Ultimate Limits of Integrated Electronics
Ultimate Limits of Integrated Electronics
Ultimate Limits of Integrated Electronics
Josquin S. Corrales
Introduction to the Physics of Semiconductors Devices (X32)
Professor Vincent Chang
UC Berkeley Extension
introduction
To begin my discussion of the foundation in current techology of semiconductors, I present the energy
band concept, where the division of energy levels exists and is mapped into principle regions. I follow
this by showing several expressions for the ideal diode device, and how it reveals its fundumental
device behavior for an input-output relationship. An explaination of the crystal structure for silicon is
presented, when doped with phosphorus and boron atoms. Lastly I briefly discuss the p-n junction
forces that exist in order to provide a jumping point into the Tunneling Diode discussion and further
advances of integrated electroncs.
current technology
energy band
For an intrinsic semiconductor, the schematic band diagram (Figure 1), shows an energy level representation of hole, electron states, and the separation by an energy gap region. EV is maximum of valence
band. EC is conduction band. Eg is the band gap energy. The basic operating principle is that at room
temperature, thermal agitation results in excitation of electrons from the valence to the conduction band,
allowing the device to conduct current.1 At a finite temperature, the number of electrons in the conduction band is equal to the number of holes in the valence band, and is equal to the intrinsic carrier concentration (ICC).2 In a pure semiconductor, the intrinsic carrier concentration equals the electron or hole
concentration.3
corralesj_x32_paper_web2.nb
Figure 3: Relationship between material energy gap and intrinsic carrier concentration.6
This expression presented in Fig 2 and 3 is captured in the ideal diode equation, and can be plotted as
in Fig 4. Here, the operating principle of p-n junctions is that they rectify, allowing current to flow easily
in only one direction. When forward bias condition occurs in the junction, the current increases rapidly
as the voltage increases.7
Printed by Wolfram Mathematica Student Edition
corralesj_x32_paper_web2.nb
This expression presented in Fig 2 and 3 is captured in the ideal diode equation, and can be plotted as
in Fig 4. Here, the operating principle of p-n junctions is that they rectify, allowing current to flow easily
in only one direction. When forward bias condition occurs in the junction, the current increases rapidly
as the voltage increases.7
150
100
50
0.1
0.2
0.5
Figure 4: The ideal diode activation graph, where x-axis represents voltage, and y-axis is current.
For the material dependence (Table 1), different materials have different band gap, and similarily different materials have different ICC.
corralesj_x32_paper_web2.nb
corralesj_x32_paper_web2.nb
Diffusion net transport of the hole group will be from high concentration of the hole region to the low
concentration on the electron side, and similarily for the high electron concentration in the electron
side.18 The built-in electric field lines are created in the depletion layer.19
corralesj_x32_paper_web2.nb
tunnel diode
A tunnel diode (TD) is a type of semiconductor utilizing a quantum mechanical effect of tunneling to
achieve very fast operation, well into the microwave frequency region (between 300 MHz and 300
GHz).23 This effect is achieved when a particle tunnels through the built-in voltage barrier, that in classical models would not be possible. In 1958, Leo Esaki discovered the TD, but today, more traditional
three terminal devices offer better level of performance,24 as well as a new class of devices known as
super-steep slope tunneling field effect transistor (TFET) Fig 12.25
corralesj_x32_paper_web2.nb
Figure 12: Advancements of state of the art of electroncs by an area of active research TFET.26
TD types of diodes are heavily doped p-n junctions, with depletion area width 10 nm wide.27,28 When a
small forward bias is applied, a forward bias conduction current will flow due to electrons in the conduction band of the n-region tunnelling to the empty states of the valence band in the p-region Fig 13.
Figure 13: Direct tunneling current starts growing with forward bias of a TD.28
As voltage increases, the energy of the majority carrier becomes equal producing a maximum tunneling
current.29 As forward bias continues increasing, the number of electrons in the n-side directly opposite
to the empty states in the valence band decreases causing a new decrease in tunneling curent. The TD
device that exhibits negative resistance, meaning when the voltage is increased the current through it
decreases.30 As more forward voltage is applied, the tunnel current drops to zero, but regular diode
current increases due to lower potential barrier. Following increased forward voltages, the TD I-V
characteristic is similar to that of a regular p-n diode Fig 14.31
corralesj_x32_paper_web2.nb
wkb approximation
To understand how an electron particle tunnels through the barrier (E < U(x)) by quantum tunneling, we
can apply the Wentzel-Kramers-Brillouin (WKB) approximation, which is a mathematical technique to
solve one-dimensional (1D) differential equations, and plane-wave ray equations.32 It is a semiclassical
calculation in quantum mechanics, where it is relatively computationally inexpensive and accounts for
quantum effects. Here, the wave function is assumed an exponential with amplitude and phase that
slowly varies compared to the de Broglie wavelength and then is semiclassically expanded.33
The WKB approximation shows that with a constant potential, the wavefunction solutions of the
Schrodinger equation are of the form of simple plane waves Fig 15. Where is called a modified
Plancks constant.34
corralesj_x32_paper_web2.nb
10
corralesj_x32_paper_web2.nb
SL is using the resonance tunnel diode phenomena lab at room temperature with the aim of achieving a
magnetic resistance ratio of 100%. If this research progresses, it will also enable the development of
corralesj_x32_paper_web2.nb
11
devices for reading information from magnetic memory.
The Lab is researching materials for solar cells. Currently, about 95% of the electricity produced worldwide from solar cells comes from silicon. Because silicon is an abundant element in the Earths crust,
and silicon technology is highly advanced, silicon is a very superior material. But during the last few
years, the amount of silicon material science research interest has increased enormously. Researchers
worldwide are studying how to make efficient solar cells using thin films. SL is working on a material
called barium silicide. Its light-absorbing performance, as expressed by the light absorption coefficient,
is hunderds of times that of silicon. With this material, solar cells can then be made hundreds to times
thinner and still maintain the same current producing performance. The SL group has confirmed that a
current actually flows under light sources, and they have reached the stage where it can function as a
solar cell (Fig 23). From this achievement, they aim to improve the efficiency of the solar cells.
conclusion
The tunneling diode, although discovered in the 1950s, represents an active area of research that has
the potential to advance the state of electronics. New mathematical models are being developed based
on the Schodingers wave equations and quantum mechanical effects. As our knowledge of these
subjects increase, there could in fact be no limit to the application and expansion of the field of
advanced integrated electronics.
references
[1] Semiconductor Devices Physics and Technology, 3rd Ed, S.M. Sze, p 29.
[2] Ibid, p 31.
[3] UC Berkeley X32: Introduction to Semiconductor Devices, Prof Vincent Chang, Lecture 2, Slide 71.
[4] Ibid, Lecture 2, Slide 62.
[5] Ibid, Lecture 2, Slide 67.
[6] Ibid, Lecture 2, Slide 69.
[7] Semiconductor Devices Physics and Technology, 3rd Ed, S.M. Sze, p 83.
[8] UC Berkeley X32: Introduction to Semiconductor Devices, Prof Vincent Chang, Lecture 2, Slide 70.
[9] Ibid, Lecture 3, Slide 20.
[10] Ibid, Lecture 3, Slide 38.
[11] Ibid, Lecture 3, Slide 51.
12
corralesj_x32_paper_web2.nb