Thomas Research Group

Interfacial States in Semiconducting BaTiO₃

Many commerical devices exist which are built on this material, such as thermal switches on motors (when the temperature rises above a set point, the resistance through the switch increases, shutting off the motor) and constant temperature heating devices (when it reaches the set point temperature, an increase in current will cause the temperature to rise, but this will cause an increase in resistance, which decreases the current, which decreases the temperature and the device becomes self-regulating at that temperature). In 1992, worldwide production of BaTiO₃ thermistors exceeded a half billion units with European companies such as Philips and Seimens leading the way.

However, in spite of all of this interest, an accurate model of the cause of the PTCR effect remains a contentious issue. One group of models assumes that a two-dimensional layer of interfacial acceptor states occurs at the grain surface, which may be intrinsic in origin (simply a result of the termination of the crystal lattice) or extrinisic due to the surface segregation of impurities or the ion absorption of O. The second model assumes that the barrier states arise from the Ba vacancy-rich region which originates due to diffusion during the sintering process. The grain boundary region is seen to be an n-i-n type junction and no experiment has clearly answered why. Current debate still centers around these ideas and further development in the PTCR materials can be aided by a clearer understanding of the origin of this effect.

We are collaborating with the group of Prof. Anatoly Belous from the Institute of General and Inorganic Chemistry in Ukraine to reach a better understanding. They fabricate these materials and have even started a small company called OXIDE to market devices such as these amongst others. We have been using Scanning Tunneling Microscopy (STM) and Spectroscopy (STS) to probe the interfacial region to find these states and more accurately characterize them. We have observed a rather anomolous behaviour when scanning the tip bias voltage during tunneling. This hysterises effect is observable on some locations on the surface and not on others. In the image on the left, the site at point B shows hysteresis while the site at A does not. We have employed a new spectroscopic method, involving a voltage step/current relaxation method. A sudden step in the tip bias voltage induces a change in the tunneling current. We measure the rate of relaxation to this new tunneling level and find that at times, the surface relaxation occurs in microseconds while at other places it is characterized by a time constant of 100's of milliseconds. It is at these locations that the hystresis is observed. The image to the right shows the two voltage step relaxation spectra for the two sites in the STM image. The current at point A response time is limited by the electronics, while that at B is controlled by the charging kinetics of the trapping sites located at B.

The above experiments were performed under atmospheric conditions but we have recently performed some further experiments in a controlled atmosphere environment. When a sample is placed into a UHV environment and heated to drive off any adsorbed gases, the resulting sample shows no hysteresis effects at any location. By contrast, when 1 torr of O₂ is admitted, the sites manifest increasing hysteresis when the bias potential is swept between -2 V and +2 V. Apparently, these trapping states are closely related to adsorbed oxygen, perhaps in an ionic form. And since, the PTCR effect only is shown when the sample has been processed in an oxidizing environment, it is reasonable to presume that these interfacial sites that have been identified are closely related to those responsible for the PTCR effect.