Introduction to Nanoelectronics. Topic Resonant Tunnel Diode. Professor R. Krchnavek. By Kalpesh Raval

Size: px

Start display at page:

Download "Introduction to Nanoelectronics. Topic Resonant Tunnel Diode. Professor R. Krchnavek. By Kalpesh Raval"

Sarah Sutton
6 years ago
Views:

1 1 Introduction to Nanoelectronics Topic Resonant Tunnel Diode Professor R. Krchnavek By Kalpesh Raval

2 2 Over the several decades, the world of electronics has been dominated by Si (Silicon) based technologies. Silicon is the principal component of most semiconductor devices, and most importantly integrated circuits (IC s). An example of such technology is the CMOS (Complementary Metal Oxide Semiconductor) circuits. CMOS technology is used in microprocessors, microcontrollers, static RAM (Read Access Memory) and other digital logic circuits. CMOS is also used for a wide variety of analog circuits such as image sensors, data converters and highly integrated transceivers for many types of communication systems. Current modern technology is based on nanoelectronics (nanoscale) where integrated circuits are reduced in size (scaling) for high speed, less power consumption and simply using less physical space IC s for enhanced and improved application performances. This involves an integration of CMOS technology with various electronic devices. As the industry begins to address the theoretical limits of CMOS scaling, the current challenge is the integration of CMOS with new types of devices. The RTD, which utilizes electron-wave resonance in double potential barriers, has emerged as one of the most important testing grounds for modern theories of transport physics, and is central to the development of new types of semiconductor nanostructure. The tunneling devices are candidates of new functional devices applicable to new integrated circuit technology in so-called Beyond CMOS region. Among the various tunneling devices, the resonant tunneling devices (tunneling diodes) are promising because of their capability for high-speed operation, large Negataive Differentiate Resistance (NDR) characteristics at room temperature, and adaptive design of device characteristics. The most common function of a diode is to allow an electric current to pass in one direction (called the forward biased condition) and to block the current in the opposite direction (the reverse biased condition). Forward bias occurs when the P-type semiconductor material is connected to the positive terminal of a battery and the N-type semiconductor material is connected to the negative terminal. This makes the P-N junction conduct as shown below. Figure 1 If a diode is reverse biased, the voltage at the cathode is higher than that at the anode. Therefore, no current will flow until the diode breaks down. Connecting the P- type region to the negative terminal of the battery and the N-type region to the positive terminal produces the reverse-bias effect. Figure 2

3 3 Current research efforts focus on adapting the characteristics of resonant tunneling diodes (high speed & negative differential resistance (NDR)) in a way that may be closely matched to the characteristics with resonant tunnel devices. Resonant tunnel diode (RTD) is a two terminal device and widely recognized as being inherently highspeed device having a load I-V curve, characterized by a region of negative differential resistance as shown below. Figure 3 The diode is capable of generating a terahertz (x10 12 Hz) wave at room temperature. These two characteristics, high speed and NDR region make them potentially very attractive as high-speed switching devices. One of the efforts in adapting the RTD structure involves integrating pairs of RTDs with a CMOS gate to achieve bistable operation. The logic cell design consists of two RTDs connected in series with a gate voltage applied in between them to control its properties. The interactions between the nanocells of RTDs are achieved via tunneling in the quantum layers. Leo Esaki invented the tunnel diode in August 1957 while working with Tokyo Tsushin Kogyo (now known as Sony) and who in 1973 received the Nobel Prize in Physics for discovering the electron tunneling effect used in these diodes. These diodes have a heavily doped p-n junction that is approximately 10 nm (100 A 0 ) wide. The heavy doping produces an extremely narrow depletion zone and results in a broken bandgap, where conduction band electron states on the n-side are more or less aligned with valence band hole states on the p-side. This results in overlapping of the two bands. SONY manufactured tunnel diodes for the first time in 1957 followed by General Electric and other companies from about Currently, they are manufactured in low volume today. Tunnel diodes are usually made from germanium, but can also be made in gallium arsenide and silicon materials. They can be used as oscillators, amplifiers, frequency converters and detectors. RTDs are formed as a single quantum well structure surrounded by very thin layer barriers known as a double barrier structure.

The majority electrons and holes are at the same energy level in the equilibrium state.

4 4 The response function of a tunnel diode having a region of NDR due to heavy doping can be described by following bias conditions. Figure - 4 Figure - 4 shows the equilibrium energy level diagram of a tunnel diode with no bias applied. Note that the valence band of the P-material overlaps the conduction band of the N-material. The majority electrons and holes are at the same energy level in the equilibrium state. If there is any movement of current carriers across the depletion region due to thermal energy, the net current flow will be zero because equal numbers of current carriers flow in opposite directions. When a small bias voltage is applied we get, Figure - 5 Figure- 5 shows the energy diagram of a tunnel diode with a small forward bias (50 millivolts) applied. The bias causes unequal energy levels between some of the majority carriers at the energy band overlap point, but not enough of a potential difference to cause the carriers to cross the forbidden gap between the valence and

The amount of overlap between the valence band and the conduction band decreases when forward bias is applied.

5 5 conduction bands. Since the valence band of the P-material and the conduction band of the N-material still overlaps, current carriers tunnel across at the overlap and cause a substantial current flow. The amount of overlap between the valence band and the conduction band decreases when forward bias is applied. As the bias is further increased, the area of overlap becomes smaller as seen in Figure-6. When the overlap between the two energy bands becomes smaller, fewer and fewer electrons can tunnel across the junction thus the current decreases as the voltage increases, which is known as the negative resistance region of the tunnel diode. Figure 6 When the bias is increased even further, the energy bands no longer overlap (Figure 7) and the diode operates in the same manner as a normal PN junction, where the current increases because of the normal conduction and no longer by tunneling since the applied bias voltage is high enough to break the barrier. Figure 7

6 During the forward bias operation, resonant tunnel diode produces a negative differential resistance region, which is a property of electrical circuit elements composed of certain materials in

6 6 During the forward bias operation, resonant tunnel diode produces a negative differential resistance region, which is a property of electrical circuit elements composed of certain materials in which, over certain voltage ranges, current is a decreasing function of voltage. This range of voltages is known as a negative resistance region. Under normal forward bias operation, as voltage begins to increase, electrons at first tunnel through the very narrow p-n junction barrier because filled electron states in the conduction band on the n-side become misaligned with empty valence band hole states on the p-side of the p- n junction. An imbalance between the carrier charges causes current to flow. As voltage increases further these states become more misaligned and the current drops this is called negative resistance, because current decreases with increasing voltage. As voltage increases further, the diode begins to operate as a normal diode, where electrons travel by conduction across the p-n junction, and no longer by tunneling through the p-n junction barrier. Thus the response function for a tunnel diode is the negative resistance region that exhibits a high peak to valley current ration (PVCR). The transmission and reflection processes of coherent electron waves through the double barrier structure can describe resonant tunneling. For better application performance, resonant tunneling requires quantum well layers and high-energy barrier tunnel layers, which are fabricated from a high-quality heterostructures. Currently, the fabrication of good heterostructure is difficult on Si substrate. In quantum mechanics, quantum tunneling (or the tunnel effect) is a nanoscopic phenomenon in which a particle violates the principles of classical mechanics by penetrating a potential barrier or impedance that is higher than the kinetic energy of the particle. In quantum tunneling through a single barrier, the transmission coefficient or the tunneling probability is always less than one for an incoming particle that has energy less than the potential barrier height. The tunneling probability decreases with increasing barrier height and width. Despite the lower transmission coefficient, it is possible that two barriers in a row can be completely transparent for certain energies of the incident particle. Figure 8 Consider a potential profile shown above that contains two barriers located close to each other, for which, one can calculate the transmission coefficient as a function of the incoming particle energy using any of the standard methods. It turns out that, for certain energies of a particle, the transmission coefficient is equal to one, meaning that the double barrier is completely transparent for particle transmission. This phenomenon is called resonant tunneling. Resonant tunneling also occurs in potential profiles with more than two barriers (i.e two wells and three barriers).

7 7 Heterostructures are structures with two or more interfaces at the boundaries between the regions of different materials. Usually, compound semiconductors from Group (III V) are widely used to fabricate heterostructures, which are used to create the double or multiple potential barriers in the conduction band or valence band. A doublebarrier structure can be grown by molecular beam heteroepitaxy (MBE). Common semiconductor compounds used for type-i (Figure-9a) heterostructures are GaAs (gallium arsenide) and AlAs (aluminum arsenide) is used for type-ii (Figure-9b) heterostructures. In type-i, the lowest conduction-band states occur in the same part of the structure as the highest valence-band states. In type-ii, the lowest conduction-band minimum occurs on one side and the highest valence-band maximum on the other with an energy separation between the two that is less than the lower of the two bulk bandgaps. In broken-gap (Figure-9c) lineup structure, the bottom of the conduction band on one side drops below the top of the valence band on the other. (Example-InAs/GaSb gap of 150meV) Figure-9 Traditionally, RTDs have been fabricated in-group III-V (B, Al, Ga, In, Tl N, P, As, Sb, Bi) material systems, which has limited their widespread applicability. Reasonably high performance III-V resonant tunnel diodes have been realized; however, such devices have not entered into mainstream applications yet because the processing of III-V materials is incompatible with Si CMOS technology and the fabrication cost is also high. Most of semiconductor optoelectronics uses III-V semiconductors so it is possible to combine III-V RTDs to make OptoElectronic Integrated Circuits (OEICS) that use the negative differential resistance of the RTD to provide electrical gain for optoelectronic devices. However, recent works have described fabrication of group IV (C, Si, Ge, Sn, Pb) devices with Si compatible materials. These include tri-state logic devices fabricated in SiGe and a Si based field induced band-to-band tunneling transistor. Current research activity involving RTDs focuses on fabrication of group IV (Si/SiGe) materials for the integration on the silicon platform. Resonant tunnel diodes can also be realized using the Si/SiGe materials system. But the performance of Si/SiGe resonant tunnel diodes was limited due to the limited conduction band and valence band discontinuities between Si and SiGe alloys. Resonant

8 tunneling of holes through Si/SiGe heterojunctions was attempted first because of the typically relatively larger valence band discontinuity in Si/SiGe heterojunctions than the conduction band

8 8 tunneling of holes through Si/SiGe heterojunctions was attempted first because of the typically relatively larger valence band discontinuity in Si/SiGe heterojunctions than the conduction band discontinuity. This has been observed, but negative differential resistance was only observed at low temperatures but not at room temperature. Resonant tunneling of electrons through Si/SiGe heterojunctions was obtained later, with a limited peak-to-valley current ratio (PVCR) of 1.2 at room temperature. Subsequent developments have realized Si/SiGe RTDs (electron tunneling) with a PVCR of 2.9 with a Peak Current Density (PCD) of 4.3 ka/cm 2 and a PVCR of 2.43 with a PCD of 282 ka/cm 2 at room temperature. Recent work has demonstrated resonant tunneling diodes on Si substrates using better quality Fluoride alloy heterostructures. RTDs composed of CaF2-barrier/CdF2- well/caf2-barrier/si heterostructures are expected to be co-integrated with Si-Large Scale Integration (LSI). The material layout and corresponding energy bands are as follows: Figure-10 CdF 2 has large electron affinity comparable to that of Si thus, it can be used as deep quantum well layers in Si and CaF 2 /CdF 2 heterostructures. The electron affinity is the energy required to remove an electron from the bottom of the conduction band to outside of a material (to a vacuum level). The alloy of CaF 2 and CdF 2 namely Ca x Cd 1-x F 2 was investigated in order to improve RTD characteristics. The barrier height of RTD was found to be lowered by using Ca 0.5 Cd 0.5 F 2 instead of CdF 2. It was also discovered that Cd-rich (80% or higher composition) alloy could be grown with good crystallinity at higher temperature than that for pure CdF 2. RTD using the Cd-rich alloy for the well layer exhibited large peak to valley current ratio (PVCR=29) at room temperature due to the good crystallinity. Addition of a small amount of CaF 2 to CdF 2 formed a Cd-rich Ca x Cd 1-x F 2 alloy and reduced the surface roughness improving crystallinity at higher temperature. The peak current of Ca x Cd 1-x F 2 alloy-barrier RTD was larger than the pure CaF 2 barrier, which indicates that the barrier height is controlled by using Ca x Cd 1-x F 2 alloy. Thus, fluoride heterostructures has large potential as a material system to produce quantum well

9 9 structures on Si substrates where peak-to-valley current ratio of 10 to a few 100 has been obtained with good reproducibility. Figure-11 Trials of co-integration of the fluoride RTDs with Si MOSFET on a chip have been carried out but the stability of electrical properties of the RTD is insufficient to realize circuit operations well at the moment. The typical instability is due to current drifting that involves charging effects of carrier traps in the heterostructures. Thus, further investigation to improve crystalline quality of fluoride heterostructure is necessary to achieve reliable circuit operation of such device. One of the issues involved in designing RTDs is scalability. Reducing the physical structure below certain dimensions introduces undesirable surface leakage current. In addition, since the peak current through an RTD depends exponentially on the barrier thickness, it is difficult to get reproducible device operation unless the gate also controls the peak current. Integrating a transistor with the series-connected double RTDs on a common substrate performs control of the peak current. This approach requires the precise control of the layer thicknesses. The control of RTD properties may require use of commercial molecular beam epitaxy (MBE) to grow complex compound semiconductor crystal structure resulting in a higher cost. However, integration of a transistor with a pair of RTDs introduces delays to the fast bistable switching times associated with capacitive charging and discharging of the transistor gate stack. Due to this delay, the operational speed of the integrated device can be slower than the intrinsic speeds of the RTDs themselves. Additional challenges include a limited I ON /I OFF ratio of 10 compared to the ratio of 10 5 required by CMOS digital circuit designers and the inherent complexity of the integrated structure, which limits the dimensional scaling of the devices. Another issue is in obtaining high peak-to-valley current ratios for fabrication of silicon or silicongermanium tunnel diodes. Although, RTD devices have several issues to overcome as described above, fabrication in a Si compatible material structure substantially reduces the integration challenges.

10 10 RTDs combined with an array of quantum dots in a CNN (cellular nonlinear network) connection pattern can be applied very effectively to image processing tasks such as edge detection, image recognition, and noise reduction. The logic cell of this structure consists of two RTDs connected in series where the interconnections between them are achieved via tunneling in the quantum layers. When two RTDs are connected in series with opposing polarities, they have two stable operating points and can switch between the two stable points very quickly if a third terminal is added which can act as a gate. This results in a monostable-bistable transition logic element (MOBILE) configuration. Recent work has demonstrated the application of MOBILE circuits to ultra high-speed analog to digital converters. MOBILEs have also been applied to high speed (80 Gb/s) multivalued quantizers and chaos generators useful for Monte Carlo simulations. Utilization of MOBILE type structures can be applied to multi-valued threshold circuits and multi-threshold gates in general. Adding a control terminal to RTDs extends their usability to a variety of applications. This approach has been used to build Resonant-tunneling transistors (RTT). RTTs have a negative transconductance that is used in logic XOR gate with only one transistor. A number of recent works explore spin-polarized resonant tunneling, which can be useful for application in spintronic devices. Another potential application for RTDs is in photodetectors for detection of photons with low dark current count rates and high efficiency. RTDs may be useful for alternative information processing applications other than Boolean logics that require high speed; low dynamic range and low peak currents provided that manufacturing issues associated with uniformity of the tunneling barrier can be resolved.

11 11 References: David Bohm, Quantum Theory, Prentice-Hall, New York, Baym Gordon, Lectures on Quantum Mechanics, Addison-Wesley Company, Canada, H. Mizuta and T. Tanoue. The Physics and applications of Resonant Tunnelling Diodes. Cambridge University Press, Cambridge Nanoelectronics and Information Technology. Rainer Wasser, ed. Wiley-VCH, Alexander Khitun and Kang L. Wang. Cellular nonlinear network based on semiconductor tunneling structure with a self-assembled quantum dot layer th IEEE Conference on Nanotechnology. K. Maezawa and T. Mizutani, Jpn. J. Appl. Phys., Vol. 32, L42, Koichi Maezawa. Resonant Tunneling Diodes and Their Application to High-Speed Circuits. CSIC 2005 Digest, IEEE. Héctor Pettenghi, María J. Avedillo, and José M. Quintana. New Circuit Topology for Logic Gates based on RTDs. Proceedings of th IEEE Conference on Nanotechnology Nagoya, Japan, July M.A. Reed, et al. Realization of a Three-terminal Resonant Tunneling Device: the Bipolar Quantum Resonant Tunneling Transistor. Appl. Phys. Lett. 54, 1989, P. Fau, et al. Fabrication of Monolithically-integrated InAlAs/InGaAs/InP HEMTs and InAs/AlSb/GaSb resonant Interband Tunneling Diodes. IEEE Trans. Electron Dev. 48, 2001,

Resonant Tunneling Device. Kalpesh Raval

Resonant Tunneling Device. Kalpesh Raval Resonant Tunneling Device Kalpesh Raval Outline Diode basics History of Tunnel diode RTD Characteristics & Operation Tunneling Requirements Various Heterostructures Fabrication Technique Challenges Application