(100) Silicon Etch‐Rate Dependence on Boron Concentration in Ethylenediamine‐Pyrocatechol‐Water Solutions

Abstract

The (100) silicon etch-rate dependence on boron concentration in ethylenediamine-pyrocatechol-~ater (EPW) solutions at 110~ has been measured by successive etching of boron-diffused silicon. The etch rate begins to decrease near 10'9cm -3 and decreases approximately as the fourth power of doping over three orders of magnitude in etch rate. The etch-rate ratio between heavily and lightly doped silicon appears insensitive to pyrazine concentration from 0.01 to 6.0g/liter E, oxygen exposure of the etch solution, and light. The etch-rate decrease is sensitive to hole concentration and not to atomic concentration of boron or stress. The etch-rate dependence may be explained by assuming that the hydrogen-evolution cathodic half-reaction of the oxidation-reduction reaction is the rate-determining step. The reason for the etch-rate decrease as the hole concentration increases is given in terms of increased Auger recombination of large numbers of electrons chemically produced at the silicon surface during etching. Excellent agreement with experiment is obtained. Similarities of this etch system and the phenomenon of silicon staining are discussed in terms of oxidation-reduction reactions. Application to control of thickness and uniformity of boron-doped silicon membranes is discussed.Boron-doped silicon membranes have been fabricated using ethylenediamine-pyrocatechol-water (EPW) etching solutions for a variey of purposes: as thin substrates to reduce backscattering in electron (1) or ion beam lithography (2), to reduce scattering of transmitted radiation in x-ray lithography (3), as thin highly doped barriers for Josephson junctions or super-Schottky diodes (4), as ink-jet nozzles (5), and for the fa~iqation of micromechanical beams used in accelerometers (6, 7). In almost all cases, boron diffusions are used to stop the EPW-etchant attack on silicon. Thus far, there has been very little work reported on the etch-rate dependence on boron or on the basic etchant system itself. Finne and Klein (8) first reported the etchant but did not find any etchrate dependence up to 8 • 1017 cm -3 boron. It was reported later that the etchant stopped at 7 • 1019 cm -3 (9), but there was no detailed measurement of the transition region. Reisman et al. (10) discovered the effect of pyrazine and oxygen-induced benzoquinone catalysts on the etch rate for lightly doped silicon in EPW solutions. This is important, since the pyrazine and quinone concentration in commercially available ethylenediamine and pyrocatechol, respectively, varies from lot to lot. They briefly report the etch-rate dependence on boron in bulk silicon wafers at two levels of heavy doping for high pyrazine/low benzoquinone etch solutions and find the etch rate to be small but finite in the 102o cm -3 range. Seidel and Csepregi (11) reported a stronger etch-rate dependence on doping in epitaxial layers than that reported by Reisman et al. They used a similar etch solution, except with somewhat lower pyrazine concentration and unknown benzoquinone content. Initial experime...

Etch rates for micromachining processing

“…The dependence of the reaction on p-type doping is explained by Seidel et al [30] and also by Raley et al [32]: At intermediate steps in the etch, four free electrons are generated that reside near the surface before being exchanged. P-type doping reduces this surface supply of electrons.…”

“…It also slows (110) (Other inorganic hydroxides [29], [30], organic hydroxides such as tetramethyl ammonium hydroxide (TMAH) [31], [33], and ethylenediamine pyrocatechol (EDP) [29], [30], an organic base, are orientation-dependent etchants similar to KOH. In the Berkeley Microlab and others, EDP has been found to be better than KOH at stopping abruptly at heavily boron-doped regions [29], [32]. TMAH has the advantages of not being a source of sodium (which contaminates the gate oxide in MOS circuitry) and not attacking aluminum when it has been "doped" with a small amount of silicon [33] …”

Introduction to MEMS

“…6.2(b) and (c). Moreover, the etching rate can be decreased by heavily doping boron into silicon [14,20] as well as electrochemically biasing a silicon p-n junction [21] . This adds an additional way of etching control besides the concentration and temperature of the etching solution.…”

Study of surface morphology and optical characterization of crystalline and multi-crystalline silicon surface textured in highly diluted alkaline solutions

“…Alkaline etching is anisotropic since it etches crystallographic planes in silicon with different rates and etches the (100) planes much faster than the (111) and (110) planes. This difference has been attributed to the different Si bond density, to the large differences in etching activation energies, to the etching retardation or halting caused by thin silicon oxide/ dioxide SiO x instantaneously grown on some planes and not on others after immersion the silicon into the etchant which retards or stops the nucleation of pyramids, and to the competition between the forward and reverse etch reactions especially when the etching is done at low KOH concentrations (Seidel et al,1990;Price et al, 1973;Palik et al,1985;Palik et al,1983;Glembocki et al,1991;Raley et al,1984;Kendall 1979;Kendall & de Guel 1985;Tan et al, 1996).…”