HAL Id: jpa-00226711
https://hal.archives-ouvertes.fr/jpa-00226711
Submitted on 1 Jan 1987
HAL
is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.
L’archive ouverte pluridisciplinaire
HAL, estdestinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.
RESONANT TUNNELING DEVICES
R. Nahory, N. Tabatabaie
To cite this version:
R. Nahory, N. Tabatabaie. RESONANT TUNNELING DEVICES. Journal de Physique Colloques,
1987, 48 (C5), pp.C5-585-C5-588. �10.1051/jphyscol:19875127�. �jpa-00226711�
Colloque C5, suppl&ment au noll, Tome 48, novembre 1987
RESONANT TUNNELING DEVICES
R.E. NAHORY and N. TABATABAIE
Bell Communications Research, Red Bank, NJ 07701-7020, U. S. A.
The physics of resonant tunneling is as yet a research topic and has been discussed at some length elsewhere in this conference. The subject of this paper, devices using the phenomenon of resonant tunneling, is one which is in its infancy. Nevertheless, it is of great interest to examine the present state and to explore some of the directions which research is taking in the study of resonant tunneling devices. Among other things, these devices offer the possibilities of very high speed and very small size, and thus advantages for very large scale integration.
In order to give some framework to this discussion we consider resonant tunneling devices in the context of some of the current trends in materials and devices. In the case of materials trends, the concept of superlattices as a new kind of tunable material has been a subject of research since the early 1970's[l]. Recent advances in materials growth techniques have rapidly advanced these ideas, as higher and higher qualities of crystals and interfaces have been obtained. In addition, during the last decade or so, the concept of heterostructure layering using bulk films (0.1
-
1 pm layer thicknesses) has advanced to a high degree. An excellent example of this is the progression of semiconductor laser structures from simple stripe geometry double heterostructures to the present sophisticated multilayer, single mode laser devices[2]. In this context, resonant tunneling devices, although first introduced in the early 1970's,[3] are still rather simple and apparently in yet an early stage. They have yet to take advantage of the extensive possibilities of layering and patterning that current materials science and fabrication technology have to offer.Trends in devices include advances toward higher and higher speed, smaller size design rules, larger and larger scales of integration for electronics, and the beginnings of integrated opto-electronic circuits. It has been pointed out that the trend to smaller and smaller devices, and the concomitant larger and larger scales of integration, eventually must end as scaling laws break down [4]. These limits could be reached before the end of this century. Accordingly, new concepts in both devices and their interconnection will be needed, for which it is urgent that research be in progress now. One of the areas where breakthroughs are being sought is in that of resonant tunneling devices, which offer hopes for very fast devices, very small devices, possibly new concepts in device interconnections, and new ways of thinking. In this context, we will briefly discuss in this paper some recent results we have obtained on triple banier, multiply resonant devices; some current published results on application of resonant tunneling to transistors; and some recent published results On the use of resonant tunneling devices themselves for multistate memories and logic functions.
Most of the experimental work on resonant tunneling devices has so far been concentrated on double bamer (single quantum well) structures. In our laboratory, we have fabricated and studied both double and triple banier GaAs/AIAs resonant tunneling structures grown by molecular beam epitaxy [5]. The triple barrier structures, shown schematically below in the figure, are of particular interest since under proper bias conditions the ground state of the pair of quantum wells becomes degenerate. The structures were made to be asymmetrical, which
Article published online by EDP Sciences and available at http://dx.doi.org/10.1051/jphyscol:19875127
C5-586 JOURNAL DE PHYSIQUE
results in a strongly asymmetric tunneling I-V characteristic with respect to bias polarity. With this design we have increased the total tunneling current at room temperature by a factor of four over a similar double bamer structure with a 70 angstrom quantum well. The triple bamer structures consisted of two GaAs layers 50 and 70 Angstroms thick and three AlAs bamer layers each 20 Angstroms thick. The I-V characteristics measured for 90 pm diameter mesa devices at both room temperature and low temperature (77K) are also shown in the figure below. In each case we observe multiple negative differential resistance regions, which are all especially pronounced at 77K. The interpretation is as follows: The first negative differential resistance region, between 0.3 and 0.5 volts bias, is attributed to tunneling through the lowest quantum states in the two GaAs wells, which are in this case aligned in energy.
This structure is in many ways similar to that observed for double bamer structures [6]. The sharp negative differential resistance observed near 0.7 volts bias is attributed to injection into the n = 2 state in the 70 Angstrom quantum well, followed by a downward transition in this well from the n = 2 to the n = 1 energy state, and then further tunneling through the n = 1 state of the 50 Angstrom GaAs quantum well. The n = 2 energy state in the thinner well is too high to be important in this process. This interpretation is supported by other measurements in the reverse current direction, which differ from the results shown because of the asymmetry [7]. We have thus identified two type of tunneling mechanisms in this triple bamer structure, one requiring degenerate ground state energies for the two quantum wells, and the other requiring transitions between the n = 2 and n = 1 energy states in one of the wells.
Roc Liq.
V .1000/dlv < W 1.000
wn Temp.: t o p
Nit. Temp.: b o t .
One could take advantage of these effects, for example for multistate memory devices, by making use of the I-V curve shown for 77K, since one can define a load line distinguishing three separate operating points o n the positive slope regions of the curves. This will be discussed further below. In addition, it might be important that the n = 2 excited state resonance of the 70 Angstrom well does not match that of the 50 Angstrom well, resulting in a tunneling sequence which requires large energy loss (transfer to the n = 1 energy state) by the tunneling electrons. If the transition between these subbands is radiative, it is hoped that these types of structures may lead to a new class of infrared emitters. T h e possibility of such light emission has been discussed theoretically many years ago [8].
T h e speed of resonant tunneling devices has been an active area of discussion. Several reports conclude that very high speeds ought to be expected, with tunneling times in the subpicosecond regime [6,9,10]. This in conjunction with the negative differential resistance region in the I-V characteristics has enabled the operation of high frequency oscillators, so far the most successful application of resonant tunneling to devices. It has been demonstrated that resonant tunneling in AlGaAs/GaAs double bamer structures takes place at terahertz frequencies [I I], while a quantum well oscillator has recently been operated a t a frequency as high as 200 GHz [12].
containing resonant tunneling structures. A resonant tunneling MESFET has for example bcen proposed and analyzed [13]. It was found that substantial modulation of the negative differential resistance could be obtained through application of reasonable gate voltages. It is anticipated that such devices could be useful in the microwave and millimeter wave regimes.
Double barrier resonant tunneling structures have been proposed as part of the base in bipolar transistors [9], and have been inserted into the emitter of resonant tunneling hot electron transistors [14]. These latter devices have been fabricated and measured, showing a peaked collector current characteristic as a function of base-emitter voltage. This device, although not yet of high quality [15], shows promise for future high speed, high density integrated circuits. For example, an exclusive-nor logic function has been demonstrated using this one transistor and three resistors, which is to be compared to the use of 8 MESFET's for the usual exclusive-or logic circuits [14].
Without the aid of transistors, multiple resonant tunneling devices haveabeen used recently in a simple integrated device to demonstrate multistate memory functions [16]. By operating two double barrier resonant tunneling devices in parallel in a three terminal structure, a double negative resistance 1-V characteristic was demonstrated. This was further operated in a simple circuit'as a 3-state memory device by constructing a circuit with a suitable load line, which intersects the I-V at 3 different, isolated points at positive slope regions. Thus three switchable stable states were defined. As in the transistor case above, this is significant in that 2 integrated resonant tunneling devices are able to replace conventional 3-state logic circuits which usually consist of 4 transistors and 6 resistors. Since a requirement for this 3-state logic function is an I-V with two negative differential regions with nearly equal current peaks, we would expect triple barrier devices as discussed above to be potentially useful for such functions also. We might speculate at this point that indeed this use of resonant tunneling devices alone to carry out logic operations, without transistors, might be a very rich direction for future applications, in view of our discussion above about future needs for large scale integrated circuits. Finally, it is worth noting in passing that several other kinds of resonant tunneling devices have been proposed, including unipolar resonant tunneling devices [I71 and a new kind of quantum well injection transit time effect device [18]. Space does not permit discussion of these devices here however.
In conclusion, resonant tunneling devices alid their applications are in their infancy. Early work however shows glimmers of great promise for new kinds of circuits and new kinds of fast devices which can perform logic operations with fewer devices than present transistor circuits. This is just the direction desired, looking for future break-throughs, which would enable progress beyond presently seen limitations in the continuing advance to larger and larger scales of integration. At present however, resonant tunneling devices are very simple and have not yet benefited from the kinds of thinking and technologies that have so affected the evolution of semiconductor lasers.
REFERENCES
[I] L. Esaki and R. Tsu, IBM J. Res. Develop. 14 , 61 (1970); L.L. Chang, L. Esaki, W.E.
Howard and R. Ludeke, J.Vac.Sci.Techno1. 10 , I1 (1973).
[2] See for example S.M.Sze, "Physics of ~erniconduct'or Devices", second edition, Wiley (1981). page 726.
-
i3] R. Tsu and L. Esaki, Appl.Phys.Lett. 22 , 562 (1973); L.I. Chang, L. Esaki and R. Tsu, Appl.Phys.Lett. 2 4 , 593 (1974).
[4] See for example G.H. Heilmeier, International Electron Devices Meeting, Technical Digest (1984), page 2.
[5] MBE material grown by M.C. Tamargo and J.L. deMiguel.
[6] N. Tabatabaie and M.C. Tamargo, International Electron Devices Meeting, Technical Digest (1 986), page 80.
C5-588 JOURNAL DE PHYSIQUE [7] N. Tabatabaie, unpublished.
[8] R.F. Kazannov and R.A. Suns, Sov.Phys.-Semiconductors 8 , 707 (1971).
[9] F. Capasso and R. Kiehl, J.Appl.Phys. 58 , 1366 (1985).
[lo] H.C. Liu and D.D. Coon, Appl.Phys.Lett. 50 , 1246 (1987); W.R. Frensley, International Electron Devices Meeting, Technical Digest (1986) page 571; B. Ricco and M.Y. Azbel, Phys.Rev B 2 9 , 1970 (1984).
[I I] T.C.L.G. Sollner, W.D. Goodhue, P.E. Tannenwald, C.D. Parker and D.D. Peck, IEEE Tr;ns.Electr.Devices ED-30 , 1577 (1983).
[12] E.R. Brown, T.C.L.G. Sollner, W.D. Goodhue and C.D. Parker, Proceedings of the 45th Annual Device Research Conference, Santa Barbara, California, (June 1987), page VIA-2.
[13] A.R. Bonnefoi, T.C. McGill and R.D. Burnham, IEEE Electron Device Letters E D L 6 , 636 (1985).
[14] N. Yokoyama, K. Imamura, S. Muto, S. Hiyarnizu and H. Nishi, Japan. J. Appl. Phys.
24 , L853 (1985).
[I51 N. Yokoyama, Extended abstracts of the 18th International Conference on Solid State Devices and Materials, Tokyo (1986), page 347.
[16] F. Capasso, S. Sen, A.Y. Cho and D. Sivco, IEEE Electron Device Letters EDL8 , 297 (1987).
[17] B. Jogai and K.L. Wang, Appl.Phys.Lett. 46, 167 (1985); R.H. Davis and H.H. Hosack, J.Appl.F?hys. 34 , 864 (1963).
[18] V.P. Kesan, D.P. Neikirk, B.G. Streetman and P.A. Blakey, IEEE Electron Device Letters EDG8 , 129 (1987).
