We report on the electroluminescence from p-type porous silicon due to minority carrier injection from an electrolyte solution. The MV ϩ• radical cation formed in the reduction of divalent methylviologen is able to inject electrons into the conduction band of crystalline and porous silicon. The electrochemistry of this redox process at silicon electrodes is briefly described, and electroluminescence due to recombination of the injected electrons with holes from the substrate is described. The results are discussed in terms of a semiconductor model. © 1995 American Institute of Physics.One of the main reasons for the extensive research on porous silicon is the prospect of integrating optical functions into all-silicon devices. To make a device capable of showing electroluminescence ͑EL͒, electrons and holes have to be injected into the porous silicon structure. Radiative recombination of these carriers will give rise to light emission. The first luminescent solid state devices based on this material 1-3 consisted of a porous layer on a crystalline substrate. A semitransparent Au or transparent conducting oxide electrode was deposited on top of the porous layer. These devices were reported to have very low quantum efficiencies (Ͻ10 Ϫ6 ) and require high operating voltages. More recently, different designs 4,5 and better contact materials 6-8 have led to some improvement in the electrical and luminescent properties. The reasons for the low EL efficiency are not understood, but it has been suggested that the ease with which carriers are injected into luminescent states and the transport mechanisms in the porous silicon play an important role. 9Much higher efficiencies were obtained using a liquid electrolyte junction which contacts the whole porous structure. [10][11][12] The mechanism of the strong visible emission from n-type porous silicon is thought to be similar to that for conventional semiconductors. Electron capture from the valence band, i.e., hole injection, is achieved by a strong oxidizing agent ͑electron acceptor͒ in the solution. At appropriate bias the injected holes recombine with electrons from the conduction band. If this recombination is radiative, EL characteristic of the semiconductor can be observed. 13A similar mechanism should hold for p-type semiconductors, only now electron injection into the conduction band is required. This is possible if the occupied energy levels of the redox couple in the solution overlap with the conduction band of the semiconductor. Thus, a strong reducing agent ͑electron donor͒ has to be used. To our knowledge there have been no reports of EL from p-type porous silicon due to minority carrier injection by a reducing agent in solution.14 Light emission during anodic oxidation of p-type porous silicon in indifferent electrolyte has been reported. 15 This has been attributed to electrons released during oxidation of Si-H surface species or of Si-Si bonds. However, this explanation may be suspect as luminescence is also observed under similar conditions at n-type porous...

From measurements using stationary and rotating disc and ring-disc electrodes, it is concluded that the reduction reactions of the divalent methylviologen cation MV 2-(to MV +" and MV °) proceed via the conduction band of both porous and crystalline silicon. The product of the second reduction step (MV °) forms a blocking layer on the electrode. The oxidation reactions of MV ° and MV ÷" take place by electron injection into the conduction band of n-type crystalline silicon. From the current-potential characteristics it follows that MV ÷" is also able to inject electrons into porous silicon formed on an n-type electrode. At p-type porous silicon, electroluminescence is observed as a result of radiative recombination of holes (majority carriers) with electrons (minority carriers) injected into the porous structure during oxidation of MV +'. Introducing ethanol into the solution leads to changes in the current-potential characteristics, which are explained by the solubility of MV ° in these solutions. The current-potential characteristics are considered on the basis of the position of the band edges, as deduced from impedance measurements, and differences between crystalline and porous silicon are discussed.