Controlled Sputtering Enables Better SiCr Films

Valery V. Felmetsger Sputtered Films Inc., Santa Barbara, Calif. -- Semiconductor International, 10/1/2000

		At a Glance
	An investigation of the influence of sputter deposition parameters on properties of thin SiCr resistive films deposited on thermally oxidized silicon wafers showed that the film's sheet resistance, uniformity across the wafer, and repeatability from wafer to wafer depended essentially on wafer and target surface temperature. The researchers also found the value and sign of the film's temperature coefficient of resistance (TCR) may be effectively controlled by varying the substrate rf bias power.

Thin films of tantalum nitride (TaN), nichrome (NiCr), and silicon chromium (SiCr) are widely used as thin-film resistors in manufacturing passive electronic components and various types of integrated circuits (ICs). Last year's cermet materials consisting of Si and Cr compounds such as Cr-SiO and CrB₂-Si-SiC are becoming more and more attractive for IC manufacturing because of their high resistance and stability. Sputter deposition technology now allows precise control of film composition, enabling manufacturers to create resistors with very low, reproducible temperature coefficient of resistance (TCR).¹

Recently, researchers reported they obtained SiCr thin-film resistors with high resistance and near-zero TCR by adjusting the target composition, sputtering and reactive gas pressure, deposition rate, and annealing temperature.² But more research on how sputter process parameters influence film properties is needed to develop specific approaches for fine tuning the film's TCR and to further advance resistor film sputter technology.

Sputter process features and investigation

We deposited resistive films with a thickness of about 60 Å and sheet resistance in the range of R_S = 1000 - 1300 W/sq on 100 mm diameter thermally oxidized silicon wafers. All depositions were done in the EndeavorAT cluster tool using the dual cathode dc magnetron series IV S-Gun with sputtering targets having a phase composition of CrSi₂-Cr-SiC. A special etch/degas module, equipped with an infrared radiant heater, enabled us to use a pre-deposition wafer degas and post-deposition anneal in a high-vacuum environment.

Thin SiCr films enable the use of low-power sputtering regimes to get a low deposition rate and ensure good film thickness control. Also, because targets for SiCr have a high electrical resistance, we wanted to avoid overheating the targets and the resultant changing of the target composition that high-power regimes can cause.

Our initial design of experiment included two cathode power levels: relatively low (300 W) and relatively high (600 W). Deposition rates for these recipes were measured as 3 and 6 Å/sec, respectively.

Each time we varied one of the main process parameters (power, Ar gas flow or substrate temperature), the targets were conditioned in discharge. We performed this step to stabilize a new sputtering surface composition and ensure a repeatable film stoichiometry. However, during deposition of a wafer sequence, we observed that average sheet resistance decreased from wafer to wafer. Theoretically, it may be explained that increasing the target temperature provokes changes in sputter yield or target surface composition. Therefore, prior to product wafer deposition we employed 10 minutes of sputtering in the working regime to allow targets to reach equilibrium temperature.

1. The higher the inner cathode (target) voltage during sputtering, the lower the film sheet resistance. (Source: Sputtered Films)

Our previous investigations have shown us that wafer degas and heating directly prior to film deposition are critical to achieve continuous, dense SiCr films with uniform, "smooth" resistance distribution on the surface, as well as a low and reproducible value of TCR. Therefore, in this work we employed a 40-sec wafer degas with a maximum temperature of 350°C.

Within a day after deposition, film sheet resistance increased by approximately 2% of the initial as-deposited value. A short (60-sec) post-deposition anneal in vacuum at a temperature of 350°C decreased the resistance by 14%. During storage in air the annealed films did not change their resistance (Table 1).

Films deposited with elevated temperature (with pre-heat) had practically the same resistance as "cold" films deposited by analogous recipes, but their sheet resistance deviation was less, and the resistance distribution on the wafer surface was essentially smoother. Films having an average resistance of 1100 W/sq showed a standard deviation of 0.45% and 1.4% for 100 and 200 mm diameter wafers, respectively (measurements were made with a CDE ResMap automatic 4-point probe).

Measuring TCR

We made TCR measurements in temperatures ranging over 20-150°C for film samples deposited at relatively high and low temperatures and cathode power of 600 and 300 W (we prepared five samples for each parameter combination). We detected that all the films had negative TCR in the range of 31 to 66 ppm/°C. Films sputtered at low cathode power and elevated temperature showed the smallest values of TCR (Table 2).

Because the data suggested the TCR value correlates with film structural features, we employed the following three approaches to develop our low TCR process:

.Deposition while applying rf substrate bias.

.Wafer pre-heat and/or heat during deposition.

.Post-deposition anneal in vacuum.

Table 3 presents data showing how substrate rf bias influenced the thermal stability of the SiCr film sheet resistance.

Films deposited with rf substrate bias had less sheet resistance and showed more thermal stability than films deposited without bias. TCR decreased about half with increasing rf power while keeping the negative sign and reaching a saturation value at relatively low power (70 W). These results indicate ion bombardment can play a positive role in the formation of thermally stable resistors. At the same time, it is interesting to note that post-deposition rf plasma treatment of deposited SiCr films led to an essential rise of the TCR.

Table 4 shows that pre-heat applied together with rf substrate bias can improve the film's thermal stability. However, excessive wafer heat during the sputter process is not acceptable because it leads to an increased film resistance. At temperatures above 450°C we discovered we were creating an insulator film.

Post-deposition anneal in vacuum has been found to be very effective for film TCR adjustment. This procedure enabled us to achieve films with a positive TCR, meaning it is possible to create films with close to zero TCR by optimizing the anneal process parameters.

Conclusion

The goal of this work was to research and develop effective approaches enabling in situ control of electrical-resistance parameters of thin SiCr composite films deposited by magnetron sputtering. We investigated sputter process features and defined the main parameters that influenced the film's sheet resistance, uniformity and repeatability. We found that dc magnetron sputtering at elevated temperatures with rf substrate bias allows fine tuning of the film's TCR. Also, we concluded that an anneal in vacuum immediately following film sputtering may be an effective approach to achieve thermally stable thin-film resistors. As a result of this research, we have developed a deposition process ensuring highly uniform 60 Å thick SiCr films with sheet resistance of 1100 W/sq and near-zero TCR.

1. Thin Film Precision Technology, Vishay Intertechnology Inc., www.vishay.com/brands/thin_film/tech_guide.html.


2. Fan Wu, A.W. McLaurin, K.E. Henson, D.G. Managhan, S.L. Thomasson, Thin Solid Films, 332 (1998) 418-422.

The author would like to thank P.N. Laptev for assistance in TCR measurements and very useful discussions. .