• PDF / 486,893 Bytes
• 6 Pages / 420.48 x 639 pts Page_size
• 21 Downloads / 249 Views

DOWNLOAD

REPORT

---

MECHANISM OF ELECTRON INJECTION DURING THE ANODIC OXIDATION OF SILICON J.-N. CHAZALVIEL AND F. OZANAM Laboratoire PMC, Ecole Polytechnique, Route de Saclay, 91128 Palaiseau, France.

ABSTRACT n-Si photoanodes have been found to exhibit photocurrent multiplication during the first seconds of exposure to a fluoride-free, acidic electrolyte. This shows that, in contrast with earlier hypotheses, photocurrent doubling is not directly related to the presence of fluoride in the electrolyte, but rather must arise from an electron injection mechanism associated with the Si-H bonds initially present at the Si surface. It also suggests that the electroluminescence which has been observed during the anodic oxidation of porous silicon most probably stems from the same electron-injection mechanism.

INTRODUCTION A strong electroluminescence from porous silicon on a p-Si substrate has been observed during its anodic oxidation in 1M aqueous KNO 3 or HCI.1 This indicates the existence of an efficient electron injection mechanism at this interface. Such mechanisms may occur in multielectron electrochemical reactions, typically when the oxidation path involves a highly reducing intermediate. 2 An alternate manifestation of such mechanisms is the associated phenomenon of photocurrent doubling, which may be observed on n-type photoanodes (see Fig.1). 2 The occurrence of an efficient electron-injection process in the anodic oxidation of silicon is yet rather surprising, as this reaction is well3 known to proceed via the valence band (i.e., four valence holes per oxidized silicon atom). It has been proposed that the electron injection might be associated with the oxidation of the Si-H bonds, Iwhich are initially present on the surface of porous Si. 4 - 6 On the other hand, electron injection in fluoride electrolytes has long been known to induce photocurrent doubling on (flat) n-Si photoanodes, and has been attributed to the attack of partially oxidized silicon atoms at the interface by fluoride ions. 7-9 At this point one might wonder whether these two observations of electron injection have a common origin. Specifically, if the oxidation of the Si-H bond results in an electron injection, such a process should occur even in a non-fluoride electrolyte, on a transient time scale. We have addressed this question by using transient photocurrent measurements.

EXPERIMENTAL We have used a cell allowing the recording of the current just after the electrode is put into contact with the electrolyte. The design, allowing this contact to be realized in -10-2 s, is shown in Fig.2. It is similar to a previous design,' 0 but it further allows illumination of the electrode through a window. The electrode was n-Si (111) with doping ND=-10 15 cm- 3 . It was carefully polished using mechanochemical polishing,1 0 cleaned with sulphochromic mixture, and rinsed for a few seconds in 40% HF before each experiment. Several acidic electrolytes have been used: IM HCl, 0.5M H 2 SO4, IM HC1O 4 and IM KCI buffered to pH=2.2 (O.lM phosphate buffer). The electrochemi