38
S. J. Mukhopadhyay et al.
device is absent. Generally, the small-signal assumptions are well valid for noise
analysis of IMPATT diodes [2].
In the earlier chapter, the potentiality of 3C-SiC-based IMPATT sources was
explored by presenting the large-signal characteristics of those at mm-wave and THz
frequencies. However, the descriptions as well as the exploration of high-frequency
performance of ATT sources stay inadequate without studying the avalanche noise
properties of those around those frequency regimes. Therefore, this chapter is dedicated to the detailed presentation of the results of a co-relative research on noise
performance of mm-wave and THz Si, 3C-SiC, and diamond-based IMPATT sources.
Again, this comparative study will be very much useful for the researchers in order
to understand the potency of 3C-SiC IMPATT sources as a suitable solid-state source
at the specified frequency bands.
2 A Brief History
At the initial stage, in the year of 1965, A. S. Tager developed a model to estimate
the noise in IMPATT sources as a function of frequency under small-signal condition [1]. Later, Hines [3] developed the theory for small-signal avalanche noise in
IMPATT oscillators. He considered the interaction of the device with external circuits
in his proposed model. He obtained the noise current spectrum of an amplifier under
small-signal condition and applied the same to find out the noise figure of read-type
avalanche diode amplifier (see Fig. 1). In his analysis, he assumed equal ionization rates for both types of charge carriers. Hines showed that the avalanche noise
increases at high signal level due to increasingly nonlinear behavior of avalanche current caused by large amount of fluctuation in the initial current which is subsequently
multiplied by avalanche process. Gummel and Blue [4] proposed a generalized smallsignal model for investigating the noise characteristics of ATT devices having any
arbitrary doping profile. They considered realistic field-dependent carrier ionization
rates in their model but assumed that the drift velocity of charge carriers are saturated and field-independent even at the edges of the space charge layer. However, the
effect of field-dependent velocity at the space charge layer edges was considered in
estimating the noise measure as a function of parasitic resistance. Figure 2 illustrates
some important findings of Gummel and Blue [4].
Earlier the noise generation under large-signal condition was analyzed by Inkson [5]. He pointed out that the noise power increases under large-signal condition.
However, the above-mentioned approximation is valid under extreme large-signal
condition and thus the theory failed to provide reliable information regarding the
enhancement in noise power with the increase of signal level. Haus et al. [6] reported
that the optimum noise measure is obtained if the field derivative of ionization coefficient is constant. They showed that the device would exhibit minimum noise for a
transit angle of 2π . Kuvas [7] in 1972 reported that the homogeneous noise spectrum of the device can be extracted by separating the large signal and noise terms. He
suggested that a complete design optimization for the power–noise characteristics
S. J. Mukhopadhyay et al.
device is absent. Generally, the small-signal assumptions are well valid for noise
analysis of IMPATT diodes [2].
In the earlier chapter, the potentiality of 3C-SiC-based IMPATT sources was
explored by presenting the large-signal characteristics of those at mm-wave and THz
frequencies. However, the descriptions as well as the exploration of high-frequency
performance of ATT sources stay inadequate without studying the avalanche noise
properties of those around those frequency regimes. Therefore, this chapter is dedicated to the detailed presentation of the results of a co-relative research on noise
performance of mm-wave and THz Si, 3C-SiC, and diamond-based IMPATT sources.
Again, this comparative study will be very much useful for the researchers in order
to understand the potency of 3C-SiC IMPATT sources as a suitable solid-state source
at the specified frequency bands.
2 A Brief History
At the initial stage, in the year of 1965, A. S. Tager developed a model to estimate
the noise in IMPATT sources as a function of frequency under small-signal condition [1]. Later, Hines [3] developed the theory for small-signal avalanche noise in
IMPATT oscillators. He considered the interaction of the device with external circuits
in his proposed model. He obtained the noise current spectrum of an amplifier under
small-signal condition and applied the same to find out the noise figure of read-type
avalanche diode amplifier (see Fig. 1). In his analysis, he assumed equal ionization rates for both types of charge carriers. Hines showed that the avalanche noise
increases at high signal level due to increasingly nonlinear behavior of avalanche current caused by large amount of fluctuation in the initial current which is subsequently
multiplied by avalanche process. Gummel and Blue [4] proposed a generalized smallsignal model for investigating the noise characteristics of ATT devices having any
arbitrary doping profile. They considered realistic field-dependent carrier ionization
rates in their model but assumed that the drift velocity of charge carriers are saturated and field-independent even at the edges of the space charge layer. However, the
effect of field-dependent velocity at the space charge layer edges was considered in
estimating the noise measure as a function of parasitic resistance. Figure 2 illustrates
some important findings of Gummel and Blue [4].
Earlier the noise generation under large-signal condition was analyzed by Inkson [5]. He pointed out that the noise power increases under large-signal condition.
However, the above-mentioned approximation is valid under extreme large-signal
condition and thus the theory failed to provide reliable information regarding the
enhancement in noise power with the increase of signal level. Haus et al. [6] reported
that the optimum noise measure is obtained if the field derivative of ionization coefficient is constant. They showed that the device would exhibit minimum noise for a
transit angle of 2π . Kuvas [7] in 1972 reported that the homogeneous noise spectrum of the device can be extracted by separating the large signal and noise terms. He
suggested that a complete design optimization for the power–noise characteristics