Strain introduction into group IV semiconductors makes them remarkably promising materials, opening possibilities of band engineering and resultant significant enhancements of electrical and optical properties toward ultra-low-power and high-speed microelectronic devices and photonic devices. So far, several techniques to induce strain have been developed. One is the growth of their heterostructures on a whole wafer, providing so-called global strain. Strain relaxed SiGe, (Si)GeSn and SiC buffer layers are used in many cases, where reduction of defects is indispensable. Other methods are local strain introduction by local formation of stressors, such as a SiN film and SiGe embedded source/drain for current strained Si MOSFETs. Every method has merits and demerits in terms of amounts of the strain, direction of the strain, uniaxial or biaxial, stability, uniformity, lateral area and defectivity. Here, various methods and materials are overviewed and compared. Since a strained Ge is particularly a promising material thanks to its inherent superior properties among group IV semiconductors, several examples of strained Ge structures are shown in the following. It has been shown that a compressively strained Ge offers much higher hole mobilities than strained Si devices [1,2]. The mobility enhancement is caused both by the band splitting and reduction of effective mass of the heavy hole. It was experimentally confirmed that the effective mass decreases with amounts of the compressive strain [3]. Extremely high hole motilities far beyond those of bulk-Ge has been obtained from strained Ge formed on SiGe virtual substrates [4,5], where contributions of parallel conductions through SiGe buffer layers are excluded by the mobility spectrum analysis. For device applications, however, the conduction in the buffer has to be eliminated since it not only reduces total mobility but also increases off-leakage currents. In addition, since thick SiGe buffer layers are problematic in terms of heat dissipation, removal of the SiGe buffers via wafer transfer onto the insulator, that is, fabrication of strained Ge-on-Insulator (sGOI), is ultimately considered to be the best solution. Moreover, it is expected that the uniaxial strain is also effective for Ge similarly as Si [6,7]. A GeSn embedded source/drain stressor is one of attractive methods to induce uniaxial stress and very large strain has been reported [8,9]. On the other hand, a tensile-strained Ge grown on a Si is also attractive structure for photonic device applications since the energy difference between direct and indirect band gaps can be reduced by the strain, and direct-gap light emission efficiency can be highly enhanced. An optical gain [10,11] and optically and electrically pumped lasers [12–14] have been demonstrated although very high thresholds were needed. To lower the threshold, fabrication of GOI is considered to be effective because defects originated from epitaxy of Ge on Si can be completely eliminated by wafer transfer.As the tensile strain is induced by utilizing thermal expansion mismatch between Si and Ge, the amount of strain is limited up to ~0.2%, which is not sufficient for direct-gap transition. To further increase amounts of the tensile strain, fabrication of micro-structures, such as micro-disks and micro-bridges, has been developed and remarkably large strains have been achieved [15,16]. In conclusion, straining of group IV materials offers highly performance-enhanced electronic and photonic devices on the Si platform, and further performance improvements can be made by reducing and eliminating crystal defects and increasing amounts of strain. [1] R. Pillarisetty et al., IEDM Tech. Dig., 6.7.1 (2010). [2] P. Hashemi et al., IEEE Electron Device Lett. 33, 943 (2012). [3] K. Sawano et al., Appl. Phys. Lett. 95, 122109 (2009). [4] M. Myronov et al., Jpn. J. Appl. Phys. 53, 04EH02 (2014). O.A. Mironov et al. Thin Solid Films 557, 329 (2014). [5] T. Tanaka et al., Appl. Phys. Lett. 100, 222102 (2012). [6] Y. Sun et al., J. Appl. Phys. 101, 104503 (2007). [7] T. Krishnamohan et al., IEDM Tech. Dig., 899 (2008). [8] B. Vincent et al., Microelectron. Eng. 88, 342 (2011). [9] S. Ike et al., Appl. Phys. Lett. 106, 182104 (2015). [10] J. Liu et al., Opt. Lett. 34, 1738 (2009). [11] X. Xu et al., Appl. Phys. Express 8, 092101 (2015). [12] J. Liu et al., Opt. Lett. 35, 679 (2010). [13] R. E. Camacho-Aguilera et al., Opt. Express 20, 11316 (2012). [14] R. Koerner et al., Opt. Express 23, 14822 (2015). [15] J. R. Jain et al., Nature Photon. 6, 398 (2012). [16] J. Petykiewicz et al., Nano Lett., 16, 2168 (2016).
Read full abstract
Apr 26, 2024
Victor Turpaud + 13
Open Access
