Geoscience Reference
In-Depth Information
Figure 1.
Different ways to measure electrical conductivity. (a) standard procedure, no load; (b)
under load; (c) measuring at two places, one under load; (d) measuring a battery current.
Though valence fl uctuations on the oxygen anion sublattice may appear to be of
interest only to the narrowest of the specialists, they can have far-reaching conse-
quences the capability of rocks to generate electric currents and, hence, to generate
electric or EM signals.
To provide some background I start with a tutorial about electrical conductivity
measurements. The electrical conductivity is typically measured with a set-up such
as depicted in Figure 1a: a voltage is applied between a pair of electrodes on opposite
surfaces of a fl at sample and the current is recorded with an ammeter. If the effect of
stress is to be studied, two pistons may be used to apply a force as shown in Figure 1b.
Figure 1c depicts a set-up, where the conductivity is measured at two places, spot 1 be-
tween the pistons and spot 2 where no stress is applied. Conventional wisdom suggests
such an experiment will lead to nothing new: the conductivity across spot 2 should not
change when stress is applied to spot 1. Figure 1d shows yet another set-up: a circuit
without a voltage source but a pair of steel pistons at the center, in electrical contact
with the rock and connected to ground through one ammeter plus a Cu stripe as a con-
tact along the periphery, connected to ground through a second ammeter. This circuit
does not apply a voltage across the rock sample and, hence, will not measure electrical
conductivity. Instead it will measure currents that fl ow out of the stressed rock under
a self-generated voltage differential between the pistons and the periphery of the rock.
Figure 2a shows a typical current-voltage (I-V) plot recorded with a set-up as in
Figure 1a. The sample is a dry gabbro. Except for the low voltage region, the I-V plot
is linear as it should be for an Ohmic response. Figure 2b gives an example for the
conductivity of dry granite measured with a set-up as shown in Figure 1b. During the
fi rst 7 min, the current is recorded without applying stress to the rock. The conduc-
tivity before loading is 0.7 × 10
−6
S/m. Upon loading, the conductivity rises sharply,
increasing by a factor of 3-4 to about 2.5 × 10
−6
S/m and then continues to increase
slowly. Stress-induced increase in the conductivity is commonly explained by better
point-to-point contacts between grains under load (Glover and Vine, 1992). However,
the next experiment will show that we are dealing with a more complex and much
more interesting phenomenon.
The data shown in Figure 2c were obtained with a set-up as depicted in Figure 1c.
We measured simultaneously the currents at spot 1 and at spot 2, that is between a pair
of electrodes away from the stressed rock volume. The bold line shows the current at
spot 1, the thin line the current at spot 2. After about 6 min, when we began applying
