Влияние температуры формирования на морфологию por-Si, получаемого методом Pd-стимулированного химического травления

The process of Pd-assisted chemical etching of silicon in a solution containing HF and H2O2 was studied. The influence of factors such as etching duration and solution temperature on the morphology of the formed layers was investigated. It is shown that in process Pd - assisted etching, Pd nanoparticles remain on the walls and bottom of the pores. Those structures, as was demonstrated in early works, have the property of electro-oxidation of ethanol, which gives grounds to assert that the formed structures are Schottky-type structures. Using the electrochemical equilibrium diagram in the Si-HF system (aq.), a model of Pd-assisted etching was determined. It is shown that the polishing dissolution of Si occurs without the formation of intermediate products (SiO2).

Модификация рельефа n-поверхности AlGaInN-светодиодов изменением состава газовой смеси при реактивном ионном травлении

The kind of profile produced during the reactive ion etching of AlGaInN light-emitting-diode (LED) heterostructures on the surface that became free after removal of the growth substrate is studied in relation to the composition of the gas mixture used in the etching process. It is shown that using a mixture composed of Cl_2 and Ar, taken in a 3:2 ratio in terms of flow rates, leads to the thinnest profile, whereas a 2 : 1 gas mixture of BCl_3 and Ar provides the largest structural elements. To study the effect of the kind of profile on the quantum efficiency (QE), flip-chip LEDs are fabricated on a silicon substrate. The LEDs are etched in different modes after the growth substrate is removed. Etching in the Cl_2:BCl_3:Ar mixture with a flow ratio of 6:10:11, which leads to intermediate sizes of the etching profile elements, is optimal for obtaining maximum light extraction from a LED chip at a wavelength of 460 nm. The variation of the kind of profile with the gas-mixture composition suggests that the profile parameters can be tuned to the wavelength used. An analysis of how the QE of LED chips depends on the etching duration in the three-component mixture under consideration results in that the optimum etching duration is estimated to be ~30 min. The results of the study can also be of use in the search for conditions minimizing the reflection of incident light by a chip, e.g., for photodetectors.

Фотоанодирование n-Si в присутствии перекиси водорода: зависимость от напряжения

The fundamental aspects of the electrochemical etching of lightly doped n -Si under backside illumination conditions are studied in a solution with a low HF concentration and a high concentration of hydrogen peroxide. The data obtained are compared with those for a control electrolyte containing no H_2O_2. The morphology of the self-organized macropores, their growth rate, porosity, effective valence, and the amount of dissolved silicon are examined in relation to the applied voltage. The anodization kinetics at low and high bias voltages is analyzed. It is found that, under the same illumination, the initial photocurrent in the peroxide electrolyte is approximately twice lower than in the aqueous electrolyte, which makes it possible to state that the quantum efficiency of the photocurrent is lower. As, however, the etching duration is made longer, the current in the peroxide electrolyte strongly increases to become higher than that in the control electrolyte based on H_2O. It is found that, in the presence of H_2O_2, the depthwise growth rate of the macropores increases by more than a factor of 2, and the porosity decreases. The vertical macropore channels have a diameter smaller than that for macropores formed in the aqueous electrolyte and their walls are poorly passivated, which causes branching and the formation of secondary mesopores, the number of which grows with increasing voltage. The effective valence of silicon dissolution in the presence of H_2O_2 decreases to less than 2. The results are interpreted in terms of the Gerischer and Kolasinski models.