Historical Background

1.1 Historical Background

In 1926 Julius Edgar Lilienfeld first described a device similar to what we now call a field effect transistor in the US-patent named “Method and apparatus for controlling electric current” [1]. However, it took about thirty more years until the first transistor was actually built; ironically, it was a bipolar junction transistor.

While the first integrated circuits only contained a few transistors, the demand for more complex circuits, and therefore a higher number of transistors, increased steadily. To accomplish the growing number, scaling became the most important topic. In 1974 Dennard et al. presented a paper where they stipulated that scaling all device dimensions and voltages by a factor of $s$ at the same time requires to scale all doping concentrations by a factor of $1 ∕s$ to maintain the same electric fields inside the device [2].

In the beginning of the 1980s, the complementary metal-oxide-semiconductor (CMOS) technology was introduced to maintain the development of the already “very large-scaled integration” (VLSI) of transistors. Besides decreasing device size, cleaner and larger fabrication plants for semiconductor manufacture (fabs) were required to increase the yield.

As the demand for faster central processing units (CPUs), larger memory cells, and other integrated circuits increased further, reliability issues concerning the product specifications became more important. In order to reduce the rate of failure of devices further, the semiconductor industry had to improve the involved production processes which often included the replacement of materials responsible for the malfunction of devices. Unfortunately the novel materials in turn caused reliability challenges. One of these reliability phenomena was originally discovered in 1966, when Miura et al. linked the generation of charge due to an electrochemical reaction to the presence of a strong electric field at the Si– $SiO2$ interface [3], a phenomenon called bias temperature instability (BTI). Despite this, BTI was nearly forgotten for some decades due to its only minor relevance for the early semiconductor industry.

Already right from the start, it was discovered that when interfacing different materials with different lattice parameters, like Si and $SiO2$ , defects maybe generated at the interface [3, 4]. This is due to the non-abrupt transition, which spans over one to two atomic layers and results in an “interface region”, where a lot of dangling bonds act as traps for electrons and holes. By annealing of the structure with hydrogen (H-passivation) the density of these dangling bonds at the interface $Dit$ can be reduced from $1012cm −2eV− 1$ to around $1010cm −2eV −1$ [5], which is a huge improvement. When placing a metal gate electrode on top of the $SiO 2$ -oxide, a metal-oxide-semiconductor (MOS) structure is formed, whose operation is explained in Appendix B. Such a MOS-structure is the central part in the metal-oxide-semiconductor field effect transistor (MOSFET), which is exemplarily shown in Fig. 1.1.

As already mentioned, newer materials entered the MOSFET-structure and especially the gate oxide. With the introduction of nitrogen into the oxide the permittivity was increased and the boron diffusion from the gate material into the bulk semiconductor was significantly reduced. At the same time BTI increased in importance¹ .

Figure 1.1: An oxide material placed between a gate contact out of metal (aluminum) or highly doped polysilicon and a semiconductor substrate is called metal-oxide-semiconductor (MOS) structure. With ever smaller MOS-structures’ as part of the metal-oxide-semiconductor field effect transistor (MOSFET) reliability issues become more important. Note that the doping of the substrate is illustrated on a logarithmic scale with emphasis on the lightly doped drain (LDD) regions between gate and the source, respectively drain regions.

[next] [front] [up]