DMOS transistors in integrated power technologies are often subject to significant self-heating and thus high temperatures, which can lead to device failure and reduced lifetime. Hence, it must be ensured that the device temperature does not rise too much. For this, the influence of the on-chip metallization must be taken into account because of the good thermal conductivity and significant thermal capacitance of the metal layers on top of the active DMOS area. In this paper, test structures with different metal layers and vias configurations are presented that can be used to determine the influence of the onchip metallization on the temperature caused by self-heating. It will be shown how accurate results can be obtained to determine even the influence of small changes in the metallization. The measurement results are discussed and explained, showing how on-chip metallization helps to lower the device temperature. This is further supported by numerical simulations. The obtained insights are valuable for technology optimization, but are also useful for calibration of temperature simulators.
DMOS transistors are often subject to high power dissipation and thus substantial self-heating. This limits their safe operating area because very high device temperatures can lead to thermal runaway and subsequent destruction. Because the peak temperature usually occurs only in a small region in the device, it is possible to redistribute part of the dissipated power from the hot region to the cooler device areas. In this way, the peak temperature is reduced, whereas the total power dissipation is still the same. Assuming that a certain temperature must not be exceeded for safe operation, the improved device is now capable of withstanding higher amounts of energy with an unchanged device area. This paper presents two simple methods to redistribute the power dissipation density and thus lower the peak device temperature. The presented methods only require layout changes. They can easily be applied to modern power technologies without the need of process modifications. Both methods are implemented in test structures and investigated by simulations and measurements.
Advanced power semiconductors such as DMOS transistors are key components of modern power electronic systems. Recent discrete and integrated DMOS technologies have very low area-specific on-state resistances so that devices with small sizes can be chosen. However, their power dissipation can sometimes be large, for example in fault conditions, causing the device temperature to rise significantly. This can lead to excessive temperatures, reduced lifetime, and possibly even thermal runaway and subsequent destruction. Therefore, it is required to ensure already in the design phase that the temperature always remains in an acceptable range. This paper will show how self-heating in DMOS transistors can be experimentally determined with high accuracy. Further, it will be discussed how numerical electrothermal simulations can be carried out efficiently, allowing the accurate assessment of self-heating within a few minutes. The presented approach has been successfully verified experimentally for device temperatures exceeding 500 ◦C up to the onset of thermal runaway.
Modern power DMOS transistors greatly benefit from the continuous advances of the technology, which yield devices with very low area-specific RDS,on figures of merit and therefore allow for significantly reduced active areas. However, in many applications, where the devices must dissipate high amounts of energy and thus are subjected to significant self-heating, the active area is not dictated by RDS,on requirements, but by the energy constraints. In this paper, a simple method of improving the energy capability and reliability of power DMOS transistors operating in pulsed conditions is proposed and experimentally verified. The method consists in redistributing the power density from the hotter to the cooler device regions, hence achieving a more homogeneous temperature distribution and a reduced peak temperature. To demonstrate the principle, a simple gate offset circuit is used to redistribute the current density to the cooler DMOS parts. No technology changes are needed for the implementation, only minor changes to the driver circuit are necessary, with a minimal impact on the additional required active area. Improvements in the energy capability from 9.2% up to 39% have been measured. Furthermore, measurements have shown that the method remains effective also if the operating conditions change significantly. The simplicity and the effectiveness of the implementation makes the proposed method suitable to be used in a wide range of applications.
The experimental characterization of the thermal impedance Zth of large power MOSFETs is commonly done by measuring the junction temperature Tj in the cooling phase after the device has been heated, preferably to a high junction temperature for increased accuracy. However, turning off a large heating current (as required by modern MOSFETs with low on-state resistances) takes some time because of parasitic inductances in the measurement system. Thus, most setups do not allow the characterization of the junction temperature in the time range below several tens of μs.
In this paper, an optimized measurement setup is presented which allows accurate Tj characterization already 3 μs after turn-off of heating. With this, it becomes possible to experimentally investigate the influence of thermal capacitances close to the active region of the device. Measurement results will be presented for advanced power MOSFETs with very large heating currents up to 220 A. Three bonding variants are investigated and the observed differences will be explained.
Large power semiconductors are complex structures, their metallization usually containing many thousands of contacts or vias. Because of this, detailed FEM simulations of the whole device are nowadays not possible because of excessive simulation time.
This paper introduces a simulation approach which allows quick identification of critical regions with respect to lifetime by a simplified simulation. For this, the complex layers are replaced by a much simpler equivalent layer, allowing a simulation of the whole device even including its package. In a second step, precise simulations taking all details of the structure into account are carried out, but only for the critical regions of interest. Thus, this approach gives detailed results where required with consideration of the whole structure including packaging. Further, the simulation time requirements are very moderate.
DMOS transistors in integrated smart power technologies are often subject to cyclic power dissipation with substantial selfheating. This leads to repetitive thermo mechanical stress, causing fatigue of the on-chip metallization and limiting the lifetime. Hence, most designs use large devices for lower peak temperatures and thus reduced stress to avoid premature failures.
However, significantly smaller DMOS transistors are acceptable if the system reverts to a safer operating condition with lower stress when a failure is expected to occur in the near future. Hence, suitable early-warning sensors are required. This paper proposes a floating metal meander embedded between DMOS source and drain to detect an impending metallization failure. Measurement results of several variants will be presented and discussed, investigating their suitability as early warning indicators.
An experimental study of a zero voltage switching SiC boost converter with an active snubber network
(2015)
This paper presents a quasi-resonant, zero voltage switching (ZVS) SiC boost converter for an output power of up to 10 kW. The converter is realized with an easily controllable active snubber network that allows a reduction of switching losses by minimizing the voltage stress applied to the active switch. With this approach, an increase of the switching frequency is possible, allowing a reduction of the system size. Experiments show a maximum converter efficiency up to 99.2% for a switching frequency of 100 kHz. A second version of the converter enables a further size reduction by increasing the switching frequency to 300 kHz while still reaching a high efficiency up to 98.4 %.
A gate driver approach is presented for the reduction of turn-on losses in hard switching applications. A significant turn-on loss reduction of up to 55% has been observed for SiCMOSFETs. The gate driver approach uses a transformer which couples energy from the power path back into the gate path during switching events, providing increased gate driver current and thereby faster switching speed.
The gate driver approach was tested on a boost converter running at a switching frequency up to 300 kHz. With an input voltage of 300V and an output voltage of 600V, it was possible to reduce the converter losses by 8% at full load. Moreover, the output power range could be extended by 23% (from 2.75kW to 3.4 kW) due to the reduction of the turn-on losses.
Gallium nitride high electron mobility transistors (GaN-HEMTs) have low capacitances and can achieve low switching losses in applications where hard turn-on is required. Low switching losses imply a fast switching; consequently, fast voltage and current transients occur. However, these transients can be limited by package and layout parasitics even for highly optimized systems. Furthermore, a fast switching requires a fast charging of the input capacitance, hence a high gate current.
In this paper, the switching speed limitations of GaN-HEMTs due to the common source inductance and the gate driver supply voltage are discussed. The turn-on behavior of a GaN-HEMT is simulated and the impact of the parasitics and the gate driver supply voltage on the switching losses is described in detail. Furthermore, measurements are performed with an optimized layout for a drain-source voltage of 500 V and a drain-source current up to 60 A.
