Clusters in a Silane Glow Discharge: Mechanism of Their Formation and How to Avoid Them
- PDF / 346,452 Bytes
- 6 Pages / 414.72 x 648 pts Page_size
- 56 Downloads / 249 Views
CLUSTERS IN A SILANE GLOW DISCHARGE: MECHANISM OF THEIR FORMATION AND HOW TO AVOID THEM S.Veptek, O.Ambacher, W.Rieger, K.Schopper and M.G.J.Veptek-Heijman, Institute for Chemistry of Information Recording, Technical University Munich, Lichtenbergstr.4, D-85747 Garching/Munich, Germany ABSTRACT The formation of clusters in glow discharge plasma processing is of great concern with respect to the production yield. Their appearance, trapping and transport in silane plasmas have been the subject of several publications, but little is known about the mechanism of their formation and growth and how to avoid them in intense discharges used for high rate deposition of amorphous silicon. We present mass spectrometric and light scattering data and theoretical modelling which show that the formation of clusters in a clean silane discharge (total impurity = 10 ppm) is due to a sequential growth of higher silanes SinH 2n+2 with a strong, catastrophic-like onset starting from pentasilane. A selfconsistent mechanistic model will be presented together with a discussion of alternative ionic mechanisms. Several possibilities of how to avoid the cluster formation at high deposition rates of a-Si will be discussed and documented. INTRODUCTION The cluster formation during plasma processing of semiconductors (etching and deposition) is of great concern with respect to the contamination of the material and degradation of the production yield. Due to the non-equilibrium nature of the glow discharge plasma (Te>>Ti) the clusters - whenever formed - will be negatively charged and consequently trapped within the positive space charge region. As the plasma is positive with respect to the electrodes and walls, the clusters have sufficient time to grow up to micrometer sizes before they are swept away either by gas flow (in typical plasma processing units with parallel plate electrodes, there are various regions of gas flow stagnation), gravity (see [1,2] and references therein) or space-charge instabilities [3]. Even if the clusters remain confined within the plasma during the process, they will fall down and contaminate the Si-wafer or the deposited film after the discharge has been switched off. Significant progress in the understanding of the behaviour of the clusters in the discharges and in the development of the techniques for their removal has been achieved during the last few years which resulted in a significant improvement of the cleanliness of the plasma etching process [1]. The mechanism of the cluster formation remains still subject of many discussions, studies and hypotheses. This is particularly embarrassing in plasma CVD where the cluster formation usually limits the maximum achieveable deposition rate. Several groups have shown that, in the case of amorphous silicon (a-Si) deposition, the clusters appear when the deposition rate is increased by increasing the plasma density ([4] and references therein). We shall show that this is not necessarily generally valid. An understanding of the reaction mechanisms of the a-Si deposition
Data Loading...
