Fig. 1

Results showing the importance of using the impedance spectroscopy (from by Karato and Dai (2009), for wadsleyite). a Typical results of the impedance spectroscopy. The correct measurements should include a wide frequency range showing a half circle, and the intercept of the half circle (shown by R (correct)) of the Z′ axis provides the estimate of resistivity (from Karato and Dai (2009)). In this figure, we show the results using one frequency (0.01 Hz) as Yoshino (Yoshino et al. 2008, 2006) did. These two methods lead systematic difference in the estimated resistivity. b A comparison of the results from the impedance spectroscopy and from one-frequency (0.01 Hz) measurements (incorrect method) and from the inversion of the impedance spectroscopy. Incorrect one-frequency method leads to the results similar to (Manthilake et al. 2009; Yoshino et al. 2008) (shown by YMMK), but these results show systematic difference from the results from the correct method. Activation enthalpy determined by the incorrect method is systematically lower than the true activation enthalpy. For dry samples where conductivity is due to electron holes, very little difference is seen (a small difference is due to the fact that Yoshino’s “dry” samples are not truly dry). The difference is larger for samples with larger water content. This leads to the apparent water content dependence of activation enthalpy