Phosphorus incorporation during Si(001): P gas-source molecular beam epitaxy:: Effects on growth kinetics and surface morphology

The effects of P doping on growth kinetics and surface morphological evolution during Si(001):P gas-source molecular beam epitaxy from Si2H6 and PH3 at temperatures T-s=500-900 degrees C have been investigated. With increasing PH3/Si2H6 flux ratio J(P/Si) at constant T-s, we observe a decrease in the film growth rate R and an increase in the incorporated P concentration C-P, both of which tend toward saturation at high flux ratios, which is accompanied by increased surface roughening and pit formation. At constant J(P/Si), R increases with increasing T-s, while C-P initially increases, reaches a maximum at T-s=700 degrees C, and then decreases at higher growth temperatures. We use in situ isotopically tagged D-2 temperature programed desorption (TPD) to follow changes in film surface composition and dangling bond density theta(db) as a function of J(P/Si) and T-s. Measurements are carried out on both as-deposited Si(001):P layers and P-adsorbed Si(001) surfaces revealing beta(1) and beta(2) peaks due to D-2 desorption from Si monohydride and dihydride species, respectively, as well as the formation of a third peak beta(3) corresponding to D-2 desorption from mixed Si-P dimers. Dissociative PH3 adsorption on Si(001) results in a decrease in theta(db) and an initial increase in P surface coverage theta(P) with increasing T-s. Saturation theta(P) values reach a maximum of similar to 1 ML at T-s=550 degrees C, and decrease with T-s>600 degrees C due to the onset of P-2 desorption. Comparison of theta(P)(T-s) results obtained during film growth with postdeposition C-P(T-s) results reveals the presence of strong P surface segregation. From measurements of theta(P) versus C-P in Si(001):P layers grown as a function of T-s, we obtain a P segregation enthalpy Delta H-s=-0.86 eV. By using the combined set of results, we develop a predictive model for C-P versus T-s and, J(P/Si) incorporating the dependence of the PH3 reactive sticking probability S-PH3 on theta(P), which provides an excellent fit to the experimental data. (c) 2008 American Institute of Physics.