Investigation of low-pressure elevated-temperature plasma immersion ion implantation of AISI 304 stainless steel

Transcription

1 Investigation of low-pressure elevated-temperature plasma immersion ion implantation of AISI 304 stainless steel Xiubo Tian and Paul K. Chu a) Department of Physics and Materials Science, City University of Hong Kong, 83 Tat Chee Avenue, Kowloon, Hong Kong Received 23 October 2000; accepted 19 February 2001 Elevated-temperature plasma immersion ion implantation can be used to improve the surface properties of austenite stainless steels. Unlike previous investigations conducted using radio frequency plasma at a moderate gas pressure Pa, we recently conducted a series of experiments at lower pressure 0.06 Pa utilizing nitrogen plasma sustained by hot filament glow discharge. The implantation voltage was varied from 8 to 25 kv and the sample temperature was kept at 360 C by adjusting the implantation current density. To elucidate the mechanism and dynamics of the process, the treated samples were characterized by Auger electron spectroscopy and Rutherford backscattering spectrometry. Our experimental results show that surface oxidation is very severe at lower pressure due to higher oxygen partial pressure. It affects the nitrogen profile and diffusion. Since typical plasma immersion ion implanters are not designed for ultrahigh vacuum, the presence of the oxygen in the residual vacuum can give rise to unexpected results, particularly under the elevated-temperature conditions, and it must be controlled properly American Vacuum Society. DOI: / I. INTRODUCTION Plasma immersion ion implantation PIII is an effective surface modification method. 1 3 The technique circumvents the line of sight and retained dose limitations inherent to conventional beam-line ion implantation, and is particularly useful for large components possessing nonplanar and complex geometries. In order to increase the thickness of the modified surface layer, elevated-temperature PIII 4,5 has been proposed and shown to be more suitable for the strengthening of materials that are difficult to nitride such as aluminum alloys 6,7 and chromium based stainless steel. 8,9 This treatment is usually conducted in rf plasma at a moderate gas pressure between 0.1 and 1.0 Pa because experiments at lower pressure are more difficult due to hardware constraints. Under these conditions, the enhancement mechanism is believed to be a synergistic effect involving implantation, adsorption, and diffusion Unlike conventional plasma nitriding which is primarily a thermal and diffusion process, a higher implantation voltage here enables the incident ions to pass through the surface oxide layer while reducing sputtering loss to attain better treatment results. In PIII processes, the nitrogen plasma provides the medium for the enhancement of the surface properties of the materials. Different plasma generation methods require different electric and magnetic fields as well as working gas pressure. Thus, plasma properties such as the plasma potential, electron temperature, density, and ion distribution depend on the formation method. For instance, the nitrogen plasma density and density of excited neutrals vary with the working pressure in rf plasma. 12 Hence, it is conceivable that the modification results also depend on the plasma formation a Author to whom all correspondence should be addressed; electronic mail: paul.chu@cityu.edu.hk method. Plasma sustained by hot filament glow discharge has good potential in the metallurgical industry due to its simplicity. In addition, PIII experiments can be conducted at much lower pressure when hot filaments are employed. In this work, we focus on the elevated-temperature lowpressure PIII technique using hot filament glow discharge plasma to study its efficacy and the enhancement mechanism. II. EXPERIMENT The samples used in this study were AISI 304 stainless steel 20 mm in diameter and 3 mm thick. One side of each sample was polished and ultrasonically cleaned before loading into the plasma immersion ion implanter. 13 The samples were first sputter cleaned and heated using argon highfrequency low-voltage plasma ion bombardment. 14 The instrumental parameters were: pulsing frequency 10 khz, pulse duration 25 s, and bias voltage 2.5 kv. Afterwards, argon was replaced by nitrogen and the nitrogen plasma was sustained by hot filament glow discharge. Three different pulsed target voltages: 8, 16, and 25 kv, were used to conduct PIII at 360 C for 90 min. The experimental parameters are summarized in Table I. During implantation, the samples were heated spontaneously by energetic ion bombardment, and the temperature that was monitored continuously using an in situ thermocouple 15 was adjusted by iteratively varying the pulsing frequency and plasma density discharge current of the hot filament glow discharge. The typical temperature variation during the treatment process is depicted in Fig. 1. The implantation voltage and target current waveforms are displayed in Fig. 2. The rise or fall time depends on the inherent capacitance of the hardware and plasma sheath, and as a result, there is a current surge when the pulsing modulator is switched on and off. 16 After the 1008 J. Vac. Sci. Technol. A 19 3, MayÕJun Õ2001Õ19 3 Õ1008Õ5Õ$ American Vacuum Society 1008

2 1009 X. Tian and P. K. Chu: Low-pressure, elevated-temperature plasma 1009 TABLE I. Instrumental parameters. Parameters Process 1 Process 2 Process 3 Temperature C Voltage kv Pulse duration s Pulsing frequency Hz Discharge current A Gas pressure mtorr treatment, the samples were characterized by Auger electron spectroscopy AES using a PHI-610 scanning Auger microscope. The depth profiles were acquired using Ar ion bombardment at an estimated average sputtering rate of 20 nm/ min. III. RESULTS The Auger results reveal that besides implanted nitrogen, the top surface has a large amount of oxygen and carbon due to gases in the residual vacuum. The nitrogen depth profiles feature thermal diffusion and a long tail depending on the implantation voltage, as shown in Fig. 3. After treatment, nitrogen can be detected to a depth of up to 150 nm, and the nitrogen distribution varies with the implantation voltage as illustrated in Fig. 4. Unlike conventional elevated temperature PIII using rf plasma 4 11 in which a higher implantation voltage leads to a lower retained nitrogen dose and thinner modified later, the surface concentration and penetration depth increase with the implantation voltage in this mode. In order to corroborate the Auger results, the samples were also measured by Rutheford backscattering RBS. The RBS conditions are: 3.7 MeV He,10 C, and using nitrogen resonance to avoid interference from oxygen and enhance the nitrogen signal. The resonance process increases the scattering yield and subsequently the sensitivity. It is commonly used to enhance the sensitivity of light elements such as O FIG. 2. Implantation voltage and current wave forms: a Voltage and b current. and N. The regular backscattered peaks of O and N are close together in a silicon substrate, but in the resonance analysis, the energies are different, and so signal separation is much easier. The spectra are exhibited in Fig. 5 indicating that the Auger and RBS results are consistent. It further confirms that a higher implantation voltage indeed leads to higher nitrogen FIG. 1. Measured temperature evolution in 16 kv PIII: 1 Argon plasma bombardment heating by high-frequency low-voltage PIII: pulsing frequency 10 khz, bias voltage 2.5 kv, and pulse width 25 ms. 2 Argon gas replaced by nitrogen kv / 300 Hz / 30 ms PIII kv / 160 Hz / 30 ms PIII. 5 Cooling in vacuum chamber at a pressure of 0.05 mtorr. FIG. 3. Elemental depth profiles in 16 kv sample acquired by sputtering Auger analysis. JVST A - Vacuum, Surfaces, and Films

3 1010 X. Tian and P. K. Chu: Low-pressure, elevated-temperature plasma 1010 FIG. 4. Effects of implantation voltage on nitrogen depth profiles. FIG. 5. RBS spectra acquired from the treated samples. The 25 kv curve was artificially raised by 500 counts to make comparison easier. FIG. 6. Oxygen depth profiles acquired from the three samples. Also shown are the calculated nitrogen projected ranges at the respective voltages and the relative doses. incorporation into the samples, and the trend is in fact different from that observed in rf PIII processes at elevated temperature. Oxygen is very abundant in the subsurface region as demonstrated in Fig. 6. A higher implantation voltage gives rise to more oxygen in this layer. The oxygen has penetrated a depth of about nm. As shown in the inserted histogram in Fig. 6, the voltage affects the oxygen incorporation more than that of nitrogen. One possible source of the oxygen is the original oxide from which oxygen can be recoiled into the substrate as a result of energetic ion bombardment. However, since the samples underwent argon plasma sputtering and heating before the nitrogen plasma was ignited, the oxide film should have been sputtered away, and the contribution of oxygen from the original surface oxide should be small. The second and more reasonable explanation is oxygen in the residual vacuum of our instrument arising from out-gassing and minor leaks in the gas lines. The existence of residual oxygen is in fact quite common in plasma processing equipment that does not operate in ultrahigh vacuum UHV conditions. The residual oxygen gas is ionized in the plasma and coimplanted into the substrate together with nitrogen. Reactive oxygen species in the plasma will also oxidize the sample surface significantly even when the voltage pulse is off. Figure 3 also reveals that the chromium and nickel atoms retreat into the bulk of the sample and there is a pileup after the oxygen peak. On the contrary, in low-temperature PIII of AISI 304, 17,18 the chromium distribution shifts towards the surface. When the implantation voltage is increased, the Ni peak shifts more deeply into the substrate and the concentration of the pileup peak becomes higher. The Ni peak always appears behind the oxygen distribution as shown in Fig. 7. It appears that oxygen shoves or snowplows Ni into the substrate. Another interesting observation is that Ni can hardly be detected in the oxygen-rich zone, 18 possibly because Ni has a smaller affinity to oxygen or nitrogen. IV. DISCUSSION Our experimental evidence shows the presence of surface oxygen incorporation, nitrogen diffusion, and redistribution of nickel and chromium during this process. The kinetics of surface oxidation depends very much on the process parameters including oxygen partial pressure, implantation energy, and ion current density. The evolution of the oxide layer is controlled by sputtering and oxidation occurring simultaneously. 19 The oxide growth rate R IG is limited by the flux of oxygen atoms from the ambient onto the surface of the substrate. This flux is given by the kinetic gas theory: R IG 2P(O 2 )/ 2 mkt, where P(O 2 ) denotes the oxygen partial pressure, k is the Boltzmann s constant, T is the temperature, and m is the mass of oxygen. Meanwhile, oxide removal is due to sputtering. The oxide removal rate R SR is controlled by the ion energy and ion flux according to R SR j ion f (E)Y o (E)dE, where j ion is the total ion flux, f (E) is the ion-energy distribution, and Y o (E) is the oxygen sputtering yield. The oxide growth and removal processes compete with each other. In our cases generally R IG R SR thereby yielding a surface oxide layer. When the implantation voltage goes up, the average value of R IG /R SR also J. Vac. Sci. Technol. A, Vol. 19, No. 3, MayÕJun 2001

4 1011 X. Tian and P. K. Chu: Low-pressure, elevated-temperature plasma 1011 FIG. 7. Oxygen and nickel depth profiles acquired from the three samples. The oxygen concentration shown here has been divided by 4. increases. A higher implantation voltage and the corresponding lower ion current in order to keep the sample temperature constant result in a decrease in R SR since both j ion and Y o (E) decrease. However, the oxygen partial pressure does not change with the voltage as the oxygen source is the residual vacuum, and consequently the variation of R IG is small. Hence, the oxide thickness increases with the voltage. In addition, while the ratio of R IG /R SR changes with the implantation voltage, the relationship of R IG R SR is always valid although it may approach a dynamic equilibrium, thereby leading to a thicker oxide layer. During conventional elevated temperature PIII treatment of stainless steel, the nitrogen plasma is usually produced by a rf plasma source at higher gas pressure compared to that used in the hot filament glow discharge experiments reported here. Thus, the oxygen partial pressure is different in these two processes. For example, if the working pressure in the vacuum chamber is 0.5 Pa for rf plasma and 0.05 Pa for hot filament glow discharge, the oxygen partial pressure is higher by a factor of 10 in the hot filament glow discharge case as less nitrogen is bled in. As reported by Parascandola et al. 20 there is indeed a difference in the oxide thickness due to this factor. In addition, hot filament glow discharge requires that the filaments be hot to emit enough electrons to sustain the plasma. In general, the needed power is close to 1500 W for four sets of filaments. It will heat the nearby vacuum wall to a higher temperature compared to rf plasma thereby accentuating out gassing. In particular, the H 2 O partial pressure has been reported to increase with the surrounding temperature. 21 This will further increase the effective oxygen partial pressure and more severe surface oxidation will ensue. Hence, it is not surprising to observe thicker oxide layers in our experiments. Owing to surface oxidation that is unfortunately common in non-uhv environment typical of PIII and more severe at lower pressure, nitrogen penetration and diffusion are hampered. As shown in Fig. 6, the oxide thickness ranges from 30 to 90 nm, depending on the implantation voltage. According to TRIM 95 simulation, 22 the projected range of N 2 is 9 and 24.5 nm for 8 and 25 kv, respectively, assuming that the top layer is composed of Fe 2 O 3. That is to say, the incident nitrogen ions are mostly confined in the oxide layer. A surface oxide layer acts as a nitrogen transport barrier, and it is easy to imagine that a higher implantation voltage will yield a larger retained nitrogen dose on account of smaller sputtering. It should be noted that the surface oxide layer is not always deleterious and may positively affect the nitriding kinetics. Since it acts as a transport barrier, it may hinder the release of incident nitrogen. 20 For instance, we have achieved higher nitriding efficiency using titanium nitride deposited using a filtered vacuum arc source as a barrier. 23 It should be noted that the plasma characteristics and surface properties of the treated samples are also influenced by the plasma generation method, working pressure, and geometry of the instrument. The investigation by Baldwin et al. 12 has indicated that the density of electronically excited ions diminishes with increasing pressure, but the density of the excited neutral molecules peaks at a pressure of 0.3 Pa. Consequently, the surface hardness of the treated samples increases with pressure. 24 Blawert et al. have also demonstrated the influence of the pressure. 25 At higher pressure, more neutrals, and high-energy ions hit the surface during and between the high-voltage pulses, thereby enhancing the total nitrogen uptake in spite of sputtering. Our experimental results show that low-pressure elevatedtemperature hot filament glow discharge induces a higher oxygen partial pressure compared to rf discharge at higher pressure, thereby leading to higher surface oxidation. Surface oxidation is unfortunately frequently encountered in PIII experiments using typically non-uhv equipment and must be taken into account in the experimental design. Due to the intricate relationship among surface oxidation, sputtering, and nitrogen retention in the surface oxide layer, the experimental conditions for low-pressure elevated-temperature hot filament glow discharge PIII, particularly the oxygen partial pressure must be selected carefully as nitrogen diffusion into the bulk of the materials is stifled by the oxidized layer. A JVST A - Vacuum, Surfaces, and Films

5 1012 X. Tian and P. K. Chu: Low-pressure, elevated-temperature plasma 1012 higher voltage may enable a larger portion of the incident nitrogen ions to penetrate through the surface oxide layer for more rapid diffusion into the bulk of the materials. A lower implantation voltage generates a bigger R SR leading to thinner oxide layer and higher ion current for the same sample temperature, consequently resulting in higher nitriding efficiency Although the present work only deals with elevated temperature, it should be mentioned that low temperature PIII using hot filament glow discharge is also a practical method due to the smaller oxidation effect at low temperature. 29,30 V. CONCLUSION We have conducted plasma immersion ion implantation of AISI 304 stainless steel at elevated temperature and low pressure utilizing nitrogen plasma produced by hot filament glow discharge. AES and RBS results reveal that the nitrogen depth profiles depend on the implantation voltage and the retained dose increases with the implantation voltages. The phenomenon is believed to be due to severe surface oxidation that is common in non-uhv PIII machines. The effectiveness of the process depends on the working gas pressure that affects the oxygen partial pressure, plasma activity, plasma density, and the plasma solid interactions. Hence, the optimal process window must be selected carefully and attention must be paid to the oxygen partial pressure in the system, particularly for elevated-temperature processes. ACKNOWLEDGMENTS The work described in this article was jointly supported by Hong Kong RGC CERG No or CityU 1003/99E and No or CityU 1032/00E as well as City University of Hong Kong SRG No The authors are also grateful to S. Mandl and R. Guenzel for helpful discussions as well as S. P. Wong and W. Y. Cheung for RBS analysis. 1 J. R. Conrad, J. L. Radtke, R. A. Dodd, F. J. Worzala, and N. C. Tran, J. Appl. Phys. 62, J. Tendys, I. J. Donnelly, M. J. Kenny, and J. A. Pollock, Appl. Phys. Lett. 53, P. K. Chu, X. Lu, S. S. K. Iyer, and N. W. Cheung, Solid State Technol. 40, S W. Ensinger, Surf. Coat. Technol , G. A. Collins, R. Hutchings, and J. Tendys, Mater. Sci. Eng., A 139, C. Blawert and B. L. Mordike, Nucl. Instrum. Methods Phys. Res. B , E. Richter, R. Gunzel, S. Parascandola, T. Telbizova, O. Kruse, and W. Moeller, Surf. Coat. Technol , S. Mandl, R. Gunzel, E. Richter, and W. Moeller, J. Vac. Sci. Technol. B 17, R. Wei, J. J. Vajo, J. N. Matossian, P. J. Wilbur, J. A. Davis, D. L. Williamson, and G. A. Collins, Surf. Coat. Technol. 83, W. Moeller, S. Parascandola, O. Kruse, R. Gunzel, and E. Richter, Surf. Coat. Technol , S. Leigh, M. Samandi, G. A. Collins, K. T. Short, P. Martin, and L. Wielunski, Surf. Coat. Technol. 85, M. J. Baldwin, G. A. Collins, M. P. Fewell, S. C. Haydon, S. Kumar, K. T. Short, and J. Tendys, Jpn. J. Appl. Phys., Part 1 36, P. K. Chu, B. Y. Tang, Y. C. Cheng, and P. K. Ko, Rev. Sci. Instrum. 68, X. B. Tian, X. F. Wang, B. Y. Tang, P. K. Chu, P. K. Ko, and Y. C. Cheng, Rev. Sci. Instrum. 70, X. B. Tian, Z. N. Fan, X. C. Zeng, B. Y. Tang, and P. K. Chu, Rev. Sci. Instrum. 70, X. B. Tian, B. Y. Tang, and P. K. Chu, J. Appl. Phys. 86, X. B. Tian and P. K. Chu, Scr. Mater. 43, B. Garke, C. Edelmann, R. Guenzel, and J. Brutscher, Surf. Coat. Technol. 93, S. Parascandola, O. Kruse, and W. Moeller, Appl. Phys. Lett. 75, S. Parascandola, O. Kruse, E. Richter, and W. Moller, J. Vac. Sci. Technol. B 17, R. Wei, Surf. Coat. Technol. 83, J. F. Ziegler, J. P. Biersack, and U. Littmark, The Stopping and Range of Ions in Solids Pergamon, New York, 1985, Vol X. B. Tian and P. K. Chu, Proceedings of the 7th International Conference on Plasma Surface Engineering PSE 2000, Garmisch- Partenkirchen, Germany, Paper Or01:8, M. J. Baldwin, M. P. Fewell, S. C. Haydon, S. Kumer, G. A. Collins, K. T. Short, and J. Tendys, Surf. Coat. Technol. 98, C. Blawert, B. L. Mordike, G. A. Collins, K. T. Short, and J. Tendys, Surf. Coat. Technol , X. B. Tian, Z. M. Zeng, X. C. Zeng, B. Y. Tang, and P. K. Chu, J. Appl. Phys. 88, P. J. Wilbur, J. A. Davis, R. Wei, J. J. Vajo, and D. L. Williamson, Surf. Coat. Technol. 83, D. L. Williamson, J. A. Davis, and P. J. Wilbur, Surf. Coat. Technol , Z. M. Zeng, B. Y. Tang, P. K. Chu, X. B. Tian, S. Y. Wang, and X. F. Wang, J. Vac. Sci. Technol. B 17, W. Ensinger, J. Vac. Sci. Technol. B 17, J. Vac. Sci. Technol. A, Vol. 19, No. 3, MayÕJun 2001