Silicide deposition methods

For some time, difficulties with the deposition of silicides were preventing successful large volume implementation. Several techniques have been tried  

For a more extensive comparison of these techniques see Ahn et al.213 and Crowder214. 

After 1983, CVD-WSi2 became because of the above mentioned reasons popular and is now worldwide in use in large volume production (almost exclusively in polycide applications). Besides CVD also sputtered WSi2 is still in use in production. In the following sections we will elaborate on the CVD technique. 

There are several methods possible with the CVD technique to come to WSix films as we will mention below. Lehrer and Pierce218, have described an interesting approach in which they do a sequential deposition of Si and W. The chemistry used was  

Both reactions were carried out in a cold wall reactor at 600°C and at atmospheric pressure. First, 400nm of silicon was deposited followed by 65nm of W. After an anneal at 1000°C in Ar for 10 min a thin film resistivity of about 100 /iflcm was obtained. This high value was due to oxygen incorporation during the W film deposition. 

---

In 1983 an extensive review of silicides for IC s applications was published by Murarka86. This work focused mainly on sputter techniques as the deposition method. As we will see, after 1983 the CVD technique became the most popular method for polycide applications. Much literature on the technique itself and on the film properties has been published. It seems appropriate here to summarize from the literature the most significant results reported after 1983 on CVD-WSix. 

Protection of niobium and its alloys from oxidation in air is accomplished by coating, e.g. with zinc deposited by holding in zinc vapour at 865°C or coating with a layer of chemically stable oxide, nitride or silicide. Silicide coatings applied by pack cementation, fused slurry or by electrolytic methods have been found to be one of the most effective means of preventing oxidation of the metal. 

Metal and polysilicon films are formed by a chemical-vapor deposition process using organometallic gases that react at the surface of the IC structure. Various metal silicide films may also be deposited in this manner by reaction with the surface of the silicon wafer to form metal silicides. Glass and pol3uner films are deposited or spin cast or both, as are photoresist films (those of a photosensitive material). This process is accomplished by applying a liquid polymer onto a rapidly rotating wafer. The exact method used varies from manufacturer to manufacturer and usually remains proprietary. 

On a metal surface, silicide layers can be formed by two methods. In the first, Si atoms are vapor deposited by heating either a well degassed silicon wafer or a silicon rod to near its melting point. In the second method the metal is heated in 10 to 50 mTorr of silane for a desired length of time, usually about 10 to 60 s at a desired temperature, usually about 300 to 700°C. The first method is better suited for studying very early stages of silicide formation, the second more convenient for growing thick layers of silicides. Chemical vapor deposition or laser enhanced chemical vapor deposition may probably be used also, but have not yet been explored. 

Figure 20.14 [13] shows a second method of converting the polygate to silicide for high-k applications. Here again, the CMOS transistors are formed with conventional poly-silicon gates. After ILDO deposition, the surface is polished back to expose the poly-silicon. Metal films are then deposited and... 

When silicon is deposited from the vapor phase at ambient temperature, it solidifies as amorphous silicon. Vapor deposited bilayers and multilayers of silicon with metals thus consist of polycrystallinc metal and amorphous silicon. The earliest observations of amorphous silicide formation by SSAR were made on such diffusion couples [2.51, 54], Similar results were also obtained earlier by Hauser when Au was diffused into amorphous Tc [2.56], Figure 2.15 shows an example of an amorphous silicide formed by reaction of amorphous silicon with polycrystallinc Ni-metal at a temperature of 350"C for reaction times of 2 and 10 s [2.55,57], The reaction experiments were carried out by a flash-healing method (see [2.55] for details). In this example, the amorphous phase grows concurrently with a crystalline silicide. The amorphous phase is in contact with amorphous Si and the crystalline silicide in contact with the Ni layer. As in the case of typical mctal/metal systems, the amorphous interlayer is planar and uniform. It is also interesting that the interface between amorphous silicon and the amorphous silicide appears to be atomically sharp despite the fact that both phases are amorphous. This suggests that amorphous silicon (a covalently bonded non metallic amorphous phase with fourfold coordinated silicon atoms) is distinctly different from an amorphous silicide (a metallically bonded system with higher atomic coordination number). These two phases are apparently connected by a discontinuous phase transformation. 

---