Advanced Lithography for Nanofabrication
- PDF / 2,280,315 Bytes
- 6 Pages / 414.72 x 648 pts Page_size
- 70 Downloads / 193 Views
57 Mat. Res. Soc. Symp. Proc. Vol. 448 01997 Materials Research Society
EXPERIMENTAL The scanning electron beam lithography (SEBL) system consists of a commercial SEM, Amray 1400, configured for external scanning. The external scanning is driven by a 16 bit D/A converter controlled by a 486-50MHz personal computer. With a writing field of 100 jm, the D/A converter has a pixel resolution of 1.5 nm which is much less than both the beam diameter (5- 10 nm with a LaB 6 emitter) and the electrical noise (S/N ratio of - 120 dB) from the scanning coils. Larger writing fields with lower pixel resolution are also possible. Typical beam currents used were on the order of 5 to 100 pA. The electron dose for our application was 1-4 pC/cm. For comparison, the typical PMMA EBL dose is only a few nC/cm. The pressure in the sample chamber is in the low 10 6 to high 10' Torr range when both the turbo pump and ion pump are used. RGA monitoring of the sample chamber did not show detectable contamination. The base line performance capabilities of the EBL were determined using conventional resist based EBL. Minimum linewidths of 50 nm were routinely achieved using PMMA and gold liftoff. Note that we used relatively thick PMMA (-250 nm) and substrate (~300 pm) for the gold liftoff. Higher resolution is possible with better noise isolation and both thinner resist and substrate. Following standard solution cleaning using trichloroethylene, acetone, methanol, and DI water rinse, the Si substrates were oxidized in a UV photoreactor to remove the remaining hydrocarbons and to oxidize any metallic impurities. The silicon hydride layers were prepared by a dilute (- 1017) HF dip of the UV oxidized samples at room temperature. After the HF dip the surface is passivated by a uniform silicon hydride layer. Hydride passivation of the surface was confirmed by scanning Auger electron microprobe analysis. No surface contaminants such as carbon or oxygen were observed. Upon electron beam irradiation, the hydrogen on the Si surface is desorbed and a highly reactive Si surface is produced that is the basis for pattern formation. The exposed area can either serve as a positive or a negative patterning mask depending on subsequent processes. One can oxidize the exposed area followed by CVD on the unexposed area. On the other hand, one can inject a source gas into the sample chamber concurrently with electron irradiation. In this work, we used wet chemical etching to transfer the oxide pattern to the Si substrate. A 101% tetramethyl-ammonium-hydroxide (TMAH) at 70'C is used as a Si etch. It is a highly anisotropic etch that etches Si(100) and Si(1 10) three times as fast as Si(1 11); for Si(1 10) the etch rate is 6.8 nm/s while for Si( 11) it is 2.2 nm/s. In addition, the etch rate for SiO2 is at least an order of magnitude slower than for Si. To elucidate the mechanism of pattern formation, the effects of electron energy. electron dose and the substrate thickness on the resulting linewidth were investigated. RESULTS Figure 1 shows an AFM image of a patter
Data Loading...
