Effect of Surface Reactivity on Watermark Formation Studied By Sessile Droplet Evaporation Approach

Abstract: Wet nanoscale etching of silicon is very dependent to surface reactive sites like step edges, defect densities, and crystallographic planes [1]. Ultra-pure water (UPW) used during rinsing and drying steps is etching silicon in the sub-nanometer range furthermore this silica residues will cause ring shape drying marks on the surface [2-3]. In this paper drying marks are characterized by physical and chemical characterization techniques  to evaluate the silicon etching process. The conversion of the  silicon hydrogen bond to silicon hydroxide is investigated on different crystal orientations to monitor the reactivity of the surface.

Experiment & Measurement: The p-type (100) and (111) 200 mm silicon wafers (SunEdison) used had a resistivity of 1-5 Ω.cm. The samples were cleaned with an HF/HCl mixture to obtain a hydrophobic surface with a contact angle of around 70^◦. Atomically smooth silicon (111) was prepared by a NH₄F (Sigma-Aldrich 99%) etching process in a closed oxygen free etching cell [4]. Double side polished silicon (100) and (111) wafer were used as a crystal in attenuated total reflection Fourier transform infrared spectroscopy (ATR-FTIR) to monitor the hydrogen termination. The 3D profile of the WM formed after evaporation of a UPW sessile droplet was measured using a KLA-Tencor High Resolution Profilometer [3].

Results & Discussion:The impact of the crystal orientation on the WM residue is studied by comparing the results of droplet drying on p-type Si (100) versus (111) in the dark (40% relative humidity). Although AFM analysis confirms a smooth surface finish (RMS = 0.2 nm) for the HF treated (100) and (111) surface, no clear atomic planes are observed. Fig. 1 shows the etch rate of silicon in the dark under clean room conditions. The etch rate is calculated based on residue volume (measured by profiliometry) divided by the droplets initial wetting area, and evaporation time. Fig. 2 shows the decrease in the amplitude of the silicon hydride peak, measured by ATR-FTIR spectroscopy, as a function of time (the rate of decrease is faster in the Si (100), case as might be expected). The fact that hydride termination of the surface disappears only slowly gives information about the etching mechanism, that will be discussed in the proceeding. The higher stability of the (111) surface is attributed to the higher number of back-bonds for Si-surface atoms in case of (111) (with 3 back-bonds) versus the (100) case (with 2 back-bonds). The experimental result confirms the chemical dissolution process but a difference in surface stability is observed. WM formation was studied in the dark since previous work [3] showed that light increases the etch rate markedly, thus complicating the interpretation. In order to investigate the impact of surface morphology on WM formation, atomically smooth surfaces were prepared by immersing a p-type Si (111) substrates (miscut angle <0.07º) in oxygen free  NH₄F. The results are shown in Figure 3. By comparing the result of HF-pretreated and NH4F-pretreated surfaces it is clear that the atomically smooth surface shows reduced reactivity compared to the rough HF last surface. Due to lower step edge density the chemical reactivity of the surface decreases and WM residue formation is suppressed.

Conclusion: This study showed the effect of surface imperfections on dissolution process by WM formation. It is shown that the silicon (111) dissolution process decreased in case of atomically smooth surface. The dissolution can be further suppressed by performing rinsing and drying process in dark ambient.

References

[1] M. Hines et al., Int. Rev. Phys. Chem. 20, (2001) 645.

[2] T. Miura et al., J. Appl. Phys. 79, (1995) 4373.

[3] A.H. Tamaddon et al. solid state phenomena, 219, (2014) 89.

[4] P. Allongue et al., Electrochimica Acta 45, (2000) 4591.