was replaced by the actual target assembly and the reference diode was kept in place. The voltage starts with a negative polarity almost simultaneously with the beginning of the laser pulse. It changes polarity after 1.5 ns, increasing to a high positive value and then decaying slowly with the decay time of the diagnostic system. The temporal correlation between the voltage signal V and the charged particle flow I are shown in Fig. 4(b). The laser intensity for this measurement was the same as for the pre-

(a)

_ /,RX'

\~

(b)

Fig. 4. (a) Drawing of an oscillogram showing a typical target-potential trace (V) with the actual laser-pulse trace (L) superimposed. The voltage ordinate is 41 kV/div. The height of the laser pulse is relative and sized so as to fit on the oscillogram. Its actual time variation has been reproduced, however. (b) Drawing of an oscillogram showing actual charged current (I) and corresponding target-potential (V) traces from a typical laser shot. The voltage ordinate is 41 kV/div while the current ordinate is 3 A/div. In both (a) and (b) the time scale (abscissa) is 1 ns/div. vious oscillogram (Fig. 4(a)) and, in fact, the peak voltages are the same. The variation in the shape from pulse to pulse always appears to occur in experiments of this type and is apparently due to noncontrollable differences from shot to shot. At such large distances from the target, the negative part of the current signal is believed to correspond to a net ion expansion from the target towards the laser, and the positive current signal to a net electron-charge layer moving in the same direction. In all the recorded current signals, the negative portion of the current signal always precedes the positive portion (see

Fig. 4(b)). The change in polarity of the current signal with time appears to be a clear indication of charge separation in the plasma as it moves through the probe. This would appear to suggest the existence of a double layer in the plasma. Since basically the same types of signals were seen with laser intensities well below 1013 W/cm2, where suprathermal electrons are not produced, it seems highly unlikely that such electrons could be the cause of our observations. A systematic analysis was made of the extremum values of the negative and positive voltages as well as of the negative and positive currents. Fig. 5 shows a plot of the two voltage values and the positive current values all plotted as a function of laser irradiance. The negative-current curve is practically identical to the plotted positive current. The curves are remarkably parallel within the almost 3 orders of magnitude of the laser irradiances used in this experiment. For laser irradiances from about 1012 to about 1014 W/cm2, the slope is

134

IEEE TRANSACTIONS ON PLASMA SCIENCE, VOL. PS-13, NO. 3, JUNE 1985 100

potential and current plasma flow have been made on alumi-

num targets at laser irradiances between 4 X 1012 and 1015

0

, 3

W/cm2. The results suggest that double layers may be generated in laser-produced plasmas. Particularly, the current signals observed can only be obtained when the space charge is separated in the double layer and it moves through the probe. If present, double layers can accelerate ions [6] and electrons on one side of the plasma while stopping faster electrons moving inside the target on the other side of the plasma. In addition, double layers might strongly influence the plasma-density profile, which in turn can influence the laser absorption process and transport phenomena. Finally, they can play a role in the formation and perpetuation of wave instabilities. REFERENCES

[ 1 J. S. Levine and F. W. Crawford, J. Plasma Phys., vol. 24, p. 359, 1980. [2] S. Towen and L. Lindberg, J. Phys. D, vol. 13, p. 2285, 1980. IL (W/cm2) [31 R. L. Stenzel, M. Ooyama, and Y. Nakamura, Phys. Fluids, vol. Fig. 5. Positive (curve (a)) and negative (curve (b)) target potentials 24, p. 708, 1981. and the positive current (curve (c)) versus the laser irradiance IL. [4] N. Sato, R. Hatakuyama, S. lizuka, T. Mieno, K. Saeki, J. Rasmussen, and P. Michelson, Phys. Rev. Lett., vol. 46, p. 1330, 1981. nearly equal to one, while for higher irradiances the slope H. Hora, Laser Plasmas and Nuclear Energy. New York: Plenum, [51 flattens to about 0.4. 1975. Using a simple model [13], a scaling law for measured po- [61 H. Hora and P. Lalousis, Laser Part. Beams, vol. 1, p. 283, 1983. G. Drouet and R. Bolton, Phys. Rev. Lett., vol. 36, p. 591, tential as well as for measured current as a function of laser ir- [7] M. 1976. radiance was derived. The scaling laws for both voltage and [8] S. Jackel, H. M. Loebenstein, A. Zigler, H. Zmora, and S. Zweigcurrent turn out to be the same. They are in good agreement enbaum, J. Phys. E, vol. 13, p. 995, 1980. with the experimental results shown in Fig. S if one assumes [9] A. M. Stefanovski, Sov. Instrum. Exp. Tech., vol. 10, p. 375, 1967. a breakpoint in laser irradiance at about 1014 W/cm2. Below [10] D. G. Pellinen and P. W. Spence, Rev. Sci. Instr., vol. 42, p. 1699, this value the assumption of laser absorption by inverse brems1971. S. B. Vasserman, Sov. Instrum. Exp. Tech., vol. 15, p. 415, 1972. strahlung is made, while above this value absorption is assumed [11] B. [121 Arad etal., J. Phys. E, vol. 15, p. 1223, 1982. to occur by resonance absorption. [13] S. Eliezer and A. Ludmirsky, Laser Part. Beams, vol. 1, p. 251, To summarize, simultaneous measurements of plasma target 1983.