(image) (image) [Previous] [Next]

Degradation of Electrical Parameters of Power Semiconductor Devices – Process Influences and Modeling

Chapter 4 Accurate high temperature measurements

The degradation mechanisms in MOSFETs listed in Section 1.1 usually have a strong temperature activation. This means that these mechanisms become more pronounced when the MOSFET is operated or stressed at temperatures above room temperature. For a detailed characterization of the device it is thus important to be able to conveniently change the temperature of the device in a fast and reliable manner.

There are two main approaches for controlling the temperature of a device during an experiment. For measurements directly on the wafer the device temperature is usually set using a thermal chuck. For packaged devices a temperature-regulated furnace is used. Both approaches suffer from the drawback that reliable changes of the temperature usually require minutes to hours. Particularly in conventional systems the metal shielding is thermally coupled to the chuck and thus follows all temperature switches. As the metal shielding also carries the needle holding manipulators the thermal expansion following a temperature switch causes the needles to move slightly with respect to the pads of the device which may result in contact loss. For a reliable temperature switch it is mandatory to wait until the expansion of the shielding is completed before a long-lasting measurement can be started. This time period can last several hours which renders measurements at different temperatures a cumbersome task.

A second major drawback of thermal chuck systems and dedicated furnaces is that the highest achievable temperature is usually limited to about 200 °C to 300 °C. It is possible to build a system which can endure temperatures up to 700 °C [BW08], but only by using expensive materials which can sustain these high temperatures within the furnace for mounting and connecting the device.

The above mentioned drawbacks can be overcome with justifiable effort using a poly-heater, an in-situ heating structure surrounding the device under test directly on the wafer [PobegenTDMR13; MT92; Mut+03; MW04; Wan+06; KWS07; Sch+07; Aic+10c]. The following Sections introduce the use of the poly-heater for measurements featuring fast and reliable temperature changes beyond temperatures provided by conventional chucks or furnaces [PobegenTDMR13].

4.1 The poly-heater

The poly-heater is a simple arrangement of two polycrystalline silicon (poly) lines around a device as sketched in Fig. 4.1.

(image)

Fig. 4.1: Schematic drawing of the poly-heater (red) with contact pads (yellow) on a part of an Si or silicon carbide (SiC) wafer (gray). Two devices (blue) are centered in the middle of the heater. Oxide layers between the poly and the metal are omitted for illustration purposes. All dimensions are in micrometer.

An electric current fed to the two contacts of the poly-heater causes Joule heating in the heater which also increases the temperature of the adjacent regions, i.e. especially the device under test (DUT). Because the chuck acts as a heat sink, the heat flows mainly towards the bottom. This introduces a temperature gradient within the whole material stack. Of course, the heater and the device need to be electrically isolated from each other to be able to independently bias the heater and the device to perform arbitrary electrical experiments. This electrical isolation introduces also a thermal isolation, as also visible in Fig. 4.2.

(image)

Fig. 4.2: Schematic cross section of the poly-heater and the device highlighting the structure between the device and the heater. The poly-heater is only partially shown and extends further to the left.

Consequently, the device temperature will generally differ from the temperature of the heater itself and must be determined individually.