Experimental Studies of Electrical Fields on a Breaking Rock
Sample
Zhenya Zhu, F. Dale Morgan, Chris J. Marone, and M. Nafi Toksöz
Earth Resources Laboratory
Department of Earth, Atmospheric, and Planetary Sciences
Massachusetts Institute of Technology
Cambridge, MA 02139
Abstract
When a rock sample is pressed by a force, the pressure on the crystal lattice generates an electrical
field around the quart grains due to the piezoelectric effect. If a rock is saturated by conductive fluid,
the relative motion between the pore fluid and the matrix solid generates an electromagnetic field due
to seismoelectric conversion, and the permeating of fluid into new microcracks made by the pressure
changes the fluid distribution and the natural potential level.
In this paper, we measure the electrical fields on dry and water-saturated Westerly granite cylinder
samples during their breaking. Experimental results show that there are two kinds of mechanisms that
generate two kinds of electrical fields during rock breaking: (1) Pressure, or rock breaking, generates
an electrical potential on the dry rock surface due to piezoelectric effect; and (2) the potential on a
dry sample due to a piezoelectric effect is small, and its polarization depends on the characteristic and
orientation of quartz grains around the measurement point. Experiments with water-saturated granite
samples record two electrical fields: An electromagnetic wave due to seismoelectric conversion, and the
dc or low-frequency electrical potential due to the piezoelectric effect, which is an important indicator of
rock breaking.
1
Introduction
We can observe variations of a natural electrical potential in the ground and electromagnetic waves in
the air during an earthquake initiated by rock breaking, which also excites acoustic waves. Fraser-Smith
et al. (1990) recorded the magnetic field before and after the Ms 7.1 Loma Prieta, California, earthquake of
October 17, 1989. The geomagnetic field at low-frequency range (0.01-10 Hz) before the earthquake, and the
electromagnetic field at high-frequency (10 Hz-32 kHz) during the earthquake, were recorded at locations
about 7 km and 54 km from the epicenter, respectively. Park et al. (1993) reviewed electromagnetic precursors
before earthquakes and discussed the possible mechanisms. They concluded that the pressure in formation
and rock breaking always generate electric, magnetic or electromagnetic fields.
If a quartz crystal is pressed by a force in a certain direction, an electrical field will be generated on
the quartz surface due to the piezoelectric effect. When an acoustic wave propagates through a rock, the
acoustic pressure generates a piezoelectric field around quartz grains. Thus a piezoelectric field is an electric
potential field on the grain surface.
When a porous rock is saturated by an electrolyte, double electric layers form at the interface between
the fluid and solid (grains). The fluid-flow driven by an acoustic wave, or breaking, can induce an electromagnetic wave in fluid-saturated porous rock due to seismoelectric conversion. The natural electrical field is
related to the adsorption, permeability, and some chemical process in fluid-saturated rock. Zhu et al. (1999,
2000) investigated the seismoelectric fields generated by acoustic waves in fluid-saturated porous media and
borehole models. Because the pressure on a rock makes microcracks within the rock, the fluid motion inside
the cracks changes the distribution of fluid and the natural potential.
1
In order to study electrical fields in a breaking rock, we perform laboratory experiments with Westerly
granite to measure the electrical signals generated in dry or wet granite samples at different pressure rates
during rock breaking.
2
Measurement System
Figure 1 shows the measurement system used in our experiments. A rock cylinder is placed between two
plates of a press. Plastic films isolate the rock sample electrically from the ground of the press. An acoustic
P-wave transducer is mounted on the base of the press. We then make three ring electrodes on the rock
sample with conducting glue and wires.
Figure 1: Diagram of setup to measure electric signals on a rock sample when it is pressed and broken. A
P-wave transducer is mounted on the base of the press. The acoustic wave received by the transducer is
recorded and applied as a trigger for the Gagescope recording system.
The core samples used in this study are Westerly granite, which comes from Westerly, Rhode Island. All
samples were cored from the same granite block using a core drill machine. The core samples were polished
on the ends with a diamond drill while squeezing on the sides of the sample with a conical base to ensure
that the ends were also parallel. The sizes of the samples are 2.54 cm in diameter and 10 cm in length. It
is composed of 27.5% quartz, 35.4% microcline, 31.4% plagioclase (with 17% anorthite), and 4.9% biotite.
Total porosity is 0.9% and density is 2.646 g/cm3 (Brace and Orange, 1968). Wong et al. (1989) found
that Westerly granite has crack apertures that vary from 0.7 to 0.002 micrometers in width and the crack
surface area per volume (Sv) was measured to be approximately 7.89 mm2 /mm3 . The dry samples have
been heated in a vacuumized oven at 120o C for eight hours. The wet samples are saturated with tap water
in a vacuumized system. The conductivity of the tap water is about 0.57mS.
The multichannel recording system, Gagescope, records the four electrical signals coming from the three
electrodes and the transducer. The electric pulse received by the transducer is also used as the trigger signal.
The recording system records the signals before and after the trigger. When an acoustic pulse through
the amplifier is higher than the threshold of the preset trigger level of the recording system, the system is
triggered and records four traces of the signals. Sometimes more than one set of the signals can be recorded
2
when more than one sound is excited before the rock is broken completely. We can select the displacement
rate of the press before measurement. Because the acoustic signals recorded by the system are converted
from acoustic waves with the transducer mounted on the press base, the arrival time of acoustic signals
and electrical signals coming from the electrodes should be a little different, even though the same source
generates these signals. In our measurements the sampling rate of the recording system is 0.2 ms. Each
trace records 1024 points. Half of them are recorded before the trigger.
3
Dry Granite Sample
We performed the measurement with a dry sample at the displacement rate of 10−2 mm/sec. Figure 2 shows
the three electrical signals received by the electrodes on the rock sample and the acoustic waveform (Figure
2d) received by the transducer (Figure 1). In order to analyze more details of these signals, we change the
horizontal (time) scale and vertical (amplitude) scale, and plot them again in Figures 3 and 4, respectively.
*‘“’” ‹•—–‰ ™˜Œ‰ šw›œ” ™œ@•Ÿž "*)&(,+.- &'&/102 '43! 5
67)8- $9)&/;:8<
p
q@r t
„…†
q
ƒz
€
{
u=r t
|~ ‚
y
~
u
ovuwr t
| }{
o
q
xy z
oeqsr t
!#"%$&'&(
= op
ovuwrq
ouwr us‡
£
AB C
³´µ
A
NF
H
I JH
EF G
°­ ¯
¦
§®
>D
B C
¡
¥
­¬ ¨«
©ª
>=A
§¨
>AB C
>@?
¥
»
D
K
¤
±²
DB C
LM
u=r q
¢
QRS
IK O
u=r u@‡
¼%ÁsÀmÂw¿Ã
ĝ¾Å—¹ ¿9ÆÈÇÉ»
Z*^_]a`cb X&\ed1f2V \hg!V ij7k8b [hk&dml&n
?
PG
ouwr ucp
u=r u@p
ˆY‰ ŠŒ‹Ž%@‹@‘
>D
B%A
>2DB DT
>2D
B D?
DB D?
U
V WYX#Z*[X&\&]
DB DT
¡¤
¡£
¡¢
DB A
¡ ¦Y¶ ¥
¡ ¦Y¶ ¦c·
¦¶ ¦ ¤
¡ ¦Y¶ ¦ ¤
¸Y¹ ºŒ½
» ¼¾@»@¿@À
¦¶ ¦_·
¦¶ ¥
Figure 2: The electrical signals (a, b, and c) and the acoustic wave (d) recorded with a dry granite sample
during its breaking. The displacement rate of the press is 10−2 mm/sec.
We recorded a very strong acoustic wave and electrical signals at different locations on the rock surface
when the rock was breaking. Changing the horizontal (time) scale (Figure 3), we may study the amplitude
and polarization of these electrical pulses, which are different from each other at different measurement
points. The largest amplitude at location #1 (Figure 3a) is negative, but that recorded at the same time
is positive at location #3 (Figure 3c). The amplitude of the electrical signal at location #2 (Figure 3b) is
smaller than the other two signals in Figures 3a and 3c.
Changing the vertical (amplitude) scale (Figure 4), we can see more details of the potential before the
sample is broken. We see that the smaller sounds also generate electrical pulses, but do not cause a big change
in the basic potential levels with the smaller amplitude (about -0.005 V). Some low-frequency variation of
the potentials is also recorded.
3
•/˜™›šœ “˜Ÿž ˜¡ ” ¢J£¥¤œ —¡¤§¦¨
'/.+-1032 %+,+4657# ,98&# :
;<.=2 )>.+4@?=A
v
wy z
w
ˆ
{y z
ƒ ‡
†~
{
„
‰Š‹
u|{Jy z
€
ƒ
 ‚€
u
w
}~ 
"
# $&%('*)+%,+-
!
uxwGy z
uv
{y {Œ
{y {ŒD
EF H
¬
¼½¾
E
TL
N
O PN
KL M
¯
°·
BIJF H
ª
®
¶µ ±´
BE
²³
°±
BEGF H
BDC
®
µ³
¶¹ ¸
I
Q
­
º»
IF H
RS
{y {Dz
«
WXY
QO U
{y {D©
Ç*ÍGÌÏÎJËÐ
Ñ¥ÉÒŸÄ Ë>ÓÕÔÖÆ
b/gfih3j a+e+k6l7_ e>mn_ o
prq=j d>q+k@s+t
C
VM
{y {ŒDŽ
{y {Dv
‘ ’”“–•*—“˜™
IF IZ
IF IZ[
IF IZ\
IF I[C
^J_ `&acb/daef
IF I[]
ª­
ª¬
ª«
IF IH
¯¿ ¯ ¬
¯¿ ¯ ¬D«
¯¿ ¯ D
¯¿ ¯ «D­
¬ À
Ã‘Ä Å”ÆÈÇÊÉÆËÌ
¯¿ ¯ «DÁ
¯¿ ¯Â
Figure 3: The same signals as in Figure 2 but with different time scale (horizontal axis).
4
&.-*,0/21 $*+*3546" +87%" 9:;-<1 (=-*3?><@
‡ˆ‰
|„
 ~€ 6s tvu txtz
~ sWtvu tvy
{| }
s6tvu tvyGz
!" #%$'&)(*$
+*,
BC BB
F
OP
B
sŠtvuy
s6tu tx‹wxz
s6tvu txwz
Œ^ Ž’‘)“”
•
tu tywz
tu tz
À)ÅxÄqÆvÃ
ÇÈ Â
ɽ Ã=ÊÌË
Í
¿
¥¦ §
¥¦«¨;©
¸¹º
¥¦ ¨
¶·
±µ² ¯
´
¬³
¥¦ ¥©
¥
±² ­° W¤ ¥^¦ ¥g©
®¯ ¤Š¥^¦¨
¬­
¤W¥^¦)¨ª©
NL KM 6A BC BBF
K A6BC BE
HI J
A6BC BEGF
A6BC BD
sWtvu txw
BC BE
SJ
NLI RQ
_.cbedgf ]*aih5j6[ alk%[ mn;o<f `lo*hqp*r
B C B
D
BC BE5F
TUV
t
‚ƒ
tu ttz

tu ty
†}
‘.”
•—–˜ 
”
™šW ”œ› vž Ÿ˜ “œŸ™¢¡
£
t u tw
tu ty;z
AWBC)E
AXBC BYDF
AXBC BDF
Z[ \^]'_.`
]*a*b
BC BED
F
¤Š¥^¦ §
BC B
F
¤Š¥^¦¨
¤W¥v¦ ¥x»x§x©
¤W¥v¦ ¥g§x©
¼^½ ¾¿ÁÀ¿ÃÄ
¥¦ ¥¨;§©
¥¦ ¥©
Figure 4: The same signals as in Figure 2 but with different time scale (horizontal axis) and amplitude scale
(vertical axis).
5
Because granite contains 27.5% quarts, at each measurement point the group of quartz grains show a
piezoelectric characteristic on the average. The pressure applied on the rock sample generates a piezoelectric
field, whose amplitude and polarization depend on the piezoelectric characteristic of the local quartz group
and its orientation related to the direction of the pressure. This field is a local electrical potential, which does
not propagate in the rock. The sum of the electrical potential is very small due to the random distribution
of the quartz in a rock. The piezoelectric effect is the main mechanism that generates electrical signals when
a dry rock sample is broken.
4
Wet Granite Sample
The wet granite sample is saturated with tap water of 0.57 mS in conductivity in a vacuumized system. Figure
5 shows the electrical signals received on the water-saturated granite sample, and the acoustic waveform
received by the acoustic transducer (Figure 1). More details of the signals and acoustic waveform are
studied by changing the horizontal or vertical scales in Figures 6 and 7.
Ž,‘’”“• Œ‘–˜—Š ‘š™Š ›Gœž• šX– Ÿ¡
$,+(*.-0/ "()(132 )546 7
89+:/ &;+(1=<:>
p
qXs t
„…†
q
ƒz
€
{
uws t
|~ ‚
y
~
u
ovuGs t
| }{
o
q
xy z
!#"%$'&(")(*
orqDs t
op
ovuGsq
ouGs uD‡
BC E
¤
´µ¶
B
QI
K
L MK
HI J
±® °
§
¨¯
?FGC E
¢
¦
­® ©¬
?B
ª«
¨©
?BDC E
?A@
¦
­«
F
N
¥
²³
FC E
OP
uws q
£
TUV
NL R
uws uX‡
½'ÂDÁÄÃGÀÅ
ƞ¿Ç˜º À;ÈÊÉ˼
],a`cb0d \(_(e3fZ _;ghZ i
jlk:d ^;k(e=m(n
@
SJ
ouGs uˆp
uws uXp
‰#Š ‹ŒŽ'XŒX‘’
?FGC'B
?F
C FW
?F
C F@
FC FX@
YGZ [6\%],^\_`
FC FW
¢¥
¢¤
¢£
FC B
¢ §#· ¦
¢ §#· §ˆ¸
§· § ¥
¢ §#· § ¥
¹#º »¾
¼ ½¿X¼XÀXÁ
§· §A¸
§· ¦
Figure 5: The electric signals (a, b, and c) and the acoustic wave (d) recorded with a water-saturated granite
sample during its breaking. The displacement rate of the press is 10−2 mm/sec.
In this case, we record stronger electrical signals, which have almost the same amplitude, shape, and phase
at the three measurement points. This means that all of the electrical signals are induced by the same source
or rock breaking and received by electrodes at different locations. Compared with the dry sample, we know
that the mechanism of energy conversion in the water-saturated sample is different from the piezoelectric
effect in a dry sample. This conversion in a wet rock sample is the electrokinetic effect in nature. The
breaking generates an electromagnetic wave due to relative fluid flow in the wet rock. The electrical signal
generated by the breaking is a radiating electromagnetic wave, which can be received simultaneously at
different places.
6
–0™šœ› ”™žŸ’ ™¡ •’ ¢L£¥¤ ˜¡¤!ž§¦¨
(0/,.2143 &,-,5768$ -:9;$ <
=>/?3 *@/,5BA?C
w
x!z {
Š‹Œ
‡
|
†
|z {
„‚ ˆ
x
‰€
v}|Lz {

„
‚ ƒ
v
x
~ €
#
$ %'&)(+*,&-,.
!
vyxIz {
vw
"
|z |x
|z |x>
GH J
«
»¼½
G
VN
P
Q RP
MN O
®
¯¶
DKLH J
©
­
µ´ °³
DG
±²
¯°
DGIH J
DFE
­
´²
µ¸ ·
K
S
¬
¹º
KH J
TU
|z |F
ª
YZ[
SQ W
|z |!wF
Å+ËIÊÍÌLÉÎ
Ï¥ÇРÉ@ÑÓÒÔÄ
c0gfih4j b,e,k7l8` e@mn` o
prq?j d@q,kBs,t
E
XO
|z |x>Ž
|z |Fw!w
‘'’ “•”—–+˜!”!™š
KH KG
KH KG\
KH KG]
KH K!E!E
_L` a;b)c0dbef
KH K!E!^
©¬
©«
©ª
KH K!u
®¾ ® ­
®¾ ® ­>ª
®¾ ® >­ ¿
®¾ ® ¬F¬
Á' ÕÄÆÅÈÇ!Ä!É!Ê
®¾ ® ¬FÀ
®¾ ® «
Figure 6: The same signals as in Figure 5 but with different time scale (horizontal axis).
7
Changing the amplitude scale, we can see more details of the potential variation before rock breaking
occurs. Before the rock sample is broken completely, we record some sounds with small or smaller amplitudes
between -0.04 second and 0.01 second (Figure 7d). The larger ones generate electrical pulses (Figures 7a, 7b,
and 7c). Since recording the small sounds around -0.04 second, the basic electrical level (the dc potential)
rose (Figures 7a and 7b) or dropped (Figure 7c). The amplitudes and polarizations of the potentials are
different from each other. This means that the potentials due to the piezoelectric effect depend not only
on the pressure, but also on the piezoelectric characteristic and orientation of the quartz group around the
measurement points. They are local electrical potentials, and the potential and its variation can be observed
before the rock breaking. Because the stress concentrates around the breaking area, we can record the local
potentials around the area. When a fluid-saturated rock sample is breaking, the moving charges in the
fluid induce electromagnetic waves, which propagate independently and can be received near or far from
the breaking area. Their frequencies are usually much higher than the natural potential and close to those
of acoustic waves. This phenomenon is very similar to that observed in the Loma Prieta earthquake of 17
October 1989 (Fraser-Smith et al., 1990).
&.-*,0/21 $*+*3546" +879" :;-<1 (=-*3?><@
SJ
LN R
IQ
OP
!" #%$'&)(*$+*,
~
q
pWqXr qts
_.cbed2f ^*a*g5h6\ a=ij\ klGm<f `=m*g?n*o
­±® «
°
¡¢ ¡¥
¡
­® ©¬
W¡%¢ ¡§¥
¨¯
ª«
¨©
A6BC BDF
[X\ ]9^'_.`^ab
B C B EDF
p6qXr qtsv
ˆ%‰ ŠŒ‹Ž)‹‘
qr qusv
qr qv
½)ÂtÁÄÃXÀÅƜ¿Çº À=ÈÊÉ˼
¡¢ ¤
²³
AYBC BZDF
p6qr qt‡stv
¡¢·¤¥
´µ¶
AWBXC)E
p†qXru
¡¢ £
B
C B E
B C BBF
B
NL KM 6A BC BBF
K A6BC BE
HI J
A6BC BEGF
A6BC BD
qr qqv
}{ z| 6p qXr qtqv
z pWqXr qXu
wx y
p6qXr qXuGv
B
C BD
B C B
EF
TUV
qr qu
‚y
}{x €
qr quv
ƒ„
.‘“’
” ‹•–W‰ ˜—Œ‰ ™Xšœ›
” ˜›•žŸ
qr qs
†¡%¢¤
W¡%¢)¤¦¥
†¡%¢ £
B C BF
†¡%¢¤
W¡X¢ ¡t¸t£t¥
W¡X¢ ¡§£t¥
¹%º »Œ¼¾½¿¼ÀÁ
¡¢ ¡¤£¥
¡¢ ¡¥
Figure 7: The same signals as in Figure 5 but with different time scale (horizontal axis) and amplitude scale
(vertical axis).
To investigate the seismoelectric field generated by rock breaking, we change the rate (displacement
speed) of the press. In the previous experiments with dry and wet samples, the displacement speed is
10−2 mm/sec. Figure 8 shows the electrical and acoustic signals with a water-saturated Westerly granite
sample at the displacement speed of 10−4 mm/sec before the rock breaking. In this case, we have only
observed the potential variation in DC or low-frequency ranges, we did not record any strong acoustic waves
or electric signals. There was no loud sound during rock breaking (Figure 8d), and only few electric pulses
were recorded in Figures 8a, 8b, and 8c, but we observed the continuous variation of the natural potential.
The amplitude and polarization of the variation are different at the three measurement points. The potentials
go up in Figure 8a and 8b, but go down in Figure 8c. In this case the pressure makes more microcracks
8
within the rock. Then the pore fluid flows into the cracks and changes the natural potential, which is related
to the adsorption potential between the fluid and solid of grains.
$,+(*.-0/ "()(1324 )65# 78+9/ &:+(1<;9=
RI
KM Q
HP
NO
|zw €
p
|z y{ 4o pWq pspu
y oVpWq pWt
vw x
o4pWq pWtEu
oVpWq psr
}~
!#"%$'&(")(*
],badc0e \(`(f3g4Z `:hiZ jkEl9e _:l(f<m(n
®²¯ ¬
¢C£ ¢¦
®¯ ª­
¡V¢Š£ ¢‡¦
±
«¬
©ª
?@ ?X
pFq p†
pFq t
¢C£ ¢¸
¢C£ ¥
½'ÂsÁÄÃWÀÅƝ¿Çº À:ÈÊÉ˼
¢
©°
>V?@ ?A
?@ ?A
YWZ [#\^],_\`a
oVpWq p‡r
pFq pr
ˆŠ‰ ‹ŒŽ'Œ‘’
¢C£ ¥
³´
>V?@ ?X
oVpWq ps†
¢C£¨¥¦
µ¶·
>V?W@'B
o…pWqt
¢C£ ¤
? @ ?FB
?F@ ??D
?
MK JL 4> ?@ ??D
J >4?@ ?B
GH I
>4?@ ?CBED
>4?@ ?A
pCq pCt
pCq ppu
x
? @ ?A
?F@ ?BD
STU
pCq pCtu
‚ƒ„
Ž,‘’”“• Œ‘–—V‰ ‘™˜‰ šW›œ• ™œ–Ÿž pCq pr
¡…¢Š£¥
¡V¢Š£'¥§¦
¡…¢Š£ ¤
?@ B
¡…¢Š£¥
¡V¢Š£ ¢‡¸
¡V¢Š£ ¢‡¤
¢C£ ¢¤
¹Šº »¼¾½¿¼ÀÁ
Figure 8: The electric signals (a, b, and c) and the acoustic wave (d) recorded with a water-saturated granite
sample during its breaking. The displacement rate is 10−4 mm/sec, which is lower than the previous rate of
10−2 mm/sec.
5
Conclusions
In this paper, we performed experiments with dry and water-saturated granite samples to measure electric
and acoustic signals when the samples were pressured and broken. The experimental results show that the
electrical signals recorded during rock breaking are generated by two different mechanisms of the piezoelectric
effect and seismoelectric conversion. The results help us to understand the whole electric and acoustic
procedures due to the piezoelectric effect and seismoelectric conversion during rock breaking.
The experiment with a dry rock sample shows that the amplitude and the polarization of the electrical potentials due to the piezoelectric effect depend not only on pressure, but also on the piezoelectric characteristic
and the orientation of the grains around the measurement points.
Rock breaking generates a relative motion between the fluid and the solid matrix in a fluid-saturated
rock. Moving charges in the fluid induce electromagnetic waves in the rock. If the breaking is big enough,
it generates a seismoelectric pulse, which is a radiating electromagnetic wave and can be simultaneously
received at different places. Some pore fluid flows into the new microcracks made with smaller breaking and
establishes new balance, which also changes the natural potential in the rock.
6
Acknowledgements
We would like to thank Dr. Steve Karner and Dr. Karen Mair for their useful discussion and assistance in
control of the press system.
9
References
W. F. Brace and A. S. Orange. Electric resistivity changes in saturated rocks during fracture and frictional
sliding. J. of Geophysical Research, 1968.
A. C. Fraser-Smith, A. Bernardi, P. R. McGill, M. E. Ladd, R. A. Helliwell, and O. G. Jr. Villard. Lowfrequency magnetic field measurements near the epicenter of the ms7.1 loma prieta earthquake. Geophysical
Research Letters, 1990.
S. K. Park, M. J. S. Johnston, T. R. Madde, D. F. Morgan, and H. F. Morrison. Electromagnetic precursors
to earthquake in the ulf band: A review of observation and mechanisms. Reviews of Geophysics, 1993.
T. F. Wong, J. T. Fredrich, and G. D. Gwanmesia. Crack aperture statistics and pore space fractal geometry
of westerly granite and rutland quartzite: Implication for an elastic contact model of rock compressibility.
J. of Geophysical Research, 1989.
Z. Zhu, M. W. Haartsen, and M. N. Toksöz. Experimental studies of electrokinetic conversions in fluidsaturated borehole models. Geophysics, 1999.
Z. Zhu, M. W. Haartsen, and M. N. Toksöz. Experimental studies of seismoelectric conversions in fluidsaturated porous media. J. of Geophysical Research, 2000.
10