REVIEW OF SCIENTIFIC INSTRUMENTS
VOLUME 74, NUMBER 10
OCTOBER 2003
Hybrid evaporation: Glow discharge source for plasma immersion ion implantation L. H. Li Department of Physics and Materials Science, City University of Hong Kong, Tat Chee Avenue, Hong Kong and Department of 702, School of Mechanical Engineering and Automation, Beijing University of Aeronautics and Astronautics, Beijing 100083, China
R. W. Y. Poon, S. C. H. Kwok, and P. K. Chua) Department of Physics and Materials Science, City University of Hong Kong, Tat Chee Avenue, Kowloon, Hong Kong
Y. Q. Wu and Y. H. Zhang Department of 702, School of Mechanical Engineering and Automation, Beijing University of Aeronautics and Astronautics, Beijing 100083, China
共Received 18 March 2003; accepted 7 July 2003兲 In order to achieve stable operation for elements with a low melting point or high vapor pressure, a quasiequilibrium evaporation–glow discharge evaporation source has been designed and investigated for plasma immersion ion implantation. The important relationship between the pressure in the evaporation chamber and the implantation chamber is studied for optimal performance. Our experimental results show that the hybrid evaporation–glow discharge source is an effective method to produce ions from materials with low melting point and high vapor pressure. © 2003 American Institute of Physics. 关DOI: 10.1063/1.1606535兴
I. INTRODUCTION
pressure in the vaporization source and that in the implantation chamber is studied and the principles and theories for the design of this type of ion source are investigated. To verify our concept, a sulfur source has been designed, and the results show that a hybrid evaporation–glow discharge ion source is indeed effective in producing ions that would have been difficult by other conventional means.
Plasma immersion ion implantation 共PIII兲 is a useful niche technology for the modification of surface properties of materials and industrial components that are large or have an irregular shape. The samples are immersed in overlying plasma from which ions are extracted and implanted into the samples. A plasma source is typically needed to supply the ions in PIII. Traditionally, ions are produced by thermionic discharge,1 pulsed high-voltage glow discharge, radio frequency 共rf兲 plasma sources,2,3 microwave plasma sources, cathodic arc metal plasma source,4,5 sputtering targets,6,7 and so on. In spite of extensive research, there is still no single source that can provide ions of all elements. For instance, it is still inconvenient or difficult to ionize solid materials with poor electrical conductivity, such as sulfur, phosphorus, and boron, poorly or semiconducting materials such as silicon, germanium, and elements with low melt point and/or high chemical activity like the group IA and IIA elements. In this article, we report on a method to create ions from solid materials possessing low melting point and high vapor pressure using a hybrid evaporation–glow discharge technique. The material is initially vaporized in a source container and then introduced via a 6 mm diam gas inlet tube into a small implantation chamber with an internal glass shield to reduce contamination. The source container has a large surface area relative to that of the orifice 共inner diameter of the gas inlet tube兲 through which the evaporated species escape. This provides quasiequilibrium and yields a steady and easily controllable implantation process. The relationship between the
II. PRINCIPLES AND THEORIES
There are two basic types of evaporation sources: quasiequilibrium and nonequilibrium source.8 We adopt the quasiequilibrium approach in this work. As shown in Fig. 1, the source consists of a glass evaporation container with a large inner surface area relative to that of the orifice through which the evaporated species escapes. In this configuration, the evaporation process approaches thermal equilibrium and the vapor pressure merely depends on the evaporation temperature and inherent physical properties of the species. The vapor is fed into the implantation chamber via a long tube and subsequently ionized by the high voltage pulses applied to the sample. The negative high voltage pulses thus serve both purposes of generating the plasma and accelerating the ions in the sheath to the substrate. In this way, inadvertent plasma reactions with the thermionic cathode or rf antenna can be mitigated. In addition, the system cost and complexity is reduced.9 Because the glow discharge process is responsible for the generation of ions in this method, the ambient pressure that determines the properties of the glow discharge is of particular importance. It is thus crucial to understand the factors affecting the pressure in the implantation chamber that is mainly determined by the source container pressure
a兲
Author to whom correspondence should be addressed; electronic mail:
[email protected]
0034-6748/2003/74(10)/4301/4/$20.00
4301
© 2003 American Institute of Physics
4302
Li et al.
Rev. Sci. Instrum., Vol. 74, No. 10, October 2003
The variation of the gas particle flow per second in the implantation chamber is determined by the following three factors: 共1兲 Q inI : particle influx rate, including argon, and sulfur; 共2兲 Q outI : particles pump-out rate; and 共3兲 Q imp : particle consumption rate during implantation. The pressure in the implantation chamber can be described as follows: d 共 P impV imp兲 ⫽Q inI⫺Q outI⫺Q imp , dt
共3兲
where V imp is the volume of the implantation chamber and t is the time. By considering the conductance of the tube, we have FIG. 1. Schematic of the hybrid evaporation–glow discharge plasma ion source ( 150 mm⫻240 mm) in the plasma immersion ion implanter.
and the size of the inlet tube. In this study, in order to investigate the relationship between the gas pressure between the source container and the implantation chamber, the following assumptions are made: 共1兲 P impⰇ P background where P imp is the pressure in the implantation chamber and P background is the background pressure; 共2兲 the pressure in the evaporation container P source is determined only by the source temperature and by the flow rate of the carrying gas; 共3兲 T tube⫽T vapor , where T tube is the temperature of the inlet tube and T vapor is that of the vapor; 共4兲 the carrying gas and vapor pressure in the source are in steady state; 共5兲 the carrying gas has the same temperature as the sulfur vapor; 共6兲 gas leaks in the evaporation and implantation chambers are negligible; and 共7兲 implantation takes place when the pressure in the implantation chamber exceeds the breakdown pressure. To calculate the pressure, we first derive a gas influx model. The model is based on the Knudsen number that is defined as the ratio of the mean free path 共兲 to the typical dimension of the flow structure 共frequently the inside diameter兲.8 The inner diameter of the inlet tube, d tube , which is 6 mm, is much less than the particle mean free path, and so K n⫽
d tube
Ⰷ1.
共1兲
共4兲
where U tube is the conductance of the tube and can be described by the well-known Poisseuille formula8 and so U tube⫽
3 2 d tube av
3L tube
共5兲
,
where L tube is the length of the inlet tube, and av is the average speed of the particles, which can be given by
av⫽
冑
8kT , m
共6兲
where k is Boltzmann’s constant and m is the particle mass. Assuming that the pumping speed is S pump , we have d 共 P impV imp兲 ⫽U tube共 P source⫺ P imp兲 dt ⫺ 共 P imp•S pump⫹Q imp兲
共7兲
and V imp
d 共 P imp兲 ⫹ 共 U tube⫹S pump兲 P imp⫽U tubeP source⫺Q imp . dt 共8兲
We assume that at a given high voltage, the breakdown pressure is P breakdown , and when P imp⬍ P breakdown , Q imp⬇0. When P imp⭓ P breakdown , because the effects of the pressure on the Q imp are neglected, P imp⫽Ce ⫺ U tube⫹S pump /V imp t ⫹
U tubeP source⫺Q imp , U tube⫹S pump
共9兲
where C is a constant. At a long enough time,
Hence, the transport of the sulfur vapor via the tube is by molecular flow. Because chamber leaks are ignored, the gas pressure in the source container is the sum of the gas pressure and sulfur stream pressure P source⫽ P ArS⫹ P SulfurS ,
d 共 P impV imp兲 ⫽U tube共 P source⫺ P imp兲 ⫺ 共 Q pump⫹Q imp兲 , dt
共2兲
where the P ArS and P SulfurS are the argon partial pressure and sulfur partial pressure in the evaporation source, respectively.
C exp⫺
U tube⫹S pump t⇒0. V imp
Hence, when P imp⬎ P breakdown , P imp⫽
U tubeP source Q imp ⫺ . U tube⫹S pump U tube⫹S pump
共10兲
Before implantation or when the high glow discharge disappears, Q imp⇒0 and
Rev. Sci. Instrum., Vol. 74, No. 10, October 2003
P imp⫽
U tubeP source ⫽ U tube⫹S pump
P source . S pump 1⫹ U tube
Hybrid evaporation–glow ion source TABLE I. Implantation parameter.
共11兲
We can thus conclude that: 共1兲 Because implantation is conducted in a pulsed manner, the number of particles as well as the pressure in the implantation chamber fluctuate during the implantation process. 共2兲 The implantation chamber can be at a high vacuum if the cathodic sheath expands to a large enough thickness such that nearly all the particles are consumed by one or several cycles of implantation. That is to say, the discharge can be prolonged provided that it has begun, even if P imp⬍ P breakdown . 共3兲 Once the particles in the chamber are used up, the glow discharge will die out. The high voltage glow discharge will not be reignited until the pressure the implantation chamber exceeds P breakdown . This phenomenon has been observed in our experiments. In order to operate the hybrid ion source in a stable manner, it is necessary that P imp⬎ P breakdown , that is, P imp⫽
P source ⬎ P breakdown , S pump 1⫹ U tube
and so U tube⬎
S pump P source P breakdown
.
共12a兲
⫺1
Hence, the conductance of the tube is a key factor when designing this type of ion source. It should be large enough so that the eventual pressure in the implantation chamber is larger than P breakdown . At the same time, from Eq. 共5兲, U tube is determined by the tube length and diameter. Equation 共12兲 shows that P source , P breakdown , and S pump must be chosen properly, and it should be noted that P source and P breakdown are determined by the physical properties of the materials, whereas S pump is determined by the pumps. To achieve stable operation, there should be enough particles in the implantation chamber after the implantation process has begun so that P imp⭓ P breakdown . That is, P imp⫽
U tubeP source Q imp ⫺ ⭓ P breakdown . U tube⫹S pump U tube⫹S pump
共12b兲
We thus have U tube⭓
S pumpP breakdown⫹Q imp . P source⫺ P breakdown
4303
共13兲
Equation 共12兲 illustrates the necessary conditions in which implantation can take place, and Eq. 共13兲 shows the critical 共threshold兲 conductance value for steady state implantation. In order to demonstrate the hybrid ion source and verify our model, we conducted experiments using sulfur as the
Samples
Implantation voltage 共kV兲
Vapor composition
Implantation pressure 共Pa兲
Time 共min兲
A
10
sulfur
0.032–0.01
60
B
15 25
sulfur⫹argon sulfur⫹argon
0.014 –0.01 0.014 –0.01
60 120
materials under different conditions, for instance, varying the tube conductance and sulfur vaporizing temperature, and also with or without carrying gases. III. EXPERIMENTAL VERIFICATION
The plasma immersion ion implanter schematically illustrated in Fig. 1 has been described in detail elsewhere.10 The substrate holder was placed in the implantation chamber with a diameter of 150 mm and height of 240 mm. The distance between the vapor outlet and substrate was 220 mm. A 100 mm silicon wafer was used as the substrate and placed on the substrate holder. Sulfur was vaporized in a quartz evaporation container and fed into the implantation chamber through a gas inlet tube with a inner diameter of 6 mm. The sulfur vapor was introduced by two ways. 共1兲 Sulfur vapor was used as the source and fed into the implantation chamber directly. 共2兲 The sulfur vapor was mixed with argon, the carrying gas, and then fed into the implantation chamber. The implantation parameters are listed in Table I. The high voltage was operated at a frequency of 100 Hz with a pulse duration of 250 s, corresponding to a duty factor of 2.5%. Figures 2共a兲 and 2共b兲 depict the 15 and 25 kV implantation current and voltage wave forms. As usual, the current wave forms exhibit two spikes at the beginning and at the end. The initial spike lasts for about 8 s with a maximum value of about ⫺2 A and the other spike lasts for about 15 s with a maximum value of about 0.5 A. These spikes are the sum of the true implanted ion current, secondary electron current, as well as system capacitance. To subtract the system capacitance, Fig. 3 shows the spike current only due to the system capacitance without plasma production. The implantation current does not show a continuously decreasing trend that is typical of the expansion of the plasma sheath at low pressure.10,11 It is also different from highpressure, high-voltage implantation in which the current wave form usually exhibits an obvious breakdown current increase.12 The sulfur depth profiles acquired by sputtering x-ray photoelectron spectroscopy are shown in Fig. 4. The sulfur distribution in sample A shows a roughly Gaussian distribution, whereas it is broader in sample B because both 15 and 25 kV were used for sample B. At first glance, it seems abnormal that the sulfur distribution in sample A 共10 kV兲 is deeper than that of sample B 共15 and 25 kV兲. However, this can be explained by factoring sputtering loss into the process. During implantation of sample B, Ar was used as the carrying gas. Because Ar has a higher sputtering yield, surface etching of sample B was more severe, thereby resulting in an apparently shallower distribution.
4304
Li et al.
Rev. Sci. Instrum., Vol. 74, No. 10, October 2003
FIG. 4. Experimental sulfur depth profiles acquired using sputtering x-ray photoelectron spectroscopy: 共A兲 10 kV without feeding gas and 共B兲 15 and 25 kV with argon feeding gas.
FIG. 2. Implantation current and voltage wave forms: 共a兲 15 kV and 共b兲 25 kV.
Our experimental results show that the hybrid evaporation–glow discharge source is effective in producing ions that would be difficult using a conventional cathodic arc plasma source, such as elements with low melting point and high gas pressure. We have formulated a theoretical model that explains our sulfur implantation experiments. It should
be noted that our model is established based on the perfect conditions and is only qualitative. Hence, even though the exact relationship between the pressure and temperature is difficult to obtain experimentally in the implantation chamber, we observe continuous flashing, that is, the inability to achieve a stable glow discharge, when the pressure conditions are not proper 共as predicted by our model兲 and also that at very low pressure, the glow discharge cannot be sustained at all. Our experiments also reveal that the temperature of the inlet tube should be kept to be equal to the vapor temperature, otherwise condensation of the evaporated species on the inner wall will block the inlet tube causing failure. ACKNOWLEDGMENTS
The work was jointly supported by Germany/Hong Kong RGC Joint Research Scheme No. G – HK001/02 共CityU designation No. 9050165兲 and City University of Hong Kong SRG No. 7001389. 1
X. B. Tian, T. Zhang, R. K. Y. Fu, and P. K. Chu, Surf. Coat. Technol. 161, 232 共2002兲. H. Bender, J. Halder, F. Hilschert, H. Klein, J. Scha¨fer, B. Seiler, and R. Thomae, Surf. Coat. Technol. 93, 213 共1997兲. 3 B. Seiler et al., Nucl. Instrum. Methods Phys. Res. B 113, 266 共1996兲. 4 I. G. Brown, Surf. Coat. Technol. 136, 16 共2001兲. 5 L. H. Li, T. Zhang, P. K. Chu, and I. G. Brown, Rev. Sci. Instrum. 73, 2971 共2002兲. 6 S. Schoser, J. Forget, and K. Kohlhof, Surf. Coat. Technol. 93, 339 共1997兲. 7 C. Ruset and E. Grigore, Surf. Coat. Technol. 156, 159 共2002兲. 8 J. E. Mahan, Physical Vapor Deposition of Thin Films 共Wiley, New York, 2000兲, p. 115. 9 A. Anders, Handbook of Plasma Immersion Ion Implantation and Deposition 共Wiley, New York, 2000兲, pp. 393–394. 10 P. K. Chu, S. Qin, C. Chan, N. W. Cheung, and P. K. Ko, IEEE Trans. Plasma Sci. 26, 79 共1998兲. 11 J. Pelletier, F. Le Coeur, Y. Arnal, A. Lacoste, and A. Straboni, Surf. Coat. Technol. 136, 7 共2001兲. 12 J. N. Matossian and R. H. Wei, Surf. Coat. Technol. 85, 92 共1996兲. 2
FIG. 3. Spike current due to the system capacitance without plasma production.
