PHYSICAL REVIEW B VOLUME 42, NUMBER 3 15 JULY 1990-II Conduction processes in the layered semiconductor compound FePS3 V. Grasso, F. Neri, S. Patane, and L. Silipigni Istituto di Struttura della Materia, Universita degli Studi di Messina, Salita Sperone 31, P. O. Box 57, I-98166 S. Agata Messina, Italy M. Piacentini Dipartimento di Energetica, Universita degli Studi di Roma I, "La Sapienza, " I-00185 Roma, Italy (Received 4 December 1989; revised manuscript received 22 February 1990) We have measured, as a function of temperature, the thermopower, the dc conductivity, and pho- toconductivity of FePS3 single crystals. For temperatures below 430 K the thermopower shows an activated behavior, whereas at higher temperatures it is almost temperature independent. The dc dark conductivity and photoconductivity are thermally activated over the entire investigated tem- perature range. Moreover, ac-conductivity measurements have been carried out as a function of both temperature and frequency from 400 Hz to 100 kHz in the 340-500-K temperature region. The transport mechanisms involved in different temperature ranges have been identified. The re- sults have been interpreted on the basis of a simplified energy-band scheme. INTRODUCTION The layered compound FePS3 has received consider- able attention in recent years for its potential use as cathode material in lithium-anode secondary batteries' and for its magnetic properties. ' The synthesis of pure FePS3 is a complicated task be- cause of the large quantities of unreacted phase (seeming- ly sulfur) remaining in the reaction tubes. ' ' This could explain the observed nonstoichiornetry since lamellar compounds, such as FePS3, could accommodate a metal excess in the van der Waals gap. Thompson and Whit- tingham reported that their FePS3 samples were metal- rich, while Foot et al. observed a slight deficiency of iron, up to about 0. 1%, which in their opinion was more in keeping with a supposed p-type semiconduction. While the electrical transport properties of NiPS& and MnPS3 have been studied and satisfactorily interpret- ed, ' '" no systematic study of these properties has been yet performed for pure FePS3. Therefore, we have inves- tigated the temperature dependence of the FePS3 conduc- tion mechanism by means of intralayer dc dark conduc- tivity and photoconductivity and thermopower measure- ments as a function of the temperature. The experimen- tal results indicate that FePS3 is a semiconductor, and they give clear evidence of the carrier type. The thermal- ly activated behavior has allowed us to derive more de- tailed information on the distribution of the electronic density of states in the energy-gap region. We have also proposed a scheme for the conduction process consistent with the transition-metal weakly interacting model' and the obtained experimental values. We have also analyzed the temperature and frequency dependence of the FePS3 ac conductivity. The experi- mental results support the conclusions drawn for dc mea- surements and provide further information about the electrical transport processes taking place near the Fermi level. EXPERIMENT In a cryostat operating from liquid-N2 temperature up to about 500 K, we have carried out extensive studies of thermopower, in-layer conductivity, both dc and ac, and photoconductivity measurements on a single crystal of FePS3 compound 0. 5 X 3 X 4. 6 mm in size. Such sam- ples, grown and supplied by Professor R. Brec of the Uni- versity of Nantes (France), are black, flexible, and hexag- onal plates. During the measurements the pressure was maintained at about 10 Torr. All measurements were repeated several times on two samples of about equal di- mensions in order to check their reproducibility. In par- ticular, the reproducibility for the dark conductivity and photoconductivity measurements was better than 13%. Before each experimental run, the specimen was heated up to 470 K in a vacuum of — 10 Torr to avoid the for- mation of free surface adsorbates. In order to collect all data and to calculate the conductivity, photoconductivi- ty, and thermopower values in real time, a HP3421A data logger, connected to a microcomputer, was used. The dc dark conductivity and photoconductivity measurements were made at heating rates less than 1 C/min to avoid undesired effects due to thermostimulated currents. The polarization voltage (about 1 V) was provided by a HP 6115 A precision voltage supply, while the current was measured by a Keithley model 616 electrometer with 10' 0 input impedance. Photoconductivity was excited by a 75-W xenon arc lamp with a measured light intensity of about 500 mWcm . The bandwidth of the incident ra- diation was determined by a visible transmitting filter. In the Seebeck coefficient measurements, the differential method, described in a previous paper, ' was adopted. The in-layer ac-conductivity measurements were per- 42 1690 1990 The American Physical Society