Silicon power devices: RF power transistor & electric power device

Introduction While wireless communications standards may come and go with developments in the latest digital coding technology, or the liberation of new fragments of the electromagnetic spectrum, the common denominator among the various communications systems is the power amplifier. Over the last few decades, the transition from vacuum tubes and other forms of amplification to solid state devices has been almost complete, especially at power levels less than 1 kW. Nowadays, at the heart of the power amplifier, we find the power transistor. In the world of RF wireless communications, the base-stations and long range transmitters use silicon LDMOS (laterally-diffused MOS) high power transistors almost exclusively. In addition to modern cellular communication systems, LDMOS devices are also used in a wide range of applications requiring radio-frequency power amplification: HF, VHF, and UHF communications systems; pulsed radar; industrial, scientific and medical (ISM) applications; avionics; and most recently in WiMAX™ communication systems. The frequency range of these applications is from a few megahertz and up to 4 GHz. While LDMOS technology is pre-eminent in high-power RF and lower frequency microwave applications, a wide variety of compound (III–V) semiconductors are used as effective power amplification devices, especially for application frequencies above about 5 GHz, and also for lower power applications such as cellular handsets, Bluetooth™ and other wireless local area networks (WLANs), which generally demand output powers of about 1 watt or below.

In electrical power devices, ac losses from a superconductor is a primary factor which determines their usefulness as commercial power equipment. For this reason, extensive studies have been carried out on the losses of Bi{sub 2}Sr{sub 2}Ca{sub 2}Cu{sub 3}O{sub 10}/Ag, [Bi(2223)/Ag], tapes. These studies were mostly limited to a single isolated tape. However, a conductor in a power device is surrounded by other conductors and the precise magnetic field distribution around it is very different from that for a single conductor carrying currents or in ac fields. Since the precise field distribution in and around a superconductor is critical in determining the losses, it is very important to measure and to understand the losses in Bi(2223)/Ag tapes which are surrounded by other tapes as in a power device. Taking this fact into consideration, recently the authors have studied ac losses in stacks of Bi(2223)/Ag tapes in parallel and perpendicular applied fields and shown that they can calculate the losses in these cases utilizing the critical state model if a number of appropriate factors about properties of the tape are taken into a consideration. However, in a power device such as a transformer, magnetic fields near the ends of a solenoid vary from parallel to perpendicular with the tape face. Thus, it is important to learn the behavior of the losses in the stacks of Bi(2223)/Ag tapes with respect to the variations in the angle between the applied field direction and the tape face. In order to accomplish this, they measured the angular dependence of the losses in the stacks which were made from two different Bi(2223)/Ag tapes. Here they report this result and discuss under what conditions they can calculate the losses with a reasonable accuracy. The angular dependence of the losses in ac applied fields were measured using a series of stacked Bi(2223)/Ag tapes having the angles with the direction of applied fields of 0, 7.5, 15, 30, 45, 60, and 90 degrees. The measured values of the losses were compared with the calculated values. The calculations were performed using the measured losses for the 0 and 90 degrees orientations and assuming the losses can be separated for the currents circulating in the plane of and across the tapes. It was shown that at very high fields, i.e., well above the full penetration fields, this assumption was justified by the observed good agreement between the measured and the calculated losses for both tapes. However, at the low fields, significant deviations between the calculated and the measured were seen for both of the tapes. One possible contributing factor to this discrepancy is the non-ideal shape of the tape cross sections. This can causes the field distribution near the edges of the slabs to deviate from that for an assumed uniform infinite slab. This makes the actual fields at the edge regions to be greater than the applied fields. Thus, the losses will be higher than those calculated. Although not having a full understanding of the losses at low fields is of concern, it will not be of a great importance in the designs of power devices in practice. This is due to the fact that the values of the losses at low fields are orders of magnitudes lower than those at high fields and thus the primary concern for the power device development is the reduction of the losses at high fields.

Japanese efforts on coated conductor processing and its power applications: New 5 year project for materials and power applications of coated conductors (M-PACC)

Development of YBCO power devices in Japan

Fifth generation communication networks promise extremely high data throughputs, low latencies, and ultra-high reliability by employing small cells in high density networks. This will increase the use of highly inefficient microwave power amplifiers, that exhibit efficiencies sometimes as low as 10% for advanced multi-carrier waveforms. To overcome these inefficiencies and enable next generation technology development, the internal distributed behaviour of the packaged microwave power transistor has to be fully characterised, which is impossible using conventional port-based measurements. In this dissertation, the development of a set of multiphysics measurement techniques that capture the distributed electrical, electromagnetic, and thermal device behaviour is described. These new multiphysics measurement techniques enable the characterisation and direct visualisation of the inefficient device performance, providing a basis for future design optimisation. The techniques combine a non-linear vector network analyser, electro-optic, load-pull and thermoreflectance measurement systems that enable large signal time-domain electrical, distributed multi-harmonic vector electric field (E-field), and transient and steady-state thermal measurements. The multiphysics measurement set-up was used to characterise various transistors and PAs including 260 W silicon (Si) laterally diffused metal oxide semiconductor (LDMOS) field effect transistors in both air-cavity ceramic and plastic packages, 25 W gallium nitride (GaN) on silicon carbide high electron mobility transistors (HEMTs), 250 W GaN HEMT and 360 W Si LDMOS Doherty power amplifiers. The E-field measurements enabled the first ever imaging of parametric odd-mode oscillations within multi-die packaged high-power microwave transistors. Additionally, the measurement system was used to verify the effectiveness of in-package oscillation suppression circuits. As high-resolution E-field measurements can be slow, a surrogate modelling-based measurement algorithm was implemented to accelerate measurements 10-fold. Transient thermal measurements helped to identify abnormal transistor finger-to-finger heating in GaN HEMTs and revealed thermal differences along the gate width in GaN-based asymmetric transistors for Doherty power amplifier applications.

RECENT progress in wide bandgap power (WBG) switches shows great potential. Silicon carbide (SiC) is a promising material for power devices with breakdown voltages of several hundred volts up to 10 kV. SiC Schottky power diodes have achieved widespread commercial acceptance. Recently, much progress has been made on active SiC switches, including JFETs, thyristors, BJTs, IGBTs, and MOSFETs. Many a great promise has been made, and wondrous claims abound, but the question remains: will they live up to the hype? We explore this question for the class of high-frequency, hard-switched converters with input voltages of up to 600 VDC and power throughputs in the kilowatt range. Experimental evidence shows that both superior efficiency and higher power density may be obtained via the use of SiC MOSFETs. A direct comparison is made using silicon power devices (IGBTs and MOSFETs) and SiC MOSFETs in a 200 kHz, 6 kW, 600 V hard-switched converter. The losses are measured and conduction and switching losses of the active devices are estimated. Total losses can be reduced by a factor of 2-5 by substitution of SiC MOSFETs for Si active power devices.

New lithographically processable materials with high ampacity are in demand to meet the increasing requirement for high operational current density at high temperatures existing in current pathways within electronic devices. To meet this demand, we report an approach to fabricate a high ampacity (∼100 times higher than Cu) carbon nanotube-copper (CNT-Cu) composite into a variety of complex nano-scale, planar and multi-tiered current pathways. The approach involved the use of a two-stage electrodeposition of copper into a pre-patterned template of porous, thin CNT sheets acting as the electrode. The versatility of this approach enabled the realization of completely suspended multi-tier, dielectric-less 'air-gap' CNT-Cu circuits that could be electrically isolated from each other and are challenging to fabricate with pure Cu or any metal. Importantly, all such complex structures, ranging from 500 nm to 20 μm in width, exhibited ∼100-times higher ampacity than any known metal, with comparable electrical conductivity as Cu. In addition, CNT-Cu structures also exhibited a superior temperature stability compared to the ∼10-times wider Cu counterparts. We believe that the combination of our approach and the properties demonstrated here are vital achievements for the future development of efficient and powerful electrical devices.

Electrochemical contributions: Hans Wenking (1923–2007)

DeviceMaker—The In-House Developed TCAD Tool to Analyse and Estimate Semiconductor Device Internal Electrical Properties, and Measured DC Performance Characteristics of Commercially Available RF|Microwave and High Power Transistors

In this paper, thermal impedances of Silicon carbide (SiC) power module assembled with a Schottky Barrier Diode (SBD)–wall–integrated trench MOSFET (SWITCH–MOS) and a conventional trench gate MOSFET (UMOS) are measured. Then they are compared to show the advantage of SWITCH–MOS for thermal impedance measurement. SiC power device is attracting attention as a candidate for future power device, since it can be operated at higher current density and higher switching speed than conventional silicon (Si) power device. The power module packaging technology for SiC power device should have low thermal impedance (both transient and steady state thermal resistance) to bring out a potential ability of the device. Thermal design often has difference from measurement due to various modeling errors such as power device heat generation, module structure, interfacial thermal resistance, and cooling environment. Therefore, accurate thermal impedance measurement is an essential technology. However, the thermal impedance measurement of SiC power modules has the significant issue due to the instability of the body diode built into the SiC–MOSFET. The thermal impedance is the junction temperature rise per heat generation amount of the device. The temperature dependence of the knee voltage of the body diode is used to measure the junction temperature of the MOSFET. Since the knee voltage of SiC–MOSFET changes depending on the gate voltage, a high negative gate bias was necessary to avoid this effect. However, the negative gate bias stress may cause serious device damage such as threshold shift. The SWITCH–MOS which has SBD built in near the MOSFET channel is expected to stabilize the junction temperature measurement and avoid the risk of device deterioration. Figure 1 shows the cross–sectional schematic of the SWITCH–MOS and the conventional UMOS. The SBD are formed on the trench sidewall in SWITCH-MOS, and the device structure is almost same to the UMOS. The temperature is detected by the PiN diode for the UMOS and by the SBD for the SWITCH–MOS. Both of them measure the temperature inside of the MOSFET cell. The appearance of the test module is shown in Fig. 2. Each type of the test module is composed of the SiC die (SWITCH–MOS, UMOS, and SBD), an SiN–AMC insulating circuit substrate and an Al baseplate. The components are attached each other using AuGe solder. The junction temperature is measured from the built–in voltage of the body diode in response to a constant current flowing from the source to the drain. In the case of the UMOS, the gate bias of −18 V is applied. The gate bias is zero for the SWITCH–MOS. The thermal impedance of each test module was measured by static mode (Fig. 3). Then, the thermal structure function was calculated by transient thermal analysis (Fig. 4). Both the thermal impedance and the structure function for the SWITCH–MOS and UMOS were in good agreement. Since the SWITCH–MOS does not need negative gate bias, the threshold shift issue does not occur in principle. In addition, it is possible to perform a transient thermal testing with a large current for a long time. Because the forward degradation does not occur even if a large current is applied to the SBD in the SWITCH–MOS. The SBD built into the SWITCH–MOS makes it easier and more accurate to measure the junction temperature of the SiC–MOSFET. The thermal impedance measurement using SWITCH-MOS not only allows the precise thermal modeling of SiC power modules to be performed, but also has the potential to be applied to the calibration of the thermal design and the evaluation of thermal properties inside of the package structure. Figure 1

As power devices are series connected for voltage sharing, loss of gate drive synchronization and/or variation in device switching time constant can cause voltage imbalance. Capacitors (in snubbers) are usually added to maintain series voltage balance, however, in snubberless designs, where active gate drivers are used for voltage balancing during transients, it is necessary to evaluate the limits of the power device under transient unsynchronized switching. In series devices, desynchronization of the gate drivers in series connected devices will cause the faster switching device into avalanche during turn-OFF. Power device failure from BJT latch-up in MOSFETs and thyristor latch-up in IGBTs can result in potentially destructive consequences for the entire converter. Failure of the power device under avalanche is exacerbated by the (i) device commutation rate (ii) device junction temperature (iii) magnitude of gate drive switching mismatch and (iv) ratio of the DC bus voltage to intrinsic breakdown voltage of the device. This paper uses experimental measurements and finite element models to investigate the limits of power device failure under transient unsynchronized switching of series connected SiC trench MOSFET and silicon IGBT devices.

RF power amplifier circuits other than Class A have a previously unreported operating characteristic which can cause emitter-base reverse breakdown once each RF cycle. The breakdown is not necessarily observable at the external terminals of the transistor; it can be deduced from the observed emitter-base terminal voltage and base current and the transistor internal lead inductances and junction capacitances. Certain symptoms of RF power transistor deterioration and failure reported by other workers appear to be consistent with the possible effects of such repetitive breakdowns. Common-base amplifiers should be more likely to suffer from this problem than common-emitter amplifiers. Transistor specialists disagree in their predictions of whether such operation would damage the RF power transistor. Controlled experiments could provide definitive information as to the effects on the transistor of such operation.

The paper provides an overview of the state of the art of RF power transistors. Several material issues relevant to RF power devices are discussed and recently published performance data of different RF transistor types is presented. Future requirements and challenges for RF transistors based on the latest issue of the ITRS (open to public access since Dec. 2005) are reviewed.

RF power amplifiers use a wide range of active devices, such as vacuum tubes and transistors (usually BJTs or MOSFETs). Vacuum tubes, still widely used in high-power RF amplifiers.

The unique properties of nanoscale materials offer excellent prospects for interfacing biological recognition events with electronic signal transduction and for designing a new generation of bioelectronic devices exhibiting novel functions. In this Highlight I address recent research that has led to powerful nanomaterial-based electrical biosensing devices and examine future prospects and challenges. New nanoparticle-based signal amplification and coding strategies for bioaffinity assays are discussed, along with carbon-nanotube molecular wires for achieving efficient electrical communication with redox enzyme and nanowire-based label-free DNA sensors.