Electromagnetic Radiation from Rock During Uniaxial Compression Testing: the Effects of Rock Characteristics and Test Conditions
1. Introduction

A number of laboratory studies on the mechanics of rock have revealed the electromagnetic behavior of rock during the failure process. For example, Ogawa et al. (1985) detected electromagnetic radiation (EMR) events at frequencies of 10 kHz to 100 kHz from granite when struck with a hammer, and Cress et al. (1987) reported EMR events of 900 Hz to 5 kHz during compression tests on several rocks. Hadjicontis and Mavromatou (1994) observed transient variations in the electric field prior to failure of a rock sample which is subjected to uniaxial compression at a variable rate. Yoshida et al. (1997) also observed EMR events during the slip failure of granite, and Rabinovitch et al. (2000) confirmed the emission of 100 kHz EMR in the pre-failure and post-failure regions. Rabinovitch et al. (1995), through observed of EMR and acoustic emissions (AE) during rock fracture processes, found that the two signals were released almost simultaneously. Several other reports have described similar findings, such as O'Keefe and Thiel (1995). Frid et al. (1999) observed that granite with higher Poisson ratio exhibits less EMR activity in triaxial compression under confining pressure.

EMR during rock failure has thus been widely recognized, and a number of researchers have attempted to apply the EMR effect as a forecasting tool for rock bursts (Frid, 1997), volcanic eruptions (Tomizawa and Yoshino, 1992) and earthquakes (Sobolev, 1975; Gokhberg et al., 1982; Warwick et al., 1982).

The present study describes observations of EMR emission from various types of rock during failure under various experimental conditions of uniaxial compressive loading. This study expands on previous work as follows. While Cress et al. (1987) measured EMR for three types of rock (granite, basalt and marble), they did not compare their findings with rock characteristics. Many other reports describe results for only one type of rock. In this study, seven types of rock are investigated, and the relationships between EMR and rock characteristics are examined. Rabinovitch et al. (2000) also recommended performing further research on EMR in the post-failure region using a high-stiffness testing machine. In their research, they used a stiff testing machine, whereas many other researchers have used low-stiffness testing machines, which induce an abrupt process in the post-failure region. Thus, the effect of loading methods has been largely ignored in previous research. The present report also describes the relationships between testing conditions, including the loading method, and the conditions of EMR generation.