Erbium Silicide Growth in the Presence of Residual Oxygen

Abstract

The chemical changes of Ti/Er/n-Si(100) stacks evaporated in high vacuum and grown ex situ by rapid thermal annealing were scrutinized. The emphasis was laid on the evolution with the annealing temperature of (i) the Er-Si solid-state reaction and (ii) the penetration of oxygen into Ti and its subsequent interaction with Er. For that sake, three categories of specimens were analyzed: as-deposited, annealed at 300{\deg}C, and annealed at 600{\deg}C. It was found that the presence of residual oxygen into the annealing atmosphere resulted in a substantial oxidation of the Er film surface, irrespective of the annealing temperature. However, the part of the Er film in intimate contact with the Si bulk formed a silicide (amorphous at 300{\deg}C and crystalline at 600{\deg}C) invariably free of oxygen, as testified by x-ray photoelectron spectroscopy depth profiling and Schottky barrier height extraction of 0.3 eV at 600{\deg}C. This proves that, even if Er is highly sensitive to oxygen contamination, the formation of low Schottky barrier Er silicide contacts on n-Si is quite robust. Finally, the production of stripped oxygen-free Er silicide was demonstrated after process optimization

Silicide Formation Process of Er Films with Ta and TaN Capping Layers

“…Combining the XRD and TEM analysis results, it can be postulated that a significant amount of inward oxygen diffusion occurred in the Ta-capped sample, and it formed a crystalline Er-oxide phase and delayed the ErSi 2−x growth during the annealing process. Therefore, the amorphous layer formed in the middle of the whole film structure is not expected to be a pure Er−Si mixed state, but an Er−Si−O mixed state (probably Er-silicate), which was observed in other experiments using a Ti capping layer, 4,6 and is also confirmed by the following XPS analysis. After annealing at a higher temperature of 600 °C, an almost complete conversion of the Er−Si mixed layer into crystalline ErSi 2−x was achieved in the TaN-capped sample with a sufficient supply of Si atoms from the substrate (Figure 2c), coinciding with the XRD results shown in Figure 1a.…”

“…Combining the XRD and TEM analysis results, it can be postulated that a significant amount of inward oxygen diffusion occurred in the Ta-capped sample, and it formed a crystalline Er-oxide phase and delayed the ErSi 2– x growth during the annealing process. Therefore, the amorphous layer formed in the middle of the whole film structure is not expected to be a pure Er–Si mixed state, but an Er–Si–O mixed state (probably Er-silicate), which was observed in other experiments using a Ti capping layer, , and is also confirmed by the following XPS analysis.…”

“…Subsequently, the capping layers with a thickness of ∼20 nm were deposited in situ using TaN (99.5%) and Ta (99.99%) targets at a sputtering power of 100 W. To initiate the Er-silicidation reaction, all the samples were annealed at different temperatures ranging from 300 to 700 °C for 1 min using a rapid thermal annealing system. In order to minimize the amount of residual oxygen, 6 the chamber was pumped down for 10 min and purged by flowing high purity N 2 (99.999%) gas for 5 min before annealing. During annealing, a high purity N 2 flow was continuously applied to maintain a process pressure of ∼1 Torr.…”

“…Several researchers found that the surface defect density could be minimized by preamorphizing the Si substrate, 10,17 or by using a specific capping layer. 4,12−14 Many reports have addressed the oxygen incorporation, 4,6 and have suggested several defect formation mechanisms during the Er-silicidation process. 9,11,13,15,16 Because these studies were carried out using one type of capping layer with different experimental conditions regarding the deposition method, annealing atmosphere, and capping layer, different types of defects were observed, and the origin of each defect formation has still been in debate.…”

“…Its unique material characteristics are particularly attractive for source/drain contact applications in both conventional metal-oxide-semiconductor field-effect transistors (MOSFETs) and Schottky barrier MOSFFETs. , However, the most difficult challenge from a process point of view is that it is extremely vulnerable to oxidation during the silicide-forming annealing process in an inert gas atmosphere, in which a small amount of residual oxygen still exists. As a consequence, Er is deposited under high vacuum conditions with oxidation-preventing capping layers, such as Ti, − TiN, , Mo, W, and amorphous-Si. − Ultrahigh vacuum (UHV) annealing can be another possible route to preventing the detrimental oxidation of Er, , but it may not be a cost-effective solution for industrial device fabrication.…”