Converting nickel film to nano particles using hydrogen plasma treatment
11/01/2007
EXCUTIVE OVERVIEW
Nickel particles on silicon wafers are popularly used as catalysts in growing carbon nanotubes (CNTs). Hydrogen plasma treatment is the typical process in converting nickel film into particles. Experimental results in this work indicate that the dominant effect of the hydrogen plasma in the treatment of nickel particles on silicon wafers is plasma-enhanced coalescence, rather than the mechanism of plasma etching. This information helps in tuning the surface morphology of the nickel catalyst process as well as in the subsequent CNT growth.
CNTs were discovered as a new carbon structure in 1991 [1] and are made of graphite sheets rolled into seamless tubes. CNTs are usually classified as 1D nano materials because of their much smaller diameter compared with their length. Due to low threshold field emission [2], high-emission current density [3], and stability [4], CNTs potentially have application as emitters in the flat panel display industry.
Several parameters affect CNTs’ density and diameter, but the particle density and diameter of the catalyst in growing CNTs are the two main factors. Methods for obtaining small-volume catalyst particles have been reported [5-7]: 1) coating the catalyst film first, followed by heat treatment, with the film coalescing into particles; 2) coating the catalyst film followed by plasma treatment, after which the film starts to form particles; and 3) producing catalyst particles from organic metal material.
Wei et al. [5] heated nickel films of different thicknesses at a temperature of 660°C for 3 min. The researchers found that for the heat-treated film, the particle diameter tends to decrease with film thickness. Choi et al. [6] used a hydrogen microwave plasma at different power levels to treat a nickel catalyst film. Dense and uniform nickel particles were clearly seen with the appropriate microwave power level and process time. Pan et al. [7] used a solution containing iron to uniformly coat a silicon wafer, and when the wafer underwent heat treatment in a hydrogen environment, the result was iron particles ranging in size from 5-50nm. By comparing the above three methods, a rather uniform particle size can be obtained by using heat or plasma treatment on a catalyst film.
Reaction mechanism
This work focuses on discovering the reaction mechanism of hydrogen plasma treatment on nickel film as it transforms into nickel particles. Plasma-surface interactions are very complex, with many of the reactions happening simultaneously. Plasma etching, an often-used plasma surface interaction in semiconductor manufacturing, demonstrates the effect of the bombardment of energetic ions on a surface. However, a plasma-enhanced chemical reaction at low temperature is another often-used interaction. The term “plasma-enhanced” refers to the use of ions, radicals and excited species in plasma that lower the activation energy of the chemical reaction.
Based on our research, it is believed possible to infer two mechanisms responsible for the conversion of film into particles when treated with a hydrogen plasma. The first mechanism is that of plasma etching during which the hydrogen plasma etches the nickel film from top to bottom, producing nickel particles. The other mechanism is called plasma-enhanced coalescence, where the hydrogen plasma etching only provides the “cracking” of the nickel film. It is our contention that the plasma etching induced the stress cracking of the film, thereby enabling the formation of the nickel particles.
For this research, the plasma reaction mechanism was studied by observing the resulting change in the surface morphology of a nickel film treated with only microwave hydrogen plasma in some cases, and plasma followed by heat treatment for other samples. Previous studies in our laboratory indicated that when starting with a 2nm thick nickel film, independent nickel particles result when the film has been treated for 10 min with fixed microwave plasma parameters: 100sccm hydrogen gas flow, 2.5 × 103Pa process pressure, and microwave power at 900W. Different film thicknesses, as well as different treatment times, were tested under the same fixed plasma parameters mentioned above. If plasma etching is the dominant process, then the resulting sizes of the hydrogen plasma-treated particles should be similar-even if the starting film thicknesses are different. This should be the case because the treatment time, which is equal to the film thickness divided by the fixed 0.2nm/min plasma treatment rate [i.e., treatment time = (film thickness)/(0.2nm/min)], should be similar for both cases. However, we observed that plasma etching is not the dominant process. Instead, the process that plays the dominant role in particle formation is that associated with the plasma-enhanced coalescence mechanism. An apparent “coalescence effect” was found to occur after the hydrogen plasma treatment of the nickel film.
Experiments
N-type silicon samples, each with an area of 0.5cm × 0.5cm were prepared. Samples were cleaned with acetone and isopropanol solutions in series. A nickel catalyst film was sputtered on samples with the same DC sputtering tool. The tool was pumped down to a base pressure of 1 × 10-3Pa. With process conditions of 3sccm Ar, 4 × 10-1Pa process pressure, and 100W DC power, different film thicknesses were obtained by varying the sputtering time. A near constant deposition rate of 1nm/3.5 sec was obtained after some pre-tests.
|
Nickel-coated samples were put into a microwave plasma system, which was pumped down to 1 ×10-3Pa. The fixed plasma treatment parameters were 100sccm hydrogen gas flow, 2.5 × 103Pa process pressure, and microwave power at 900W. The samples were subjected to different plasma and heat treatment recipes as listed in the table. Samples 1-3 only received a plasma treatment. The film thickness divided by the corresponding treatment time of samples 1-3 is equal to the expected plasma treatment rate noted before: 0.2nm/min.
In addition to the same plasma recipe as samples 1-3, samples 4-6 received a 500°C heat treatment for one hour. Samples 7-10 only received a plasma treatment. The corresponding film thickness of samples 7-10 is greater than the product of treatment time and the expected 0.2nm/min treatment rate. The surface morphology of the samples was checked by field emission scanning electron microscope (SEM) (JEOL JSM-6700). The role of hydrogen plasma in treating nickel film was judged by comparing the morphology of samples 1-3. The appearance of samples 1-3 should be similar if the plasma etching is the dominant mechanism. However, if the particle size sequence were found to be such that the size of particles in sample 1 was less than the size of particles in sample 2, and the size of particles in sample 2 was in turn less than those in sample 3, then the conclusion would be that plasma-enhanced coalescence is the dominant process during the hydrogen plasma treatment. To restate the comparison process: the effect of the heat treatment on the nickel film was investigated by comparing the morphology change between samples 1 and 4, 2 and 5, and 3 and 6. The role played by the plasma was further checked by observing the morphology of samples 7-10. If plasma etching is the dominant mechanism, the film cannot be totally etched away because the film thickness with respect to samples 7-10 is greater than the product of the plasma treatment rate (0.2nm/min) × the corresponding treatment time. Still, a continuous film will be observed. On the other hand, if plasma-enhanced coalescence dominates, instead of a continuous film, independent particles may appear (samples 7-10).
Results
Plasma treatment. After plasma treatment was applied to sample 1, the nickel film had transformed to similar-shaped, isolated, round particles that were 20nm dia. Isolated polygon-shaped particles with a diameter of 93.5nm were seen for the plasma-treated sample 2. Particle size increases up to 150nm in dia. were observed in plasma treated sample 3.
The above results imply that the plasma-enhanced coalescence mechanism dominates the conversion of nickel film into particles when using a hydrogen plasma treatment. The surface morphology data from samples 1-3 after plasma treatment shows an apparent coalescence effect [8], with the particle size growing with increasing plasma treatment time. If plasma etching was the main mechanism, the same plasma treatment “rate” should produce a similar particle size; however, this was not observed for samples 1-3.
Plasma+heat treatment. Surface morphology for samples 4-6, which were processed with plasma followed by heat treatment, is shown in Figs. 1a-c, respectively. With the morphology information for samples 1-6, the particle size corresponding to different thicknesses is shown in Fig. 2. For thicknesses of 10nm and 20nm, heat treatment makes the particle size increase, but this does not happen at a film thickness of 2nm. This increase in the particle size corresponding to 10nm and 20nm film thicknesses implies that these solid particles melt during 500°C heat treatment. The melting point of bulk nickel is 1453°C [9], but the heat treatment temperature is only about one-third of that (500°C). This phenomenon can be explained by Lindemann’s effect [10]: when the film thickness decreases to the nanometer size range, the melting point of the solid material may decrease to less than one-third of its original value. The morphology for samples 1 and 4 is similar, which implies that the 20nm particle size in Fig. 1a that corresponds to the 2nm film thickness does not increase after heat treatment. This behavior can be explained by the “critical radius” concept and in this work, the phenomenon usually occurred when the film thickness was in the range of several nanometers. The critical radius phenomenon is related to the Gibbs free energy change mechanism [11]. In order for particle growth to occur, the isolated particle size should be greater than the critical radius. Since the SEM images in Fig.1 are in two dimensions, the vertical length could not be seen. For a 2nm thickness sample, the vertical height of the produced particles is believed to be similar to the original film thickness located inside the critical radius range.
Figure 2. Particle size vs. different nickel film thicknesses treated with only plasma and plasma plus heat treatment |
Morphology of samples 7 and 9 was shown in Figs. 3a and 3b. Isolated particles can be clearly observed in Figs. 3a and 3b, even after only a 10 min plasma treatment. This phenomenon further indicates that the mechanism of plasma-enhanced coalescence plays a greater role in forming particles out of the nickel film than does plasma etching. As mentioned previously (refer to “Reaction mechanism” section), a 2nm nickel film thickness needs a 10 min plasma treatment time to change the film into particles. If the plasma etching mechanism dominates, samples of film thicknesses such as those in the range of 10 and 20nm (e.g., samples 7 and 9) that are subjected to a 10min. plasma treatment, should have a thickness of 2nm etched away; the remaining films after etching (i.e., 8nm and 18nm thick films) at the bottom of the original film stack will still be films, that is, the material will not have formed particles nor begun the process of particle formation. However, the reality does not match with this projected scenario as observed in Figs. 3a and 3b.
Figure 3. SEM images of nickel film: a) 10nm thickness, plasma treatment for 10min; b) 20nm thickness, plasma treatment for 10min. |
By comparing the morphology data of two groups: samples 7 and 8, and samples 9 and 10, it was observed that for the same thickness of nickel film, the longer the plasma treatment time, the rounder the shape of the nickel particles that are formed. This result again indicates that the plasma-enhanced coalescence mechanism is the dominant one when using a hydrogen plasma treatment. The morphology of samples 1 and 4 is similar, and even though the nickel film is subjected to the same plasma treatment time as samples 1, 7, and 9, the particle size increases as one can see when comparing the particles’ shapes in Figs. 1a, 3a, and 3b; the particles become progressively less round. It can be concluded that more energy is needed to complete the coalescence reaction when the film is thicker, than when it is thinner.
Conclusion
The mechanism associated with the formation of nano particles after a nickel film is treated with a hydrogen plasma has been studied. By cross-comparison of ten different sets of nickel film samples after plasma treatment alone, or plasma plus heat treatment, it was concluded that hydrogen plasma plays a much stronger role in the formation of particles (called plasma-enhanced coalescence), than does plasma etching. Heating to a temperature of 500°C, which is one-third of the melting temperature of nickel, can soften nickel nano particles, and this observation agrees with the Lindemann effect.
Acknowledgments
The author sincerely appreciates Jian-Hua Lee for assisting with instrument operation, and Tien-Chia Lin for valuable discussions in nickel film characteristics. Part of this work was supported by the National Nano Device Laboratories, Taiwan.
References
- S. Iijima, “Helical Microtubules of Graphitic Carbon,” Nature, 354, pp. 56-58, 1991.
- P. Collins, P. Avouris, “Nanotubes for Electronics,” Scientific American, pp. 38-45, Dec. 2000.
- T.W. Ebbesen, P.M. Ajayan, “Large-scale Synthesis of Carbon Nanotubes,” Nature, 358, pp. 220-222, 1992.
- J.H.Yen, I.C. Leu, C.C. Lin, “Effect of Catalyst Pre-treatment on the Growth of Carbon Nanotubes,” Diamond and Related Materials, 13, pp. 1237-1241, 2004.
- Y.Y. Wei, G. Eres, V.I. Merkulov, D.H. Lowndes, “Effect of Catalyst Film Thickness on Carbon Nanotube Growth by Selective Area Chemical Vapor Deposition,” Appl. Phys. Lett., 78, pp. 1394-1396, 2001.
- W.S. Choi, S. Choi, B. Hong, J. Lee, “Growth of Carbon Nanotubes on Glass Substrate by MPECVD,” Material Science and Engineering, C 26, pp. 1215-1218, 2006.
- Z.W. Pan, S. S. Xie, B.H. Chang, L.F. Sun, W.Y. Zhou, G. Wang, “Direct Growth of Aligned Open Carbon Nanotubes by Chemical Vapor Deposition,” Chemical Physics Letters, 299, pp. 97-102, 1999.
- S. Wang, P. Wang , O. Zhou, “Effects of NH3 Plasma Pre-treatment on the Growth of Carbon Nanotubes,” Diamond & Related Materials, 15, pp. 361-364, 2006.
- W.F. Smith, Foundations of Materials Science and Engineering, McGraw-Hill, 878, 2004.
- F.A. Lindemann, “Über die Berechnung Molecularer Eigenfrequenzen,” Z. Phys., 11, pp. 609-612, 1910.
- D.L. Smith, Thin-Film Deposition Principles and Practice, McGraw-Hill, Singapore, 148, 1999.
Shang-Chou Chang received his PhD at National Tsing Hua U. and is assistant professor in the Department of Electrical Engineering at Kun Shan U., No. 949, Da-Wan Rd., Yung-Kang City, Tainan Hsien, Taiwan; ph 886/6-2050-518, fax 886/6-2050-298, e-mail [email protected].
Tien-Chia Lin received his PhD at National Cheng Kung U. and is assistant professor in the department of electrical engineering, Kun Shan U.
Jian-Hua Lee received his MS at Kun Shan U. and is currently in military service.