14. SOME ASPECTS OF NANOELECTRONICS DEVELOPMENT IN RUSSIA


I.A. Obukhov
"Delta" Research Institute
2 Schelkovskoe Sh.
Moscow105122, Russia
e-mail: anv@srl.phys.msu.su or obukhov@interf.mx.orc.ru

This report reviews some Russian research in the field of designing and creating nanodevices. It analyzes the results of more than 10 years' efforts to create resonant-tunneling structures and microwave devices on this basis. It also examines the modern state and future directions for using quantum wires (QW) as basic elements for designing nanodevices.

Introduction

The development of nanoelectronics in Russia began in the middle of the 1980s with work on resonant-tunneling diodes (RTD). Among those conducting research in this area are Yu.S. Tikhodeev, O.T. Gavrilov, and I.I. Kvjatkevitch at the "Pulsar" Research Institute in Moscow; Yu.A. Ivanov at the experimental scientific-industrial corporation "Special Electronic Systems" in Moscow; and I.A. Obukhov at the "Delta" Research Institute in Moscow. Somewhat later, P.N. Luskinovich, I.A. Ryjikov, and V.D. Frolov at the "Delta" Research Institute began to examine the technological applications of scanning tunneling microscopes (STM) in order to serve the needs of electronics.

The work on RTDs has given Russian researchers significant practical experience in creating quantum devices. Moreover, on the basis of the experimental data, some theoretical ideas were developed. This experience led to understanding transport physics in structures with quantum properties and to describing the characteristics of devices adequate for experiment. The experience has also appeared useful for designing next generation quantum devices, devices based on QWs.

For a long time, some authors hoped to apply STM methods to electronic technology. A set of projects on the formation of quantum electronic circuits using the STM already existed (P.N. Luskinovich et al. 1995). The projects were assumed to lead to the actual creation of such circuits by the mid-1990s. However, the goal has not been reached. It is possible to say only that STM appears to be a useful tool for nanolithography and for creating QWs and devices based on QW.

The Resonant-Tunneling Diode: Its Current State and its Long-term Perspective

The functioning of the RTD is based on the well-known effect of resonant tunneling. If for electrons, the potential shown in Fig. 14.1 is created, then increased transmission of electrons through the barriers can occur for particular alignments (or resonances) of the energy levels (Fig. 14.2). These resonances are in turn controlled by the voltage bias applied across the entire structure: as the energy levels come into alignment, or resonance, peaks in the transmission current occur (see Vp, in Fig. 14.3). As one continues to increase the voltage, the energy levels will no longer be resonant and the current will diminish, until the device reaches a local minimum (valley) in current, at a voltage Vv. Therefore, it is possible to produce current-voltage characteristics with multiple values of voltage for a given current value. Such RTDs can be formed by using structures such as those shown in figures 14.4 and 14.5, grown through the use of molecular beam epitaxy (MBE). Both theoretical estimations and experimental realizations have suggested that the RTD has applications as a high frequency device (Solner 1983; Brown 1988; Tikhodeev 1973). Circuits utilizing RTD devices have demonstrated operation at frequencies greater than 100 GHz.


Fig. 14.1. Potential RTS.


Fig. 14.2. Electron transmission coefficient.


Fig. 14.3. I-V characteristics of RTD.


Fig. 14.4. RTS on n+ -substrate.

In Russia the work on RTDs began in 1987. Dr. Yu.S. Tikhodeev was its initiator and scientific chief. The first sample devices were obtained in 1990. The main research was carried out at the "Pulsar" Research Institute (Moscow). Interesting experimental and theoretical results were obtained by the scientific groups of Prof. A.S. Tager at the State Research and Production Corporation "Istok," Frjazino (Tager 1987; Tager 1988; Kal'fa et al. 1993); Prof. V.G. Mokerov at the Institute of Radio Engineering and Electronics, Russian Academy of Sciences, Moscow; and Prof. Yu.V. Kopaev at the Lebedev Physical Institute, Russian Academy of Sciences, Moscow.

Figures 14.6 - 14.8 present the experimental I-V characteristics of RTDs developed in the 'Pulsar' Research Institute. One can clearly see the variation in values of peak voltage, Vp, with a range of ~1.4-7 volts. This variation, as well as the hysteresis shown in Fig. 14.7, may arise from influences of charge transport from the substrate. Other than the expected effect of the substrate, in determining the series resistance to the RTD, modeling and experiments have shown other influences of the substrate on the performance of the RTDs (Gavrilov et al. 1993). For example, Fig. 14.9 shows the distribution of the current density in a direction perpendicular to the RTS-n-substrate interface. The non-uniform distribution of the current density is a result of the non-monotonic I-V characteristics of the RTD, and in turn, influences the I-V performance of the RTD itself. Experimental confirmation of the theoretical predictions was made by changing the geometry of the RTD contacts, and also by changing the doping of the substrate underlying the RTD structures. Another important influence of the substrate on RTD performance may be in its action as a 'resonator' for fluctuations. The resonance frequence will vary, depending on the device geometry and its conductivity (Gavrilov et al. 1996). Experimental demonstration of this effect was given by the generation of 'white noise' in the frequency range 10 MHz - 25 GHz. The white noise was observed only when the voltage bias was placed in the RTD's region of negative differential conductivity. Total radiated power was as high as 20 W, and the spectrum of noise power calculated was equal to 2.91 x 105 kBT.

This conclusion was confirmed experimentally by using specially created test structures with various top surface contacts (see figures 14.4 and 14.5) to identical RTS (see figures 14.6 and 14.7). Reduction of the contact area results in the improvement of the RTD characteristics: shift of Vp, an increase in current density, and an increase in the threshold frequency of generation.


Fig. 14.5. RTS on semi-insulating substrate.


Fig. 14.6. I-V characteristics of RTD on n+-substrate with se = 200 and diameter of contact to RTS: 1— 100 µm; 2— 80 µm; 3— 40 µm.


Fig. 14.7. I-V characteristics of RTD on n+-substrate with se = 15 and diameter of contact to RTS: 1— 100 µm; 2— 80 µm; 3— 40 µm.


Fig. 14.8. I-V characteristics of RTD on semi-insulating substrate (diameter of contact to RTS is 80 µm).


Fig. 14.9. Distribution of normal current density to boundary between RTS and n- -substrate.


Fig. 14.10. Potential relief for electron in junction between 2 InSb QW with width 70 nm and 10 nm.

Prof. Yu.A. Ivanov and his collaborators have investigated the circuitry applications of RTD at the experimental scientific-industrial corporation "Special Electronic Systems" (Georgievskyi et al. 1996). Different types of generators and mixers were created and researched. These devices demonstrate satisfactory characteristics, but the frequency range is only between 1 GHz and 50 GHz. The limited range is caused by some technological problems and by the limits of the measuring equipment. The noise characteristics of the devices and the RTD reaction to external radiation have been investigated as well.

To conclude, it is important to note that work on the RTD problem has already passed from the research stage to the search for technical applications. The achievements of Russian researchers in the discussed area are impossible to characterize as remarkable, but some interesting original results have been obtained. This research may be the basis for future development.

Devices Based on Quantum Wires: Prospects and Problems

The next interesting direction of Russian research in the field of nanometer scale science is the design of devices based on QW studied by I. A. Ryjikov and I. A. Obukhov at the "Delta" Research Institute in Moscow. Two circumstances have generated interest in these investigations:

  1. the opportunity to create superhigh-frequency (>100 GHz) electronic devices with planar topology
  2. the prospect of increasing the elements' integration level in circuits up to sizes of the order 1010 pieces on a square centimeter

A quantum wire (QW) is a structure where electrons are confined in two directions, with transport allowed in a third dimension. At a given temperature, T, the quantized energy levels will be apparent if the difference between those energy levels is greater than kBT. A simple calculation shows that this condition will be satisfied if the critical dimension of the structure, L, is less than L0, where

L0 = (3h2/8m* kBT)1/2

For electrons in GaAs at room temperature (T = 300 K), L0 = 25.3 nm; in InSb, L0 = 57.9 nm (Table 14.1).


Table 14.1
Physical parameters important for quantum wires of various materials

T=300 K

L0(e)/L0(h) nm

Lrel(e)/ Lrel(h) nm

ni cm -3

m*e / m0

m*h/ m0

m e cm2/V s

m h cm2/V s

t e s

t h s

Organic material

5.0-10.0/ ...

1.5-2.0/ 1.5-2.0

1016 - 1022

1

1

40 - 100

40 - 100

2.5 x

10-14

2.5 x

10-14

Si

15.1/ 16.5

10.0/ 6.3

1.6 x 1010

0.19

0.16

1500

600

1.58 x10-13

5.3 x 10-14

GaAs

25.3/ 19.1

23.8/ 5.2

1.1 x 107

0.068

0.12

8500

400

3.21 x10-13

2.7 x 10-14

InAs

46.7/ 10.3

47.3/ 5.6

 

0.02

0.41

3.3 x 104

460

3.7 x 10-13

1.06 x10-13

InSb

57.9/ 8.5

72.7/ 7.1

1.7 x 1016

0.013

0.6

7.8 x 104

750

5.68 x10-13

2.52 x10-13

One candidate device is the quantum field effect transistor (QFET). The topology and calculated static characteristics of this device are demonstrated in figures 14.11 and 14.12. Note that although the specific transconductance of the QFET is very high (105 A/V-cm2), the total transconductance is only 10-7 A/ V-cm2. Such values are typical for devices based on QW and are a consequence of their small cross-sectional areas (~ 10-12 cm2).


Fig. 14.11. Quantum field effect transistor.


Fig. 14.12. I-V characteristics of GaAs QFET for Ly = 30 nm and 1— Vg = 0V, 2— Vg = -0.05 V, 3— Vg = -0.1 V, and 4— Vg = -0.3 V.

One can modulate the potential landscape seen by the electrons through the lateral modulation of the size of the quantum wire. Therefore, when the quantum wire diameter is scaled from 70 nm to 10 nm, the potential landscape for electrons in InSb may appear as shown in Fig. 14.10. The manner in which the transition is made from the wider to the more narrow quantum wire has a critical impact on the transmission of electrons from one region to the next. An important length scale is then the relaxation length:

Lrel = (ht /m*)1/2

This principle has been further employed in the proposal for a relaxation quantum transistor (RQT, see Fig. 14.15) (Obukhov 1993, Obukhov 1996). The device operation is similar to that of a conventional bipolar transistor, where an external signal (through the base) is used to modulate charged carriers flowing between the collector and emitter. Calculated current gain of the RQT is ~200-1000. The specific capacitances of these QW devices are of the order of 0.1-0.2 µF/cm2. We can use these values to extrapolate the maximum operating frequencies of the devices: 1011 Hz for the QFET and 1013 Hz for the RQT.


Fig. 14.13. Relaxation quantum diode.


Fig. 14.14. I-V characteristics of InSb RQD with L1 = 70 nm and Lt = 10 nm.

Table 14.1 lists some of the important physical parameters that might influence the performance of a QW device: the critical linear scale, L0 (for electrons and holes), the relaxation length Lrel (for electrons and holes), the intrinsic carrier concentration, expected effective mass and mobilities of electrons and holes, and lifetimes for momentum relaxation. All values are expected to pertain to operation at room temperature.

There are two principal obstacles to the creation of the QW devices described. First, there are the technological difficulties in fabricating devices at these dimensions. Although individual devices may be realized, there are numerous issues related to scaling up such approaches to the level of manufacturability. However, progress in the field indicates that there are ways of working to solve these technological issues. Secondly, however, there are problems relating to the basic physics limitations of devices at these dimensions.

The reduction of the number of possible electron states in QW helps to reduce their scattering and, hence, to increase momentum relaxation time. This is the positive effect that results in increasing the electron mobility in QW in comparison with volumetric material. However, the electron concentration in quantum-dimensional structures decreases and, as a result, the conductivity increases insignificantly. These increases do not appear to be enough to compensate for the negative influence of the small areas of QW cross-sections. Therefore, QW and devices based on QW are structures with high resistance, and, as a consequence, with high levels of thermal noise. So for QFET, the level of thermal noise reaches 100 to 1000 nV/(Hz)1/2. For RQD and RQT, this value is less: 40 - 100 nV/(Hz)1/2. But it is too large a noise in comparison to the best microelectronic devices, for which typical values are 0.3 - 1 nV/(Hz)1/2.


Fig. 14.15. Relaxation quantum transistor.


Fig. 14.16. I-V characteristics of InSb RQT in circuit with common base: 1—Veb = 0.05 V; 2— -Veb = -0.02 V 3— Veb = -0.05 V; and 4— Veb = -0.07 V.

It might be possible to solve the problem of large resistance and significant thermal noise by using extremely pure materials with very small numbers of scattering centers. One might use an approach such as employed in modulation-doped heterostructures, in which the impurity is spatially separated from the charged carriers. Nonetheless, in nanosized structures, specific quantum noise may be observed. As recent research has shown, quantum fluctuations of the electron electrochemical potential lead to the appearance of voltage noise in which spectral density is the proportional factor:

(Lrel/L)2

where L is the typical length of the device's active area. As Lrel ~ 10-6 cm, this noise is practically invisible in devices with macroscopic sizes. But for the elements considered here, L may be of the order Lrel, and even much less than Lrel. In such a situation, quantum noise may prevail. To suppress it would hardly be possible. Therefore it is necessary to realize that in nanoelectronic circuitry, the noise level will be high.

Conclusion

To conclude, the two directions for nanoelectronics development in Russia are connected by the names of the scientists and by the general methodology and experimental bases of the research.

The work on RTD is now at a stage of practical readiness for experimental production. The work on devices based on QW is in its initial stage. Scientists hope that they understand what research is necessary, but experimental research has only just begun. Which results will be obtained and which will be theoretical at the end remains unclear.


Fig. 14.17. I-V characteristics of InSb RQT in circuit with common emitter: 1— Vbe = 0.00 V; 2— Vbe = 0.01 V; 3— Vbe = 0.02 V; and 4— Vbe = 0.03 V.

References

Brown, E.R., W.D. Goodhue, and T.C.L.G. Solner. 1988. J. Appl. Phys. 64:1519.

Gavrilov, O.T., I.A. Obukhov, S.A. Dianov. 1993. Proc. of the second international conference on nanometer scale science and technology (NANO-11). Moscow, p. 860.

Gavrilov, O.T., I.I. Kvjatkevitch, I.A. Obukhov. 1996a. Sixth international Crimean conference on microwave & telecommunication technology. Conference proceedings (in Russian). Sevastopol: Weber Co. p. 308.

Gavrilov, O.T., I.I. Kvatkevitch, I.A. Obukhov, Yu.A. Matveev. 1996b. Tech. Phys. Lett. 22:311.

Gavrilov, O.T., I.I. Kvjatkevich, and I.A. Obukhov. 1997. Seventh international Crimean conference "Microwave & Telecommunication Technology." Conference proceedings (in Russian). Sevastopol: Weber Co. p. 401.

Georgievskyi, A.M., D.V. Gromov, et al. 1996. Mikroelektronika (in Russian), 25:249.

Kal'fa, A.A., A.R. Kryukov, and A.S. Tager. 1993. Symposium on Nanostructures: physics & technology. Abstracts. St. Petersburg, p.65.

Luskinovich, P.N., V.D. Frolov, E.F. Sheka, et al. 1995. JETP Lett. (in Russian) 62:881.

Obukhov, I.A. and S.A. Dianov. 1993a. Proceedings of the second international conference on nanometer scale science and technology (NANO-II). Moscow. p. 775.

Obukhov, I.A. 1993b. Tech. Phys. Lett. (in Russian) 19:537.

_____. 1996. Sixth international Crimean conference on microwave & telecommunication technology. Conference proceedings (in Russian). Sevastopol: Weber Co. p. 55.

_____. 1997. Seventh international Crimean conference "Microwave & Telecommunication Technology." Conference proceedings (in Russian). Sevastopol: Weber Co., p. 383-386.

Solner, T.C.L.G. et al. 1983. Appl. Phys. Lett. 43:588.

Tager, A.S. 1987. Elektronnay Tekhnika. Ser. 1. Elek-tronika SVCH (in Russian) 9:19.

_____. 1988. Elektronnay Tekhnika. Ser. 1. Elek-tronika SVCH (in Russian) 2:17.

Tikhodeev, Yu.S. 1973. Elektronnay Tekhnika. Ser. 2. Poluprovodnikovie Pribori (in Russian) 1:3.


Published: August 1997; WTEC Hyper-Librarian