Direct coupling of pulsed radio frequency and pulsed high power in ...
REVIEW OF SCIENTIFIC INSTRUMENTS 79, 043501 共2008兲
Direct coupling of pulsed radio frequency and pulsed high power in novel pulsed power system for plasma immersion ion implantation Chunzhi Gong,1 Xiubo Tian,1,a兲 Shiqin Yang,1 Ricky K. Y. Fu,2 and Paul K. Chu2 1
State Key Laboratory of Advanced Welding Production and Technology, School of Materials Science and Engineering, Harbin Institute of Technology, 150001 Harbin, People’s Republic of China 2 Department of Physics and Materials Science, City University of Hong Kong, Tat Chee Avenue, Kowloon, Hong Kong
共Received 18 January 2008; accepted 2 March 2008; published online 11 April 2008兲 A novel power supply system that directly couples pulsed high voltage 共HV兲 pulses and pulsed 13.56 MHz radio frequency 共rf兲 has been developed for plasma processes. In this system, the sample holder is connected to both the rf generator and HV modulator. The coupling circuit in the hybrid system is composed of individual matching units, low pass filters, and voltage clamping units. This ensures the safe operation of the rf system even when the HV is on. The PSPICE software is utilized to optimize the design of circuits. The system can be operated in two modes. The pulsed rf discharge may serve as either the seed plasma source for glow discharge or high-density plasma source for plasma immersion ion implantation 共PIII兲. The pulsed high-voltage glow discharge is induced when a rf pulse with a short duration or a larger time interval between the rf and HV pulses is used. Conventional PIII can also be achieved. Experiments conducted on the new system confirm steady and safe operation. © 2008 American Institute of Physics. 关DOI: 10.1063/1.2906220兴 I. INTRODUCTION
Plasma immersion ion implantation 共PIII兲 is an advanced surface modification method that circumvents the line-ofsight restriction that plagues conventional beam-line ion implantation. In conventional PIII, the sample is enshrouded by gaseous plasma generated by external plasma sources. When a pulsed negative voltage is applied to the sample, positive ions in the plasma are attracted and implanted into the sample surface. Therefore PIII is often utilized to treat complex-shaped objects.1–3 Implantation uniformity is a critical factor determining not only the efficacy of the treatment process but also the acceptance of PIII in industrial applications. Hence, theoretical and experimental studies have been performed to improve the lateral ion flux uniformity on samples with an irregular shape.4–6 In most PIII systems, the plasma sources are located away from the sample holder, typically along the vacuum chamber wall, and so it may be difficult to obtain a high density and uniform plasma in the vicinity of the sample holder because of plasma diffusion from the source to the sample surface.7,8 This may lead to the nonuniform implantation for samples with irregular shape due to the effects of the large plasma sheath induced by the low plasma density near the sample holder. This nonuniformity is more severe for cavity structures such as trenches and inner bores due to the locally lower plasma density.9,10 The lower the plasma density, the more is the ion flux nonuniformity. Nishimura et al.11,12 have developed a power supply system that incorporates both radio frequency 共rf兲 and highvoltage input via a single cable and feedthrough. The highvoltage modulator is based on a hard-tube switch, and the a兲
Author to whom correspondence should be addressed. Tel.: 045186418784. Electronic mail: [email protected].
0034-6748/2008/79共4兲/043501/5/$23.00
sample also serves as the antenna to excite the plasma. When the rf pulses are turned off, negative high-voltage pulses are applied via the same feedthrough to the sample to conduct PIII. In this way, higher plasma density and uniformity may be achieved due to spontaneous plasma ignition. This system is thus particularly suitable for samples with complex threedimensional geometries as well as electrically insulating materials such as polymers and ceramics due to the discharge possibility during rf excitation. In this paper, we report the design and operational behavior of this novel hybrid system that couples pulsed rf and high voltage. Different from the hardware used by Nishimura et al., the core coupling circuit of our system is composed of three parts: the voltage clamping unit, rf matching unit, and low pass filtering unit. This system offers safer and more flexible control of the rf due to the lower potentials. In this system, the pulsed rf discharge acts not only as the plasma source but also the seed plasma for high-voltage glow discharge. II. HARDWARE A. Main electrical circuit
Figure 1 depicts the circuit schematic in the system that directly couples high-voltage 共HV兲 pulses and pulsed rf via a single feedthrough. The hybrid system consists of a HV modulator, rf power supply, timing unit, and coupling circuit. The driving and timing control unit is connected to the HV pusler and rf power supply to modulate the duration, frequency, and relative phase of the rf and HV pulses. The HV pulse can be applied to the sample after the rf pulse. For safety purposes, the coupling circuit is composed of three parts, as shown in Fig. 1. Part A is the voltage divider to ensure lower-voltage operation of the rf system independent of the HV pulse. Part B is the rf matching circuit to ensure
79, 043501-1
© 2008 American Institute of Physics
043501-2
Gong et al.
Rev. Sci. Instrum. 79, 043501 共2008兲
FIG. 1. Circuit schematic of the hybrid system that directly couples HV pulses and pulsed rf.
optimal rf output. Part C is the low pass filter that eliminates the effects of the rf on the HV modulator. The rf matching circuit is a -type circuit in which VC1 and VC2 are the tuning capacitors and L3 is a fixed inductance. Capacitor C1, capacitor C2, and inductor L1 form the voltage divider network to reduce HV risks in the rf power supply. The HV pulse is applied to the sample via a low pass filter 共L2兲 that prevents the rf from interfering with the HV pulse generator. The novel design ensures the low-potential regulation of the rf system even in the presence of the HV pulse, thereby offering safer and more flexible control. B. Circuit simulation
In order to achieve better operating characteristics, the hybrid system should provide a high substrate current from the rf pulse and high substrate voltage supplied by the HV pulse generator. The PSPICE software is utilized to investigate the effects and response characteristics of individual components in order to optimize the circuit. An equivalent plasma load circuit is composed of resistors and capacitors, as shown in section D in Fig. 1. A pulsed rf with inherent frequency of
FIG. 2. rf current as a function of C1.
13.56 MHz and a HV pulse with a ceiling voltage of 35 kV, pulse width of 20 s are employed in the simulation. Figure 2 shows the dependence of the rf current via capacitor C1. When the capacitance is small, the transmission capability of the rf current is very sensitive to the variations in the capacitance C1. The larger the capacitance, the higher is the transmission capability and the higher the efficiency of the rf power system. However, the capacitance C1 should not be larger due to the effects of the voltage divider. In this case, it has to be more than 100 pF. Figure 3 shows the clamping voltage across capacitor C2 and inductor L1. A higher capacitance C2 may help to decrease the clamping voltage due to the small ratio of C1 to C2. This bodes well for the safe operation of the rf system. However, the rf current through capacitor C2 may substantially increase, leading to possible damage of the rf device if C2 is too large. Here, the proper value of capacitance C2 is larger than 100 pF. In order to prevent overcurrent in the rf system, the capacitor C2 is connected with inductor L1 in series. The effects of inductor L1 on the clamping voltage are shown in Fig. 4. The clamping voltage increases with larger L1. The value of L1 is
FIG. 3. Clamping voltage of C2 and L1 as a function of C2.
043501-3
Novel pulsed power system
FIG. 4. Clamping voltage of C2 and L1 as a function of L1.
chosen in the range of 100– 500 H. The HV pulse is applied to the sample through a low pass filter 共L2兲 which prevents the rf from affecting the HV pulse generator. Figure 5 shows the current from the rf to HV pulse generator as a function of L2. A larger L2 reduces this influence, but it also increases the rise time of the HV pulse and subsequently lower implantation energy. Therefore, the optimal value of L2 is more than 100 H. III. OPERATIONAL CHARACTERISTICS
The present system allows more flexible process control via changeable instrumental parameters such as the pulsing duration, frequency, and relative phases. Depending on the instrumental conditions, the hybrid system can be operated in two modes. A. HV glow discharge
As aforementioned, pulsed HV glow discharge can be accomplished under proper conditions. When the rf pulse width is small or the interval between the rf pulse and HV
FIG. 5. Current from rf to HV pulse generator as a function of L2.
Rev. Sci. Instrum. 79, 043501 共2008兲
FIG. 6. 共Color online兲 Oscilloscope traces of the current pulses 共below兲 and voltage pulse of 12 kV 共above兲 with the rf pulse width of 1 ms and without rf.
pulse is large, HV glow discharge can be easily induced. The remnant plasma acts as the seed plasma for easy igniting. Figure 6 displays the typical discharge current behavior when a pulsed 12 kV potential is applied together with a rf pulse width of 1 ms at argon gas pressure of 0.7 Pa. There is only displacement current without rf. In contrast, introduction of the rf evidently promotes the electrical current in the glow discharge. When the pulse is turned on, there is a time delay in which the electrons of the remaining plasma are repelled and ions are accelerated to the sample leading to secondary electron emission. These secondary electrons ionize the working gas. When there is sufficient ionization, a self-sustaining discharge occurs. As illustrated in Fig. 6 a shorter fall time is observed for the HV due to a large density of the remnant plasma after the voltage pulse has been turned off. Figure 7 exhibits the effects of the rf pulse width on the discharge current. The first peak in the time-dependent currents hardly changes, suggesting that the influence of the initial plasma is small. However, a different density of the remnant plasma may induce a different discharge behavior.
FIG. 7. 共Color online兲 Discharge current for different rf pulse widths.
043501-4
Gong et al.
FIG. 8. 共Color online兲 Discharge current for different time intervals between pulsed rf and pulsed HV.
For a smaller rf pulse duration, the delay time in the HV glow discharge is larger, and it takes a longer time to reach the steady discharge current. In contrast, the discharge current rapidly increases when the rf pulse duration is long 共e.g., 2.5 ms兲. This is because a different pulse duration may lead to a different density of the remnant plasma. Figure 8 displays the influence of the interval between rf and HV pulses on the discharge behavior with a pulse voltage of 12 kV and pulsing frequency of 88 Hz. A shorter time interval gives rise to a smaller time delay for a steady discharge current. It can be easily understood that the combination of ions and electrons is small for a small interval. Therefore, the remnant plasma density is still high when the HV pulse is applied. With decreasing interval 共e.g., 4 ms兲, there is an interesting phenomenon in which the current abruptly changes. This is similar to that in conventional PIII without HV glow discharge. The results imply that the density of the remnant plasma is large enough to sustain PIII. rf pulse discharge is an effective means to create the seed plasma in this hybrid system. The plasma produced by the rf pulse does not completely decay after the rf pulse is turned off and, hence, there remains remnant plasma in the chamber.13,14 For a larger rf pulse width or shorter time interval between the pulsed rf and pulsed HV, the remnant plasma 共seed plasma兲 increases. When the HV pulse is applied, ions are accelerated and secondary electrons are copiously created. This leads to the onset of HV glow discharge. The discharge current may reach saturation depending on the applied voltage as well as the pressure of the working gas being ionized by the secondary electrons.15–17
Rev. Sci. Instrum. 79, 043501 共2008兲
FIG. 9. 共Color online兲 Oscilloscope traces of the current pulses 共below兲 and voltage pulse of 10 kV 共above兲 with the rf pulse width of 6 ms and without rf.
single pulse, in contrast with the gradual increase to saturation in HV glow discharge. This is a typical current configuration in conventional PIII featured by a higher plasma density near the sample. Regular PIII requires a higher plasma density achieved by a longer rf pulse width and smaller time interval between the rf and HV pulses. Figure 10 illustrates the effects of the rf pulse width on the implantation current at a pulse voltage of 10 kV and gas pressure of 0.6 Pa. A larger pulse width may lead to a higher total implantation current as indicated by the initial peaks. However, the effect is slight suggesting that there is a minimum rf pulse width. A longer pulse width does not give rise to an appreciable contribution. Figure 11 describes the response of the implantation current to the applied voltage with the rf pulse width of 6 ms, interval time of 2.5 ms, pressure of 0.6 Pa, and frequency of 66 Hz. A HV leads to a higher implantation current. These observations are common for regular PIII. In contrast to HV glow discharge, rf pulse discharge is an effective means to perform
B. PIII
HV glow discharge occurs with a lower density of the remnant plasma. When the average power of the rf pulse is high, that is, when the plasma density is sufficiently high, regular PIII can be conducted. Figure 9 shows the typical implantation current for an applied HV pulse of 10 kV and a rf pulse width of 6 ms. The initial peak current is much larger than the displacement current without the rf pulse. More importantly, the current gradually diminishes in a
FIG. 10. 共Color online兲 Dependence of implantation current on rf pulse width.
043501-5
Rev. Sci. Instrum. 79, 043501 共2008兲
Novel pulsed power system
Hong Kong Research Grants Council 共RGC兲 Central Allocation Votes 共CAV兲 No. CityU 1/06C. 1
FIG. 11. 共Color online兲 Implantation current for different pulse voltages.
PIII. However, a minimum plasma density must be available before the negative potential is applied to the sample. This requires a larger rf pulse duration and smaller time interval between the rf pulse and HV pulse or higher rf power 共not investigated here兲. ACKNOWLEDGMENTS
The work was jointly supported by Natural Science Foundation of China 共No. 10575025 and No. 10775036兲 and
V. D. D. Jabon, J. D. Moreno, and P. A. Tsygankov, Rev. Sci. Instrum. 73, 828 共2002兲. 2 X. D. Chen, C. C. Ling, S. Fung, C. D. Beling, Y. F. Mei, R. K. Y. Fu, G. G. Siu, and P. K. Chu, Appl. Phys. Lett. 88, 132104 共2006兲. 3 R. K. Y. Fu, X. B. Tian, and P. K. Chu, Rev. Sci. Instrum. 74, 3697 共2003兲. 4 E. Stamate, N. Holtzer, and H. Sugai, Appl. Phys. Lett. 86, 261501 共2005兲. 5 H. Ghomi, M. Sharifian, A. R. Niknam, and B. Shokri, J. Appl. Phys. 100, 113301 共2006兲. 6 X. B. Tian and P. K. Chu, Appl. Phys. Lett. 81, 3744 共2002兲. 7 X. B. Tian, R. K. Y. Fu, P. K. Chu, A. Anders, C. Z. Gong, and S. Q. Yang, Rev. Sci. Instrum. 74, 5137 共2003兲. 8 L. H. Li, R. W. Y. Poon, S. C. H. Kwok, P. K. Chu, Y. Q. Wu, and Y. H. Zhang, Rev. Sci. Instrum. 74, 4301 共2003兲. 9 H. H. Tong, R. K. Y. Fu, D. L. Tang, X. C. Zeng, and P. K. Chu, J. Appl. Phys. 92, 2284 共2002兲. 10 T. K. K. Dixon, S. H. Pu, R. K. Y. Fu, F. Y. Jin, and P. K. Chu, Appl. Phys. Lett. 90, 131503 共2007兲. 11 Y. Nishimura, A. Chayahar, Y. Horino, and M. Yatsuzuka, Surf. Coat. Technol. 156, 50 共2002兲. 12 Y. Nishimura, R. Ohkawa, H. Oka, H. Akamatsu, K. Azuma, and M. Yatsuzuka, Nucl. Instrum. Methods Phys. Res. B 206, 696 共2003兲. 13 J. Reijonen, K. N. Leung, and G. Jones, Rev. Sci. Instrum. 73, 934 共2002兲. 14 O. Demokan and S. Akman, J. Appl. Phys. 92, 4245 共2002兲. 15 P. L. Kellerman, S. Qin, M. P. Bradley, and K. Saadatmand, Rev. Sci. Instrum. 73, 837 共2002兲. 16 S. Radovanov, L. Godet, R. Dorai, Z. Fang, B. W. Koo, C. Cardinaud, G. Cartry, D. Lenoble, and A. Grouillet, J. Appl. Phys. 98, 113307 共2005兲. 17 S. Mukherjee, M. Ranjan, R. Rane, N. Vaghela, A. Phukan, and K. S. Suraj, Surf. Coat. Technol. 201, 6502 共2007兲.
Direct coupling of pulsed radio frequency and pulsed high power in novel pulsed power system for plasma immersion ion implantation Chunzhi Gong,1 Xiubo Tian,1,a兲 Shiqin Yang,1 Ricky K. Y. Fu,2 and Paul K. Chu2 1
State Key Laboratory of Advanced Welding Production and Technology, School of Materials Science and Engineering, Harbin Institute of Technology, 150001 Harbin, People’s Republic of China 2 Department of Physics and Materials Science, City University of Hong Kong, Tat Chee Avenue, Kowloon, Hong Kong
共Received 18 January 2008; accepted 2 March 2008; published online 11 April 2008兲 A novel power supply system that directly couples pulsed high voltage 共HV兲 pulses and pulsed 13.56 MHz radio frequency 共rf兲 has been developed for plasma processes. In this system, the sample holder is connected to both the rf generator and HV modulator. The coupling circuit in the hybrid system is composed of individual matching units, low pass filters, and voltage clamping units. This ensures the safe operation of the rf system even when the HV is on. The PSPICE software is utilized to optimize the design of circuits. The system can be operated in two modes. The pulsed rf discharge may serve as either the seed plasma source for glow discharge or high-density plasma source for plasma immersion ion implantation 共PIII兲. The pulsed high-voltage glow discharge is induced when a rf pulse with a short duration or a larger time interval between the rf and HV pulses is used. Conventional PIII can also be achieved. Experiments conducted on the new system confirm steady and safe operation. © 2008 American Institute of Physics. 关DOI: 10.1063/1.2906220兴 I. INTRODUCTION
Plasma immersion ion implantation 共PIII兲 is an advanced surface modification method that circumvents the line-ofsight restriction that plagues conventional beam-line ion implantation. In conventional PIII, the sample is enshrouded by gaseous plasma generated by external plasma sources. When a pulsed negative voltage is applied to the sample, positive ions in the plasma are attracted and implanted into the sample surface. Therefore PIII is often utilized to treat complex-shaped objects.1–3 Implantation uniformity is a critical factor determining not only the efficacy of the treatment process but also the acceptance of PIII in industrial applications. Hence, theoretical and experimental studies have been performed to improve the lateral ion flux uniformity on samples with an irregular shape.4–6 In most PIII systems, the plasma sources are located away from the sample holder, typically along the vacuum chamber wall, and so it may be difficult to obtain a high density and uniform plasma in the vicinity of the sample holder because of plasma diffusion from the source to the sample surface.7,8 This may lead to the nonuniform implantation for samples with irregular shape due to the effects of the large plasma sheath induced by the low plasma density near the sample holder. This nonuniformity is more severe for cavity structures such as trenches and inner bores due to the locally lower plasma density.9,10 The lower the plasma density, the more is the ion flux nonuniformity. Nishimura et al.11,12 have developed a power supply system that incorporates both radio frequency 共rf兲 and highvoltage input via a single cable and feedthrough. The highvoltage modulator is based on a hard-tube switch, and the a兲
Author to whom correspondence should be addressed. Tel.: 045186418784. Electronic mail: [email protected].
0034-6748/2008/79共4兲/043501/5/$23.00
sample also serves as the antenna to excite the plasma. When the rf pulses are turned off, negative high-voltage pulses are applied via the same feedthrough to the sample to conduct PIII. In this way, higher plasma density and uniformity may be achieved due to spontaneous plasma ignition. This system is thus particularly suitable for samples with complex threedimensional geometries as well as electrically insulating materials such as polymers and ceramics due to the discharge possibility during rf excitation. In this paper, we report the design and operational behavior of this novel hybrid system that couples pulsed rf and high voltage. Different from the hardware used by Nishimura et al., the core coupling circuit of our system is composed of three parts: the voltage clamping unit, rf matching unit, and low pass filtering unit. This system offers safer and more flexible control of the rf due to the lower potentials. In this system, the pulsed rf discharge acts not only as the plasma source but also the seed plasma for high-voltage glow discharge. II. HARDWARE A. Main electrical circuit
Figure 1 depicts the circuit schematic in the system that directly couples high-voltage 共HV兲 pulses and pulsed rf via a single feedthrough. The hybrid system consists of a HV modulator, rf power supply, timing unit, and coupling circuit. The driving and timing control unit is connected to the HV pusler and rf power supply to modulate the duration, frequency, and relative phase of the rf and HV pulses. The HV pulse can be applied to the sample after the rf pulse. For safety purposes, the coupling circuit is composed of three parts, as shown in Fig. 1. Part A is the voltage divider to ensure lower-voltage operation of the rf system independent of the HV pulse. Part B is the rf matching circuit to ensure
79, 043501-1
© 2008 American Institute of Physics
043501-2
Gong et al.
Rev. Sci. Instrum. 79, 043501 共2008兲
FIG. 1. Circuit schematic of the hybrid system that directly couples HV pulses and pulsed rf.
optimal rf output. Part C is the low pass filter that eliminates the effects of the rf on the HV modulator. The rf matching circuit is a -type circuit in which VC1 and VC2 are the tuning capacitors and L3 is a fixed inductance. Capacitor C1, capacitor C2, and inductor L1 form the voltage divider network to reduce HV risks in the rf power supply. The HV pulse is applied to the sample via a low pass filter 共L2兲 that prevents the rf from interfering with the HV pulse generator. The novel design ensures the low-potential regulation of the rf system even in the presence of the HV pulse, thereby offering safer and more flexible control. B. Circuit simulation
In order to achieve better operating characteristics, the hybrid system should provide a high substrate current from the rf pulse and high substrate voltage supplied by the HV pulse generator. The PSPICE software is utilized to investigate the effects and response characteristics of individual components in order to optimize the circuit. An equivalent plasma load circuit is composed of resistors and capacitors, as shown in section D in Fig. 1. A pulsed rf with inherent frequency of
FIG. 2. rf current as a function of C1.
13.56 MHz and a HV pulse with a ceiling voltage of 35 kV, pulse width of 20 s are employed in the simulation. Figure 2 shows the dependence of the rf current via capacitor C1. When the capacitance is small, the transmission capability of the rf current is very sensitive to the variations in the capacitance C1. The larger the capacitance, the higher is the transmission capability and the higher the efficiency of the rf power system. However, the capacitance C1 should not be larger due to the effects of the voltage divider. In this case, it has to be more than 100 pF. Figure 3 shows the clamping voltage across capacitor C2 and inductor L1. A higher capacitance C2 may help to decrease the clamping voltage due to the small ratio of C1 to C2. This bodes well for the safe operation of the rf system. However, the rf current through capacitor C2 may substantially increase, leading to possible damage of the rf device if C2 is too large. Here, the proper value of capacitance C2 is larger than 100 pF. In order to prevent overcurrent in the rf system, the capacitor C2 is connected with inductor L1 in series. The effects of inductor L1 on the clamping voltage are shown in Fig. 4. The clamping voltage increases with larger L1. The value of L1 is
FIG. 3. Clamping voltage of C2 and L1 as a function of C2.
043501-3
Novel pulsed power system
FIG. 4. Clamping voltage of C2 and L1 as a function of L1.
chosen in the range of 100– 500 H. The HV pulse is applied to the sample through a low pass filter 共L2兲 which prevents the rf from affecting the HV pulse generator. Figure 5 shows the current from the rf to HV pulse generator as a function of L2. A larger L2 reduces this influence, but it also increases the rise time of the HV pulse and subsequently lower implantation energy. Therefore, the optimal value of L2 is more than 100 H. III. OPERATIONAL CHARACTERISTICS
The present system allows more flexible process control via changeable instrumental parameters such as the pulsing duration, frequency, and relative phases. Depending on the instrumental conditions, the hybrid system can be operated in two modes. A. HV glow discharge
As aforementioned, pulsed HV glow discharge can be accomplished under proper conditions. When the rf pulse width is small or the interval between the rf pulse and HV
FIG. 5. Current from rf to HV pulse generator as a function of L2.
Rev. Sci. Instrum. 79, 043501 共2008兲
FIG. 6. 共Color online兲 Oscilloscope traces of the current pulses 共below兲 and voltage pulse of 12 kV 共above兲 with the rf pulse width of 1 ms and without rf.
pulse is large, HV glow discharge can be easily induced. The remnant plasma acts as the seed plasma for easy igniting. Figure 6 displays the typical discharge current behavior when a pulsed 12 kV potential is applied together with a rf pulse width of 1 ms at argon gas pressure of 0.7 Pa. There is only displacement current without rf. In contrast, introduction of the rf evidently promotes the electrical current in the glow discharge. When the pulse is turned on, there is a time delay in which the electrons of the remaining plasma are repelled and ions are accelerated to the sample leading to secondary electron emission. These secondary electrons ionize the working gas. When there is sufficient ionization, a self-sustaining discharge occurs. As illustrated in Fig. 6 a shorter fall time is observed for the HV due to a large density of the remnant plasma after the voltage pulse has been turned off. Figure 7 exhibits the effects of the rf pulse width on the discharge current. The first peak in the time-dependent currents hardly changes, suggesting that the influence of the initial plasma is small. However, a different density of the remnant plasma may induce a different discharge behavior.
FIG. 7. 共Color online兲 Discharge current for different rf pulse widths.
043501-4
Gong et al.
FIG. 8. 共Color online兲 Discharge current for different time intervals between pulsed rf and pulsed HV.
For a smaller rf pulse duration, the delay time in the HV glow discharge is larger, and it takes a longer time to reach the steady discharge current. In contrast, the discharge current rapidly increases when the rf pulse duration is long 共e.g., 2.5 ms兲. This is because a different pulse duration may lead to a different density of the remnant plasma. Figure 8 displays the influence of the interval between rf and HV pulses on the discharge behavior with a pulse voltage of 12 kV and pulsing frequency of 88 Hz. A shorter time interval gives rise to a smaller time delay for a steady discharge current. It can be easily understood that the combination of ions and electrons is small for a small interval. Therefore, the remnant plasma density is still high when the HV pulse is applied. With decreasing interval 共e.g., 4 ms兲, there is an interesting phenomenon in which the current abruptly changes. This is similar to that in conventional PIII without HV glow discharge. The results imply that the density of the remnant plasma is large enough to sustain PIII. rf pulse discharge is an effective means to create the seed plasma in this hybrid system. The plasma produced by the rf pulse does not completely decay after the rf pulse is turned off and, hence, there remains remnant plasma in the chamber.13,14 For a larger rf pulse width or shorter time interval between the pulsed rf and pulsed HV, the remnant plasma 共seed plasma兲 increases. When the HV pulse is applied, ions are accelerated and secondary electrons are copiously created. This leads to the onset of HV glow discharge. The discharge current may reach saturation depending on the applied voltage as well as the pressure of the working gas being ionized by the secondary electrons.15–17
Rev. Sci. Instrum. 79, 043501 共2008兲
FIG. 9. 共Color online兲 Oscilloscope traces of the current pulses 共below兲 and voltage pulse of 10 kV 共above兲 with the rf pulse width of 6 ms and without rf.
single pulse, in contrast with the gradual increase to saturation in HV glow discharge. This is a typical current configuration in conventional PIII featured by a higher plasma density near the sample. Regular PIII requires a higher plasma density achieved by a longer rf pulse width and smaller time interval between the rf and HV pulses. Figure 10 illustrates the effects of the rf pulse width on the implantation current at a pulse voltage of 10 kV and gas pressure of 0.6 Pa. A larger pulse width may lead to a higher total implantation current as indicated by the initial peaks. However, the effect is slight suggesting that there is a minimum rf pulse width. A longer pulse width does not give rise to an appreciable contribution. Figure 11 describes the response of the implantation current to the applied voltage with the rf pulse width of 6 ms, interval time of 2.5 ms, pressure of 0.6 Pa, and frequency of 66 Hz. A HV leads to a higher implantation current. These observations are common for regular PIII. In contrast to HV glow discharge, rf pulse discharge is an effective means to perform
B. PIII
HV glow discharge occurs with a lower density of the remnant plasma. When the average power of the rf pulse is high, that is, when the plasma density is sufficiently high, regular PIII can be conducted. Figure 9 shows the typical implantation current for an applied HV pulse of 10 kV and a rf pulse width of 6 ms. The initial peak current is much larger than the displacement current without the rf pulse. More importantly, the current gradually diminishes in a
FIG. 10. 共Color online兲 Dependence of implantation current on rf pulse width.
043501-5
Rev. Sci. Instrum. 79, 043501 共2008兲
Novel pulsed power system
Hong Kong Research Grants Council 共RGC兲 Central Allocation Votes 共CAV兲 No. CityU 1/06C. 1
FIG. 11. 共Color online兲 Implantation current for different pulse voltages.
PIII. However, a minimum plasma density must be available before the negative potential is applied to the sample. This requires a larger rf pulse duration and smaller time interval between the rf pulse and HV pulse or higher rf power 共not investigated here兲. ACKNOWLEDGMENTS
The work was jointly supported by Natural Science Foundation of China 共No. 10575025 and No. 10775036兲 and
V. D. D. Jabon, J. D. Moreno, and P. A. Tsygankov, Rev. Sci. Instrum. 73, 828 共2002兲. 2 X. D. Chen, C. C. Ling, S. Fung, C. D. Beling, Y. F. Mei, R. K. Y. Fu, G. G. Siu, and P. K. Chu, Appl. Phys. Lett. 88, 132104 共2006兲. 3 R. K. Y. Fu, X. B. Tian, and P. K. Chu, Rev. Sci. Instrum. 74, 3697 共2003兲. 4 E. Stamate, N. Holtzer, and H. Sugai, Appl. Phys. Lett. 86, 261501 共2005兲. 5 H. Ghomi, M. Sharifian, A. R. Niknam, and B. Shokri, J. Appl. Phys. 100, 113301 共2006兲. 6 X. B. Tian and P. K. Chu, Appl. Phys. Lett. 81, 3744 共2002兲. 7 X. B. Tian, R. K. Y. Fu, P. K. Chu, A. Anders, C. Z. Gong, and S. Q. Yang, Rev. Sci. Instrum. 74, 5137 共2003兲. 8 L. H. Li, R. W. Y. Poon, S. C. H. Kwok, P. K. Chu, Y. Q. Wu, and Y. H. Zhang, Rev. Sci. Instrum. 74, 4301 共2003兲. 9 H. H. Tong, R. K. Y. Fu, D. L. Tang, X. C. Zeng, and P. K. Chu, J. Appl. Phys. 92, 2284 共2002兲. 10 T. K. K. Dixon, S. H. Pu, R. K. Y. Fu, F. Y. Jin, and P. K. Chu, Appl. Phys. Lett. 90, 131503 共2007兲. 11 Y. Nishimura, A. Chayahar, Y. Horino, and M. Yatsuzuka, Surf. Coat. Technol. 156, 50 共2002兲. 12 Y. Nishimura, R. Ohkawa, H. Oka, H. Akamatsu, K. Azuma, and M. Yatsuzuka, Nucl. Instrum. Methods Phys. Res. B 206, 696 共2003兲. 13 J. Reijonen, K. N. Leung, and G. Jones, Rev. Sci. Instrum. 73, 934 共2002兲. 14 O. Demokan and S. Akman, J. Appl. Phys. 92, 4245 共2002兲. 15 P. L. Kellerman, S. Qin, M. P. Bradley, and K. Saadatmand, Rev. Sci. Instrum. 73, 837 共2002兲. 16 S. Radovanov, L. Godet, R. Dorai, Z. Fang, B. W. Koo, C. Cardinaud, G. Cartry, D. Lenoble, and A. Grouillet, J. Appl. Phys. 98, 113307 共2005兲. 17 S. Mukherjee, M. Ranjan, R. Rane, N. Vaghela, A. Phukan, and K. S. Suraj, Surf. Coat. Technol. 201, 6502 共2007兲.
