Low-Energy Plasma Focus as a Tailored X-Ray Source

Journal of Fusion Energy, Vol. 19, No. 2, June 2000 (䉷 2002)

Low-Energy Plasma Focus as a Tailored X-Ray Source M. Zakaullah,1 K. Alamgir,1 M. Shafiq,1 M. Sharif,1 A. Waheed,2 and G. Murtaza3

A low-energy (2.3 kJ) plasma focus energized by a single 32-␮F capacitor charged at 12 kV with filling gases hydrogen, neon, and argon is investigated as an X-ray source. Experiments are conducted with a copper and an aluminum anode. Specifically, attention is given to tailoring the radiation in different windows, e.g., 1.2–1.3 keV, 1.3–1.5 keV, 2.5–5 keV, and Cu-K␣ line radiation. The highest X-ray emission is observed with neon filling and the copper anode in the 1.2–1.3 keV window, which we speculate to be generated due to recombination of hydrogenlike neon ions with a few eV to a few 10s of eV electrons. The wall-plug efficiency of the device is found to be 4%. The other significant emission occurs with hydrogen filling, which exhibits wall-plug efficiency of 1.7% for overall X-ray emission and 0.35% for Cu-K␣ line radiation. The emission is dominated by the interaction of electrons in the current sheath with the anode tip. The emission with the aluminum anode and hydrogen filling is up to 10 J, which corresponds to wall-plug efficiency of 0.4%. The X-ray emission with argon filling is less significant.

KEY WORDS: Plasma focus; tailored X-ray source.

1. INTRODUCTION

per shot in the energy range of 2.5 keV. They reported a temperature of about 1.5 keV and the size of the micropinches as 40–250 ␮m. It was suggested that by using other gases like neon and xenon, the total radiation could be increased. Liu et al. [3] examined the performance of plasma focus with 14-kV charging voltage and 2.9-kJ stored energy operated with neon and measured the X-ray yield of 6 J/shot in 4␲ steridian. Beg et al. [4] studied 2-kJ, 200-kA plasma focus using a gas having Z ⱕ 18 and reported the X-ray yield. By using the Ross filter technique, a maximum X-ray yield of 16.6 J of 0.7–1.5 keV photons with a pulse width of 10–15 ns was reported when neon was used as the filling gas. A time-integrated pinhole image showed a plasma column under the soft filter, while hot spot structure was observed under the hard filter. This behavior indicated the transition of the plasma from a column to a series of hot spots. Lee et al. [5] studied the plasma focus with peak current of 320 kA when the capacitor bank (4 ⫻ 7.8 ␮f) was charged to 14 kV. They reported 100 J of soft

There is continued interest in operating plasma focus with different gases for higher soft X-ray yield. Lebert et al. [1] suggested plasma focus devices for compact ultraviolet and soft X-ray sources. It was reported that with correct optimization, both continuous radiation and narrow-band line radiation could be tailored to specific applications. In particular, X-ray lithography, X-ray microscopy, X-ray contact microscopy, and X-ray photoelectron spectroscopy were demonstrated. It was concluded that, in principle, nearly contradictory demands could be met with the same source. Bayley et al. [2] studied the plasma focus by injecting argon through a hole in the anode. They found that the micropinches emit several joules of soft X-rays 1

Department of Physics Quaid-i-Azam University 45320 Islamabad, Pakistan. 2 PINSTECH, P.O. Box 2151, 44000 Islamabad, Pakistan. 3 Salam Chair in Physics, Government College, 54000 Lahore, Pakistan.

143 0164-0313/00/0600-0143/0 䉷 2002 Plenum Publishing Corporation

144 X-rays per shot, a 4% wall-plug efficiency in 4␲ -geometry, which may be defined as the ratio of the energy emitted as X-rays in 4␲ -geometry to the energy stored in the capacitor bank. They also studied another system with 12 capacitors, each having 0.6-␮f capacitance with peak current of 400 kA at 11.5-kV charging voltage into the water-cooled electrode, giving 18-J X-rays with 1% wall-plug efficiency. Keeping the working pressure and charging voltage constant, the variation in the X-ray yield was not more than 0.35%. A sufficient flux was generated in 300 shots for recording an X-ray lithograph for demonstration. Lee et al. [6] demonstrated a 1.6-kJ plasma focus as an electron source for microlithography. Resolution better than 0.5 ␮m was obtained. The total energy in the beam and the energy of electrons was estimated from the lithograph to be greater than 20 mJ per shot, with energy higher than 20 keV, and ⬎1 J with more than 10 keV of energy. Filippov et al. [7] reported that a plasma focus with a Filippov-type electrode could convert 10% of the capacitor bank energy into K-shell lines of Ne radiation. This result, when compared to the data of a plasma focus operated at 5 kJ, indicates a Y ⬃ I p3.5–4 radiation scaling and is in conformity with the resistive heating mechanism of Ne plasma. Measurement of the X-ray energy flux was carried out with specially developed sensors, based on the explosive evaporation of the filtered targets. Lebert et al. [8] studied the X-ray emission from pinch plasma devices with pinch currents ranging from 200–400 kA operated with pure high-Z gases with temporal, spatial, and spectral resolution. If operated using elements having Z ⬍ 18, K-shell emission is observed from columnlike structures having a volume of several hundred micrometers in diameter and several millimeters in length. For Z ⬎ 18, emission with h␯ ⬎ 1 keV is observed from micropinches. For argon, both modes of operation can be observed. The occurrence of a specific mode depends on the initial gas pressure. The transition regime between column and micropinches was investigated for argon. Bourgade et al. [9] modified a plasma focus device (27 kJ, 40 kV, and 400 kA) in order to enhance soft X-ray emission. A hollow anode was used to prevent hard X-ray bremsstrahlung emission, and a lot of filling gases were chosen for their characteristic lines in the keV and sub-keV ranges. The X-ray energy emitted from neon gas discharge was about 20 J/4␲. A similar X-ray yield in Hand He-like ion emission lines was measured with filling gases O2, N2, and CH4. A significantly high yield (250 J/4␲ ) was measured in the sub-keV range. Hirano and Kitaoka [10] investigated soft X-ray emission from a plasma focus in which 1 MA was discharged into an

Zakaullah et al. argon admixture with hydrogen. A time-resolved soft ˚ K-lines and dynamic behavX-ray pinhole image near 4 A ior of the plasma were observed simultaneously. It was found that the soft X-ray images could be typically divided into three categories. This fact made the soft X-ray emission from the plasma focus device not reproducible. One category showed a highly conspicuous and filamentary source whose radius was 0.5 mm, whereas the radius observed by the Moire´-Schlieren technique was 1.5 mm. The filament had a brightness several times greater than that of the other categories. The plasma was disrupted by growth of the m ⫽ 0 instabilities. The plasma sustaining time was shorter than the characteristic time for radiation collapse. The Pease-Braginskii current was evaluated to be 0.8 MA. It was concluded that the radiative collapse did not occur, although the discharge current exceeded the Pease-Braginskii current. Chee Meng et al. [11] studied the variation of Xray emission in a low-energy plasma focus over a range of pressures. The working gases were an argon and hydrogen mixture. Soft X-rays originating from the plasma and from electron-beam activity on the copper anode were also observed. They identified three pressure regimes. In the first regime, both the plasma X-rays and the copper line radiation were weak. In the second regime, the X-ray emission was intense and the contribution from the copper lines was strong. In the third pressure regime, the plasma X-rays were intense while the contribution from the copper X-rays was weak. However, no attempt was made to obtain information about the X-ray yield. Bergmann et al. [12] investigated the scaling of the K-shell line emission of intermediate-valued atomic number pinch plasmas generated in devices with pinch currents ranging between 180–310 kA. It was shown that the transient ionization dynamics could be made visible in the scaling of the 1s–2p transitions of hydrogenlike ions and the 1s2 –1s2p transitions of the heliumlike ions. Kato et al. [13] developed a bright and reliable Xray source for lithography. The X-ray source size for neon was 1 mm in diameter and 10 mm in length. A 0.4 ␮m fine pattern was printed with this source. They obtained an X-ray intensity of 5 mJ/cm2/shot at a distance of 25 cm from the source with an irradiance of 10 mW/ cm2 at the 2-Hz repetition rate. It was concluded that the plasma focus is a promising X-ray source for lithography from the viewpoint of intensity, resolution, and lifetime. This paper presents the results of a series of experiments to investigate the scope of plasma focus as an X-ray source in the 1–9 keV window. Specifically, the emission is studied with a (1) Cu-anode and argon, neon, and hydrogen filling, and (2) Al-anode with hydrogen and neon filling.

Low-Energy Plasma Focus

145 every 5–6 shots the previous gas was purged out and fresh gas was filled. We have recorded 5–7 shots for each measurement at different filling pressures. For timeresolved X-ray measurements, the Quantrad Si pin-diodes of 125 ␮m active-layer thickness along with suitable absorption filters were used. These detectors were normalized against each other by masking each with an identical filter. To keep the X-ray flux reaching the diode to a safe level, a 0.6-mm thick brass sheet with 2-mm diameter holes at the center was used to cover the diode. The detectors were placed at a distance of 17.5 ⫾ 0.2 cm from the anode axis, and at a height of 1.5 ⫾ 0.1 cm from the anode tip. The electrical signals from the X-ray detectors along with the high-voltage (HV) probe signal were recorded by a four-channel 200-MHz Gould 4074A digital storage oscilloscope; hard copies of the oscillogram were plotted using a HP7475 plotter.

Fig. 1. The schematic of experimental setup and different diagnostics.

In Section 2, the experimental setup and diagnostics are described. Section 3 presents the experiment with argon gas, Section 4 with neon gas, and Cu electrodes, and Section 5 depicts the results of a copper anode with hydrogen filling, by enhancing the interaction of the current sheath with the anode tip. Section 6 describes experiments with an anode made of aluminum, with both hydrogen and neon as the filling gases. Section 8 discusses the results and concludes the study.

2. EXPERIMENTAL SETUP AND DIAGNOSTICS In this experiment we have used a Mather-type Plasma Focus system, which is reported elsewhere in detail [14–15]. It is energized by a 32-␮F, 15-kV (3.6 kJ) single capacitor, charged at 12 kV (2.3 kJ), giving peak discharge current of about 190 kA. A triggertrontype pressurized sparkgap is used as a switch. The schematic arrangement of the electrode system made of copper and the diagnostics is given in Fig. 1. The electrode system is comprised of a copper rod 152 mm in length and 18 mm in diameter as the anode. Six copper rods of 9 mm in diameter surround it such that the ratio of cathode to anode radii is 3.2. The experiments were conducted with a conventional cylindrical anode, as well as with one slightly tapered toward the open end [16]. The plasma focus chamber was evacuated up to 10⫺2 mbar using a rotary vane pump. To reduce the impurity effect, after

3. ARGON K-SERIES LINE RADIATION Using argon as the filling gas, attention was focused on finding the pressure range for the highest argon K-series line emission. Specifically, the radiation emission in 4␲ -geometry and the efficiency of the system to this end were estimated. This experiment was conducted with a conventional cylindrical anode, which was engraved about 25 mm at the open end to minimize its interaction with the current sheath. Selection of Filters For spectral analysis of X-rays emitted from hot plasmas, one may use dispersive techniques, which are based on delicate equipment. However, simple spectral analysis may be conducted by using a Ross filter pair of appropriate absorbers along with detectors like pindiodes. When X-rays pass through foils of different material, the intensity of radiation is attenuated according to the relation, I ⫽ I0e⫺␮(E)t, where I0 is the intensity of incident radiation flux, I is the intensity of transmitted flux, and ␮(E ) is the absorption coefficient of the filter. The absorption coefficient ␮ is a function of energy, as the high-energy photons experience less absorption whereas low-energy photons experience high absorption. Further, the transmission of photons undergoes a sudden jump when the energy of incident photons equals the ionization energy from K, L, M, etc. energy levels of the filter material. Thus, every element exhibits transmission windows. This property in different elements was exploited to select a pair of filters for detecting radiation in a narrow

146

Zakaullah et al.

Fig. 3. X-ray emission (area under the curve of the signals) recorded by the X-ray detectors with Ti and Mo filters.

2.95–4.4 keV. One may estimate the line radiation emission in the said energy interval by the relation Y⫽

4␲Qexp d⍀S(E )T(E )

(1)

where ● Qexp ⫽

Fig. 2. (a) The transmission of Ti (37.5 ␮m) and Mo (20 ␮m) Ross filter set; (b) response of X-ray detectors along with respective filters.

band, known in the literature as Ross filters [17]. By reviewing the transmission windows of different commercially available filters, the Ross filter pair selected was Mo (20 ␮m) and Ti (37.5 ␮m). The transmission curves are presented in Fig. 2a. We used Quantrad Si pin-diodes with an active-layer thickness of 125 ␮m. The detectors’ response, along with transmission characteristics of the filters, is given in Fig. 2b. For the evaluation of the curves presented in Fig. 2, data for absorption coefficients was taken from the Handbook of Spectroscopy [18]. The energy of the atomic argon K␣ line is 2.95 keV [19], whereas the ionization potential of Ar XVIII is 4.43 keV [20]. The energy of K␣ lines arising from the multiply stripped argon ions will lie within the energy interval of

冮 V R (Coulombs) dt

● d⍀ is the solid angle subtended by the detector at the focus region. ● S(E ) is the average sensitivity of the detector in the 2.9–4.4 keV interval. Its value is taken to be 0.2 Coulomb per joule (from the Quantrad brochure). ● T(E ) is the average transmission of the filter in the said interval; one may select 0.1, the FWHM value of the transmission curve. ● R ⫽ 50 ⍀ in the present experiment.

Results Figure 3 shows [21] the variation of signal intensity (area under the curve) recorded by the detectors masked with Ti and Mo filters, respectively. The detector with the Mo filter records higher intensity at filling pressures less than 1 mbar, whereas the situation is reversed for filling pressures exceeding 1 mbar. This anomaly may be resolved if one concentrates on Fig. 2b. The response of the detector with the Mo filter is slightly higher for 7–13 keV. This energy range covers the Cu-K line series.

Low-Energy Plasma Focus

147

Fig. 4. Variation of argon K-series line radiation yield and the system efficiency with filling pressure.

Zakaullah et al. [16] reported that the X-radiation emission from a plasma focus is much higher at pressures less than 1 mbar and that the radiation is generated mainly by the interaction of fast electrons in the current sheath with the anode rim. The field of view of the detectors in the experiment covers the anode rim, and hence the radiation, which consists of Cu-K-series line radiation, is recorded. For pressures exceeding 1 mbar, the intensity of the detector with the Ti filter surpasses that of the detector with the Mo filter, which indicates the filling pressure region for argon line radiation emission in this device. In this arrangement, the difference in XRD signals under the Ti filter and the Mo filter corresponds to Xrays with a dominant contribution of argon line radiation. Figure 4 presents the line radiation yield and efficiency of plasma focus for argon line radiation emission. The highest line radiation yield is recorded at a filling pressure of 1.5 mbar, which is about 30 mJ in 4␲ -geometry. The total discharge energy per shot in this experiment is 2.3 kJ. One may evaluate the system efficiency for line radiation, which comes out to be approximately 0.0015%. The X-ray emission of photons of energy ⬎3 keV in 4␲ -geometry at 0.5 mbar is recorded at about 0.7 J, and the machine efficiency at about 0.028%. The X-ray emission in this environment is not predominantly from hot plasma, but rather from the interaction of energetic electrons either in the current sheath or the electron beam generated in the focus region [22–23]. Note that in this experiment, the anode is engraved about 5 cm deep to minimize the interaction of energetic electrons with the anode. To identify the regions of Cu line radiation emission

Fig. 5. Variation of signal intensity at different axial positions and argon filling pressures: (a) with 10-␮m Co filter, (b) with 12.5-␮m Ni filter.

along the focus axis, and the variation of its intensity at different filling pressures, three pin-diodes are placed at a height of 0.1, 2.6, and 5.1 cm from the anode tip. In front of each pin-diode, a brass barrel 6.5 cm in length 2.5 mm in diameter is fitted so that only the X-rays emitted just in front of it reach the detectors. The data is recorded by masking the detectors with a 12.5-␮m Ni filter, and then again with a 10-␮m Co filter. Figure 5 depicts signal intensity at different positions and filling pressures. One can see that the Cu line radiation emission, which is the difference of signal intensity with Ni and Co filters, is the highest close to the anode and that about 38% of the emission is Cu-K␣ radiation.

148

Zakaullah et al.

Table I. Neon Ionic States of Interest and Their Emission Around 1 keV Ionization state Ne IX Ne Ne Ne Ne Ne Ne

IX X X X X X

Configuration 1S2 –1S 2P or 1S2 –1S 2P Ionization energy 1S–2P 1S–3P 1S–4P 1S–5P Ionization energy

Wavelength ˚ A

Energy (keV)

13.44 13.55

0.9226 0.9151 1.196 1.0219 1.2111 1.2773 1.3079 1.362

12.134 10.239 9.708 9.481

4. NEON RADIATION This experiment was carried out using neon as a filling gas and with two anode configurations, the conventional cylindrical anode and one tapered toward the open end.

Fig. 6. Response of detectors along with different filters for emission from neon gas plasma.

Selection of Filters

Results

Table I presents X-ray line emissions from different neon ionic species. The energy of K␣ lines from neutral neon atoms (Ne I) is 0.8486 [19] and from heliumlike ˚ ) keV or 0.9151 keV. Simiions (Ne IX) 0.9226 (13.44 A larly, the K␣ energy of hydrogenlike ions (Ne X) is 1.0219 ˚ ) [20]. The ionization energy of heliumlike keV (12.134 A ions (NE IX) is 1.196 keV and that of hydrogenlike ions (Ne X) is 1.362 keV. In this experiment, to detect the Xrays, the pairs of Ross filters selected are Mg (100 ␮m), Al (50 ␮m), Co (20 ␮m), and Ni (17.5 ␮m), along with Quantrad Si pin-diodes with a 125-␮m-thick active layer. Using X-ray absorption data from the Handbook of Spectroscopy [18], the response of detectors with the different filters is evaluated and given in Fig. 6. The absorption edges of Mg and Al lie at 1.305 keV and 1.56 keV, respectively. The detector with the Mg filter offers a transmission window from 1.2–1.30 keV and with the Al filter from 1.2–1.56 keV. Therefore, the difference in detector signals (area under the curve) with Mg and Ni filters corresponds to radiation of 1.2–1.3 keV. Similarly, the difference in signals with Al and Mg filters corresponds to soft X-rays in the 1.3–1.56 keV range. The Ni filter allows the transmission of Cu-K␣ radiation, which is 8.047 keV in energy but effectively stops the soft X-rays ⱕ 4 keV in energy. The transmission windows of Co (20 ␮m) and Ni (17.5 ␮m) are almost identical, except that the former discriminates and the latter allows the transmission of Cu-K␣ radiation.

Figure 7 presents the variation of X-ray emissions vis-a`-vis the neon gas filling pressure. For a conventional cylindrical anode, the emission in the 1.2–1.3 keV range is the highest and approaches 1.2 J/sr. In the 1.3–1.56 keV range, the highest emission is 20–22 mJ/sr., whereas the Cu-K␣ emission is just 10–12 mJ/sr. For the tapered anode, the X-radiation emission is much higher. It is in excess of 5.5 J/sr. in the 1.2–1.3 keV range, 80mJ/sr. in the 1.3–1.56 keV window, and 15–17 mJ/sr. for the CuK␣ line radiation. The radial collapse of the current sheath in front of the anode rapidly changes the system inductance, which, in turn, generates high voltage. The voltage may be higher than the charging voltage of the capacitor bank. The HV probe may be used to monitor the developed transient voltage and hence help to examine the quality of the focused plasma filament. One may consider the event of the HV probe spike as the time of maximum compression. The information about the delay of X-ray emission with reference to the HV probe spike could be interesting, and that is presented in Fig. 8. The error bars are not added to avoid overlapping. There is a mismatch in the Mg and Ni filters’ transmission in the 8–20 keV region, just after the Ni absorption edge. To confirm that there is no appreciable emission in this interval, an experiment with the tapered anode was conducted by masking the pin-diode with 10-␮m Mo and 30-␮m Ni filters. X-ray emission in this interval was found to be just 2 mJ/sr., which confirmed that the

Low-Energy Plasma Focus

149

Fig. 8. Time delay of main X-ray pulse with the HV probe spike.

Fig. 7. Variation of X-ray emission in different energy windows visa`-vis the neon filling pressure for cylindrical and tapered anode configurations.

consideration is justified for the difference between the Mg and Ni filters corresponds to radiation of 1.2–1.3 keV.

The highest X-ray emission in this experiment was observed in the energy range of 1.2–1.3 keV. The previous experimental results from different laboratories [3–8] reported that the radiation from plasma focus operated with neon gas is dominantly line radiation, specifically Ly-␣ and He-␣ [3]. As described in Table I, the energy ˚ ) and of the Ly-␣ (Ne-X) line is 1.0219 keV (12.134 A ˚ the He-␣ (Ne-IX) line was 0.9226 keV (13.44 A). The result of this experiment was not in agreement with the previous work. The ionization energy of Ne-IX was 1.196 keV. Recombination of a few eV to a few 10s of eV electrons with the Ne-IX ions will emit photons in the 1.2–1.3 keV range. This hypothesis is further supported by the observation that the X-rays at 3.0–3.5 mbar operating pressure are emitted 45 ⫾ 5 ns after the event of maximum compression. A typical oscillogram at 3.0 mbar Ne pressure is given in Fig. 9. Figure 8 shows that the delay of the X-ray pulse with reference to the HV probe signal changes with the change in operating pressure, from ⫺50 ns to ⫹200 ns. It is speculated that X-ray emission before the maximum compression and formation of focus plasma is due to interaction of the current sheath with the anode tip. The emission at the event of maximum compression is bremsstrahlung and comes from hot plasma. The emission after maximum compression corresponds to recombination radiation. At a filling pressure of 5–6 mbar, this delay approaches zero. Probably, at high pressure the recombination becomes dominant earlier. However, there is a possibility for

150

Zakaullah et al.

Fig. 9. A typical oscillogram of X-ray signals with Mg (100 ␮m), Al (50 ␮m), and Ni (17.5 ␮m) filters along with HV probe spike at 3.0 mbar Ne filling.

plasma heating even after the maximum compression due to the energetic electron beam. In addition, time-resolved measurement along with simultaneous electron beam detection may resolve the paradox. In Fig. 9, a very small spike before the main X-ray pulse is seen, which synchronizes with the HV probe spike. The Ross filter pair used in the experiment, Mg (100 ␮m) and Al (50 ␮m), will discriminate the Ly-␣ and He-␣ line radiation. One may speculate that the small spike in the X-ray signal at the event of the HV probe spike is due to the above-mentioned line radiation. The experiment was repeated with Mg (25 ␮m) and Al (12 ␮m) filters, which offer a transmission window from about 0.8 keV. Now, the first peak became significant also. The line radiation emission at the event of the HV probe spike, which is thought to be due to He-like K␣ line radiation, at different filling pressures is plotted in Fig. 10. The total X-ray emission in 4␲-geometry and the plasma focus efficiency for total X-ray generation was also evaluated and results are presented in Fig. 11. Figure 12 depicts the X-ray pinhole images recorded with 200-␮m pinholes, masked with 25-␮m Mg, 12-␮m Al, and 5-␮m Ni filters, with both tapered and cylindrical anodes. The height of the pinhole aperture is adjusted to 5 mm below the cylindrical anode tip and 5 mm above the tapered anode tip. This mounting helps to minimize the entrance of those X-rays in the pinhole camera, that are emitted from the anode tip through the thick, target bremsstrahlung mechanism. In the case of a cylindrical anode, a well-defined, approximately 5–6-mm-long pinch plasma column is seen. Under Al (12 ␮m) and Mg (25 ␮m) filters, the

Fig. 10. Ne radiation emission that synchronizes with the HV probe spike vis-a`-vis filling pressure for cylindrical and tapered anodes. It is speculated that the emission is from the heliumlike K␣ line, 0.9226 ˚ ) or 0.9151 keV (13.55 A ˚ ). keV (13.44 A

plasma column is bright and highly unstable. The instability structure is not symmetrical about the pinch axis. It reflects that Rayleigh–Taylor-type instabilities are active. Under the Ni (5 ␮m) filter, an elongated faint spot is visible. Thus, photons of energy ⱖ4 keV are emitted from that region. It is not clear with the present diagnostics whether a part of the pinch region gets preferentially heated due to some anomalous mechanism. There is also a possibility that the sputtered copper from the anode tip is swept to this region and is responsible for X-rays of ⱖ4 keV. In the case of the tapered anode, the focus filament is much larger. The X-ray yield is higher by an order of magnitude. The total X-ray emission in 4␲ geometry exceeds 80 J with the tapered anode, which corresponds to an efficiency of 4%. In the same system with a conventional cylindrical anode, the emission drops to 7 J, with wall-plug efficiency of 0.35%. If one compares with other low-energy systems, it is 6 J per shot for input energy of 2.9 kJ [3], 100 J for input energy of 3 kJ [5], and 16 J for input energy of 2 kJ [4]. The different parameters of these machines are summarized in Table II. It is evident that the X-ray yield varies significantly with changes in different device parameters. A comprehensive study to optimize the interrelated parameters is essential for the use of plasma focus as an intense X-ray source for different applications. For the tapered anode, the emission is much higher. This source may be

Low-Energy Plasma Focus

151

Fig. 12. X-ray pinhole images with 200-␮m-diameter apertures masked with Mg (25 ␮m), Al (12 ␮m), and Ni (5 ␮m) filters for (a) tapered anode, (b) cylindrical anode.

Fig. 11. Total X-ray emission and the system efficiency of the plasma focus with the neon gas for (a) tapered anode, (b) cylindrical anode.

of interest for different applications like X-ray backlighting, X-ray contact microscopy, and lithography for electronic device manufacturing. In this experiment, the spectral response of pindiodes was evaluated using the X-ray absorption data [18]. Upon comparison with the experimental curves provided by the Quantrad, an excellent agreement is recorded. The results of measurement with the cylindrical anode record the X-ray yield, which is consistent with that of previous studies [4,8–9]. Obviously, there is no reason to speculate that the measurement with the tapered anode may not be correct.

5. Cu-K␣ EMISSION This experiment was conducted with the tapered anode, which is not engraved at the open end. This modification enhances the interaction of electrons in the current

Table II. X-Ray Yield and Parameters of Some Low-Energy Plasma Focus Devices

Experiment 1. Tapered anode 2. Cylindrical anode Liu et al. [3] Lee et al. [5] Beg et al. [4]

Soft Peak X-ray Input discharge Charging Parasitic yield energy current voltage inductance 80 J 7.2 J 6J 100 J 16 J

2.3 2.3 2.9 1.9 2

kJ kJ kJ kJ kJ

190 190 200 400 200

kA kA kA kA kA

12 kV 12 kV 14 kV 11.5 kV 40 kV

80 nH 80 nH 110 nH ? ?

sheath or the electron beam in the focus region with the anode tip. As described in Sections 3 and 4 earlier, the interaction of energetic electrons with the anode tip, and hence the emission of Cu-K␣ radiation, is the highest at lower filling pressures. In an attempt to lower the mass density of the working gas and, hence, to operate the PF at the lowest possible pressure, the experiment was conducted with hydrogen as a filling gas.

152

Zakaullah et al.

Selection of Filters For conducting this experiment to measure Cu-K␣ radiation, the filters selected were Co (40 ␮m), Ni (42.5 ␮m), and Cu (42 ␮m), respectively. The absorption edge of the Co filter lies at 7.709 keV, which stops the Cu-K␣ line radiation of energy 8.047 keV, whereas the Ni filter with the absorption edge at 8.33 keV allows the transmission of Cu-K␣ line radiation but stops Cu-K␤ line radiation of energy 8.905 keV. The Cu filter has an absorption edge at 8.98 keV and allows the transmission of both K␣ and K␤ lines. Results Figure 13 presents the transmission curves of the respective filters along with the response of 125-␮m active layer Quantrad pin-diodes. The difference in response of the diodes with Ni and Co filters corresponds to Cu-K␣ radiation, whereas the difference in the signals with Cu and Ni filters may provide an estimate of CuK␤ radiation. Figure 14 presents the estimate of Cu-K␣ line radiation emission (J/sr.) and the number of Cu-K␣ photons/sr. at different filling pressures [25]. The highest average Cu-K␣ yield of 0.4 J/sr. was recorded at a pressure of 0.75 mbar, which is equivalent to 3 ⫻ 1014 Cu-K␣ photons/sr. Some shots also produced a yield of 0.7 J/sr. at this filling pressure. We have used Agfa X-ray dental film to study the X-ray fluence anisotropy. Four X-ray film detectors were mounted at 30⬚, 50⬚, 70⬚, and 90⬚ with respect to the anode axis, as shown in Fig. 1. Each light-tight covered film detector was masked with 10␮m Co, 12.5-␮m Ni, and 10-␮m Cu filters. The film detectors were exposed to X-rays for a single shot with the help of a Wilson seal and a shutter arrangement [14]. The thickness of the polyester base and the emulsion composition were analyzed by the scanning electron microscope (SEM). Similarly, the detailed composition of the light-tight film cover was also determined. Using the data [18], the response of the film cover along with different filters was determined and is presented in Fig. 15. The density of the film detector with reference to a standard density wedge was analyzed with the help of a He-Ne laser and photodetector, and the angular variation of the Cu-K␣ flux was determined. In this analysis, it was ensured that the variation of film density with log (I) was linear, where I corresponds to the incident radiation flux. The Cu-K␣ radiation flux increased just 2 times from the side-on direction (␪ ⫽ 90⬚) to ␪ ⫽ 30⬚ with respect to the anode axis. It was not possible to put a radiation detector in the end-on direction (␪ ⫽ 0⬚). The fluence anisotropy [26] of characteristic X-rays emitted from

Fig. 13. Transmission curves of the Cu (42 ␮m), Ni (42.5 ␮m), and Co (40 ␮m) filters along with the detector response.

thick targets is a function of incident electron energy. It may vary from 1.5 to 80 for a copper target. It is speculated that the X-ray generation is mainly due to 12-keV electrons in the current sheath, the charging voltage of the plasma focus. Probably the contribution of the energetic electron beam generated in the focus region for Cu-K␣ emission is not significant. From this data, one may infer that the X-ray emission or, at least, ratio of K␣ to continuum may increase with the increase in plasma focus operating voltage to 40 kV [27]. Using the X-ray flux in the side-on direction and the fluence anisotropy data, the Cu-K␣ emission in 4␲ -geometry is estimated to be about

Low-Energy Plasma Focus

153

Fig. 16. X-ray pinhole images with tapered copper anode at 0.75 mbar hydrogen pressure with 200-␮m-diameter aperture masked with Cu (10 ␮m), Ni (12.5 ␮m), and Co (10 ␮m) filters.

Fig. 14. Variation of Cu-K␣ radiation (J/sr.) and number of Cu-K␣ photon/sr. vis-a`-vis filling pressure.

Figure 16 presents the X-ray pinhole images at 0.75 mbar operating pressure, recorded with 200-␮m-diameter apertures. The pinholes were masked with Cu (10 ␮m), Ni (12.5 ␮m), and Co (10 ␮m) filters, respectively. From these images, one can see that the X-rays are emitted from the anode tip. No pinch plasma column is visible under these filters. It may not be surprising that the energy of X-rays from the focus filament is in the range of 1–1.5 keV [28], for which the filters used are opaque. The image under the Co filter is faint as compared to images under the Ni and Cu filters, which are almost identical. The Co filter stops Cu-K␣ line radiation, the Ni filter allows transmission of Cu-K␣ line radiation but stops CuK␤, and the Cu filter allows both the K␣ and K␤ lines but stops the transmission of high-energy radiation. Thus, the lower intensity of the image under the Co filter as compared to the images under the Ni and Cu filters represents the more significant contribution of Cu-K␣ radiation.

6. EXPERIMENT WITH ALUMINUM ANODE Fig. 15. Response of the film detectors masked with Co (10 ␮m), Ni (12.5 ␮m), and Cu (10 ␮m) filters along with 210-␮m light-tight film cover.

8 J. The ratio of Cu-K␣ to continuous radiation is also a function of incident electron energy, which varies from 0.05 to 0.75. It is about 0.2 for 12 keV incident electron energy. These observations reveal that up to 40 J of energy is being emitted as X-ray photons in 4␲ -geometry. This corresponds to a system efficiency of 1.7%. This efficiency is expected to increase further with the charging voltage. Efforts to estimate the contribution of Cu-K␤ to the emitted radiation flux remained inconclusive.

In the previous experiments described in Sections 3–5, the central electrode of the PF machine was made of copper, the metal commonly used for machine electrodes. This experiment was performed with an aluminum anode, and the X-ray emission data was recorded both with neon and hydrogen as the filling gas. The use of aluminum instead of copper was undertaken to enhance X-ray emission in the 1–1.5 keV window. Selection of Filters In order to detect the X-rays, the same pin-diode and filter thicknesses of Mg (100 ␮m), Al (50 ␮m), and Ni (17.5 ␮m) were used. The Mg (100 ␮m) and Al (50

154

Fig. 17. X-ray emission at different neon filling pressures with aluminum anode: (a) in the 1.2–1.3 keV window, (b) Al-K␣ line radiation.

␮m) filter pair helps to estimate the Al-K␣ line emission, since the absorption edge of Mg (100 ␮m) lies at 1.305 keV and that of Al (50 ␮m) lies at 1.56 keV. The difference in area under the curves of the Al and Mg filters gives Al-K␣, whereas the difference in area under the curves of Mg and Ni filters provides an estimate of soft X-rays excluding Al-K␣. Results Figure 17 summarizes the emission in the 1.2–1.3 keV and 1.3–1.5 keV windows for both the hydrogen and the neon filling; the latter (1.3–1.5 keV) presumably is the aluminum K␣ line radiation. For hydrogen filling, the highest emission in the 1.2–1.3 keV range approaches as high as 0.9 J/sr. at 1.5 mbar. This emission flux corresponds to about 10 J in 4␲ -geometry, in excess of 0.4%

Zakaullah et al.

Fig. 18. Total X-ray emission and generation efficiency of the system with an aluminum anode and hydrogen and neon as working gases.

wall-plug efficiency. The Al-K␣ line emission attains maximum value of about 120–130 mJ/sr. at 2.0 mbar hydrogen pressure. In the experiment with neon gas, the emission in the 1.2–1.3 keV window was the highest at 0.5 mbar, and equalled approximately 200 mJ/sr. The maximum Al-K␣ emission was about 60–65 mJ/sr. at a pressure of 0.5 mbar. The X-ray yield in 4␲ -geometry did not exceed 2.5 J, which is equivalent to 0.1% wallplug efficiency. Figure 18 presents the variation of total X-ray emission and X-ray generation efficiency vis-a`-vis neon and hydrogen filling pressures. Total X-ray emission is the highest at 0.5 mbar for neon filling and is equal to 2.5 J. For hydrogen filling, the total X-ray emission attains the highest value at about 10 J at 1.5 mbar pressure. The ratio of Al-K␣ line emission to total X-ray emission was found to be dependent on the filling pressure, which is depicted in Fig. 19. It is the highest for hydrogen and has the value of 0.6 at 0.5 mbar; it decreases steadily

Low-Energy Plasma Focus

155

Fig. 19. Ratio of Al-K␣ line emission to total X-ray emission vis-a`vis filling pressure.

and becomes as low as 0.1 at 1.75 mbar and then increases again to 0.35 at 2.0 mbar. For neon filling, the highest ratio of 0.5 is observed at 0.5 mbar, and it decreases to 0.25 at 1.5 mbar. Another small peak is recorded at a filling at 2.5 mbar. Figure 20 presents the pinhole camera images of the focus region, recorded by masking the pinhole apertures with Al (12 ␮m), Ni (5 ␮m), and Mg (25 ␮m) filters, respectively. The images with the Ni filters reveal that emission in the 4–8 keV windows is much higher with hydrogen filling and occurs from the anode tip. No emission in this window is observed from the pinch filament. From the analysis of pin-diode signals, the emission in 4␲ -geometry in the 4–8 keV window was estimated. The highest emission was found to be 0.7 J with neon filling and 3.2 J with hydrogen filling. The results are presented in Fig. 21. Therefore, the images recorded by the pinhole camera and the pin-diode measurements support each other. The Al filter offers a 1.2–1.5 keV window and the Mg filter provides a 1.2–1.3 keV window. The image with the Al filter is like a cone, with the base at the anode tip. It is speculated that the aluminum is sputtered and swept to the focus region, which gives rise to emission, mainly the Al-K␣ line. A well-defined pinch filament is formed with the neon filling, but no such filament is observed with the hydrogen gas. It reveals that the X-ray emission with hydrogen filling occurs from the anode tip as a result of interaction with the energetic electrons in the plasma current sheath. With the neon

Fig. 20. X-ray pinhole images of focus region masked with Al (12 ␮m), Ni (5 ␮m), and Mg (25 ␮m) filters for (a) hydrogen at 1.0 mbar, (b) neon at 3.0 mbar.

Fig. 21. The X-ray emission in the 4–8 keV window vis-a`-vis filling pressure, with neon and hydrogen as the working gases.

156 filling, the sputtered aluminum atoms are responsible for emission in the broad region in front of the anode. The emission is predominantly the Al-K␣ line. Under the Mg filter, a bright and well-defined filament is formed in front of the anode tip. No aluminum line lies in the Mg window of 1.2–1.3 keV. As discussed in Section 4, the dominant emission arises due to recombination of hydrogenlike neon (Ne IX) ions. It is speculated that the same emission mechanism is responsible here.

8. DISCUSSION AND CONCLUSIONS This paper presents the results of a series of experiments undertaken to study the plasma focus as a bright and high-efficiency X-ray source. The X-ray emission was found to be strongly dependent on the shape of central electrode, which is the anode in these experiments. The X-rays in the 1.2–1.3 keV window are generated from the focus plasma pinch filament, and the highest efficiency is obtained with neon filling and a copper anode. The recombination of hydrogenlike neon ions (Ne IX) with a few eV to a few 10s of eV electrons is thought to be responsible for the emission. The experiment with hydrogen provides a significantly high X-ray yield. However, the emission does not occur from the pinch column, but from the anode tip by the impact of energetic electrons in the current sheath. The contribution of X-ray generation by the impact of a well-known energetic electron beam [23] is not significant. The X-ray generation efficiency of plasma focus is higher than that of typical X-ray tubes. However, the lesssignificant role of an energetic electron beam compared to energetic electrons in the current sheath reveals that one must operate the plasma focus at a higher voltage to obtain high-energy X-rays with enhanced efficiency. Many experiments [4,16] have been conducted with argon filling, with the expectation that X-ray emission would be appreciably high. However, the absolute X-ray yield and system efficiency for X-radiation generation were not determined. These experiments demonstrate that pure argon gas does not provide significant X-ray yield. However, it remains to be established to what extent the mixture of hydrogen with argon may be suitable as a high-efficiency soft X-ray source. In conclusion, plasma focus would be a potential candidate for a tailored high-efficiency X-ray source in the 1.2–1.3 keV energy window with neon and in the Cu-K␣ line with hydrogen. The first form of radiation may find application for lithography in electronic chip manufacture. Similarly, the Cu-K␣ line source is quite distinct from conventional X-ray tubes. With the use of CCD cameras, it may be helpful to record XRD in a very

Zakaullah et al. small timespan of a few 10s of nanoseconds. The work to this end is in progress.

ACKNOWLEDGMENTS This work was partially supported by Quaid-i-Azam University Research Grant, Pakistan Science Foundation Project No. C-QU/Phys (108), Pakistan Atomic Energy Commission Project for Plasma Physics, and the Abdus Salam International Center for Theoretical Physics, Trieste, Italy, Project AC-7 Islamabad.

REFERENCES 1. R. Lebert, A. Engel, K. Bergmann, O. Treichel, C. Gavrilescu, and W. Neff, Proc. 4th Int. Conf. on Dense Z-Pinches, edited by N. R. Pereira, J. Davis, and P. E. Pulsifer, AIP CP 409, p. 291 (1997). 2. J. M. Bayley, G. Decker, W. Kies, M. Malzig, F. Muller, P. Rowekamp, J. Westheide, and Y. V. Sidelnikov, J. Appl. Phys. 69, 613, (1991). 3. M. Liu, X. Feng, S. V. Springham, and S. Lee, IEEE Transactions on Plasma Science 26, 135 (1998). 4. F. N. Beg, I. Ross, J. F. Worley, A. E. Danger, and M. G. Haines, J. Appl. Phys. 88, 3225 (2000). 5. S. Lee, P. Lee, G. Zhang, X. Feng, V. A. Gribkov, M. Liu, A. Serban, and T. K. S. Wong, IEEE Transactions on Plasma Science 26, 1119 (1998). 6. P. Lee, X. Feng, G. X. Zhang, M. H. Liu, and S. Lee, Plasma Sources Sciences Sci. Technol. 6, 343 (1997). 7. N. V. Filippov, T. I. Filippova, I. V. Khutoretskaia, V. V. Mialton, and V. P. Vinogradov, Phys. Lett. A 211, 168 (1996). 8. R. Lebert, A. Engel, and W. J. Neff, J. Appl. Phys. 78, 6414 (1995). 9. J. L. Bourgade, C. Cavailler, de. J. Mascureau, and J. L. Miquel, Rev. Sci. Instrum. 57, 2165 (1986). 10. K. Hirano and H. Kitaoka, Japan. J. Physical Soc. 65, 139 (1996). 11. C. Mang Ng, S. P. Moo, and C. S. Wong, IEEE Trans. Plasma Sci. 26, 1146 (1998). 12. K. Bergmann, R. Lebert, and W. Neff, J. Phys. D: Appl. Phys. 30, 990 (1997). 13. Y. Kato, I. Ochiai, Y. Watanabe, and S. J. Murayama, Vac. Sci. Technol. B 6, 195 (1988). 14. M. Zakaullah, Ijaz. Akhtar, A. Waheed, Khalid. Alamgir, Anwer Zeb Shah, and G. Murtaza, Plasma Sources Sci. Technol. 7, 206 (1998). 15. M. Zakaullah, Ijaz Akhtar, G. Murtaza, and A. Waheed, Phys. of Plasmas 6, 3188 (1999). 16. M. Zakaullah, I. Ahmed, A. R. Omer, G. Murtaza, and M. M. Beg, Plasma Sources Sci. Technol. 5, 544 (1996). 17. D. J. Johnson, Rev. Sci. Instrum. 45, 191 (1974). 18. J. W. Robison, Hand Book of Spectroscopy 1 (Cleveland, Ohio, CRC Press Inc., 1974). 19. Douglas Vaughan, X-ray Data Booklet (California: Lawrence Berkeley Laboratory, Center for X-ray Optics, 1986). 20. L. Kelly Raymond, Atomic Emission Lines below 2000 Angstrom Hydrogen through Argon, NRL Report 6648 (Washington, DC, 1974). 21. M. Zakaullah, Khalid. Alamgir, G. Murtaza, and A. Waheed, Plasma Sources Sci. Technol. 9, 592 (2000). 22. M. Zakaullah, Imtiaz Ahmed, M. Shafique, G. Murtaza, M. Yasin, and M. M. Beg, Physica Scripta 57, 136 (1998). 23. J. R. Smith, C. M. Luo, M. J. Rhee, and R. F. Schneider, Phys. Fluids 28, 2305 (1985).

Low-Energy Plasma Focus 24. F. N. Beg, I. Ross, and A. E. Dangor, 1997 Proc. 4th Int. Conf., edited by N. R. Pereira, J. Davis, and P. E. Pulsifier, 339 (1997). 25. M. Zakaullah, K. Alamgir, M. Shafique, S. M. Hassan, M. Sharif, and A. Waheed, Appl. Phys. Lett. 78, 877 (2001).

157 26. V. Metchnik and S. G. Tomlin, Proc. Phys. Soc. 81, 956 (1963). 27. P. Tothil, Brit. J. Appl. Phys. 1, 1093 (1968). 28. M. Zakaullah, G. Murtaza, S. Qamar, Imtiaz Ahmad, and M. M. Beg, Physica Scripta. 53, 360 (1996).