Effect of Epitaxially Deposited Gold Nanoparticles on Electroless Metallization of Silicon Wafers

Abstract

Listen

Translate article

Surface metallization of silicon substrates, that is, adhesive metal film formation on silicon is important for obtaining infallible electrical contacts (electrodes) in various devices such as solar cells and power devices. Autocatalytic electroless deposition, which is a conventional method to metallize non-conductive substrates, has several advantages like simplicity of process, uniformity of deposited films, and the covering of complicated structures. The autocatalytic electroless deposition is expected to replace sputtering process of back metal on power devices and screen printing process of electrodes on solar cells. This process requires a surface-activation pretreatment for metallization of non-conductive substrates (1). Adhesion of metal films obtained on silicon substrates by using conventional pretreatments is much lower than that is necessary for the electrical contacts of the devices. We recently developed a new surface-activation process for the direct electroless deposition of adhesive metal films on silicon substrates (2 - 5). This process consists of two steps: Step 1) formation of gold nanoparticles by electroless displacement deposition; and Step 2) metal film formation on the silicon surface by autocatalytic electroless deposition. In previous study, we reported that single crystalline gold nanoparticles are epitaxially deposited on silicon substrates (4) and silicon-gold alloy is formed at the interface between gold and silicon (3, 4). Those epitaxially grown gold nanoparticles cause the high adhesion of electrolessly deposited nickel films on silicon substrates (4). In addition, an amorphous layer exists between the metals and silicon substrate (3) and the contact resistance of the deposited metal films was approximately 1 m Ω cm2for highly doped silicon wafers (5). In this study, we investigate the reliability and interfacial structure change of electrodes produced on silicon substrates by the present method. At the first step of the metallization process, gold nanoparticles were deposited on p-Si (mirror polished Cz (100) wafer, 0.5~10 W cm) substrates by electroless displacement deposition using a HAuCl4 solution containing HF. At the second step, the nickel-phosphorous alloy film was electrolessly formed on the silicon substrates by using a solution containing NiSO4 as a metal salt and NaPH2O2as a reducing agent. The electrodes were formed on the silicon substrates by preparing a photoresist pattern before metallization. Contact resistance between the electrodes and silicon substrates was measured by a transfer length method (TLM) (5). For a reliability test, the electrode formed on silicon substrates were heat-treated at 423 K in the atmosphere for 3~2000 h. The contact resistivity was decreased from 2.36 to 1.22 W cm2by heat treatment at 423 K. The adhesion of metal films on silicon substrates was unchanged by the reliability tests for 3~2000 h. Figure 1 shows cross-sectional TEM image of the interfacial structure between the deposited metal film and silicon substrate after the test for 500 h. Amorphous layers of nickel silicide were more deeply formed into the silicon substrate than gold silicon alloy. The results of XPS measurements supported this structure and indicated the thickness of amorphous layers was almost unchanged between 500 and 2000 h of the test. The formation of nickel silicide was predicted to decrease the contact resistivity and prevent diffusion of gold. In conclusion, reliable electrodes are formed on silicon substrates by electroless deposition catalyzed with epitaxially deposited gold nanoparticles using electroless displacement deposition. ACKNOWLEDGEMENT The present work was partly supported by KAKENHI (26289276) from JSPS. REFERENCES 1. M. Paunovic and M. Schlesinger; Fundamentals of Electrochemical Deposition, 2nd ed., Wiley, NY (2006).2. S. Yae, M. Enomoto, H. Atsushiba, A. Hasegawa,N. Fukumuro, S. Sakamoto, and H. Matsuda; ECS Trans., 53 (6), 99 (2013). 3. H. Atsushiba, Y. Orita, S. Sakamoto, N. Fukumuro, and S. Yae; ECS Trans., 61(10), 9 (2014). 4. N. Yamada, H. Atsushiba, S. Sakamoto, N. Fukumuro, and S. Yae; ECS Trans., 69(39), 59 (2015). 5.Y. Orita, H. Atsushiba, M. Enomoto, T. Kimura, N. Fukumuro, H. Takagami, K. Kato, S. Sakamoto, M. Hirata, and S. Yae; ECS Trans., 61(10), 25 (2014). Figure 1