Earthquake lights and the stress-activation of positive hole charge carriers in rocks

Abstract

Earthquake-related luminous phenomena (also known as earthquake lights) may arise from (1) the stress-activation of positive hole (p-hole) charge carriers in igneous rocks and (2) the accumulation of high charge carrier concentrations at asperities in the crust where the stress rates increase very rapidly as an earthquake approaches. It is proposed that, when a critical charge carrier concentration is reached, the p-holes form a degenerated solid state plasma that can break out of the confined rock volume and propagate as a rapidly expanding charge cloud. Upon reaching the surface the charge cloud causes dielectric breakdown at the air–rock interface, i.e. corona discharges, accompanied by the emission of light and high frequency electromagnetic radiation.

Full text

Earthqua ke Ligh ts and th e Stress-Activation of Positive Hole
Charge C arriers in Rocks
France St-Laurent1, John S. Derr2, and Friedemann T. Freund3,4
1 125, 68e Avenue, LaSalle, Qc H8P 3H1, Canada, france.st-laurent@csmb.qc.ca
2 US Geological Survey, Albuquerque Seismological Laboratory, Albuquerque, NM 87198-2010,
USA, derr@usgs.gov
3 Department of Physics, San Jose State University, San Jose, CA 95192-0106, USA.
4 NASA Goddard Space Flight Center, Planetary Geodynamics Laboratory, Greenbelt, MD 20771,
USA : ffreund@mail.arc.nasa.gov
Abstract
Earthquake-related luminous phenomena (also known as earthquake lights) may arise from (1) the
stress-activation of positive hole (p-hole) charge carriers in igneous rocks and (2) the accumulation
of high charge carrier concentrations at asperities in the crust where the stress rates increase very
rapidly as an earthquake approaches. When a critical charge carrier concentration is reached, the
p-holes form a degenerated solid state plasma that can break out of the confined rock volume and
propagate as a rapidly expanding charge cloud. Upon reaching the surface the charge cloud causes
dielectric breakdown at the air-rock interface, i.e., corona discharges, accompanied by the emission
of light and high frequency electromagnetic radiation.
Earthquake Lights
Earthquake-related luminous phenomena, also known as earthquake lights, EQLs, have been
reported since ancient times [Derr, 1973; Tributsch, 1983]. In 1931, based on over 1500 reports
from several various events in Japan, Musya [Musya, 1931] stated: “The observations were so
abundant and so carefully made that we can no longer feel much doubt as to the reality of the
phenomena.” Nonetheless, doubts persisted in the scientific community at least until the late 1960s
when EQLs were photographically documented during an earthquake swarm near Matsushiro,
Japan. Yasui, a seismologist at the Kakioka Magnetic Observatory, studied reports from many
people in the surrounding area, including sketches and photographs [Yasui, 1973], and concluded
that most of the observations cannot be accounted for by atmospheric lightning, zodiacal light,
auroras, meteors or by any other known sources. Similar observations were made in Mexico
[Araiza-Quijano and Hernández-del-Valle, 1996] and in many other seismically active
regions of the world [Lomnitz, 1994].

St-Laurent et al.: Earthquake Lights
2
Figure 1: Sketch of a luminous phenomenon as reported by J.A. Dallaire.
St-Laurent [St-Laurent, 2000] critically evaluated numerous reports of EQLs associated with the
M=6.5 mbLg Saguenay earthquake, Québec Province, Canada, on 25 Nov. 1988, which occurred
during darkness, at 18:46 local time. The reports confirm the diversity of the observed luminous
phenomena. One report that is particularly well supported provides insight into the processes that
seem to have taken place in the Saguenay region, close to the 29 km deep hypocenter.
The earthquake was associated with the Saguenay graben, which runs roughly SE-NW and is
nearly perpendicular to the St. Lawrence river, meeting it about 150 km northeast of Québec city.
The Saguenay graben south wall delineates the northeastern edge of the Jacques Cartier block, a
horst structure in the 600-900 m high Laurentian Plateau and exceeding it by 100-400 m.
The Saguenay earthquake was remarkable in several respects. First, it occurred outside the
previously recognized seismic zones in this part of the Precambrian craton, which forms the
Canadian Shield. Second, with a focal depth of 29 km it is one of the deepest intracontinental
crustal earthquakes outside of plate convergence zones. Third, the high Lg-wave energy led to
prolonged (2 min) and strong shaking in the epicentral region. The relatively low aftershock activity
during the next six months (only 84 events, all smaller than M=3.6 except for one M=4.1) implies
that a large fraction of the energy stored before the earthquake was released during the main shock.
There is, however, yet another reason for calling the Saguenay earthquake remarkable: A large
number of luminous phenomena. reported from a wide area around the epicenter to the INRS,
Université du Québec at Sainte-Foy, and to the Université du Québec at Chicoutimi [Ouellet, 1990].
The earliest report dates from 25 days before the 25 November 1988 main shock when two
individuals driving on Route 175 through the Laurentide Park observed at about 18:30 local time
three luminous “masses” that rose from the ground. The area was 15 km from the nearest
settlement, but close to the future epicenter. More luminous phenomena were reported during the
mb=4.3 (mbLg =4.8) foreshock of 23 November 1988, at the time of the 25 November 1988 main
shock and during aftershocks. The phenomena were variously described as globular luminous
masses, bands or rays or as intense atmospheric illuminations lasting from several seconds to
several minutes. The phenomena were reported from places as far as 205 km from the epicenter,
though the majority fell within a 50 km radius. All cases of intense atmospheric illumination were

St-Laurent et al.: Earthquake Lights
3
reported from within a radius of 35 km of the epicenter and most were coseismic or nearly
coseismic with the mb=4.3 foreshock of 23 November 1988 [St-Laurent, 2000].
The Ouellet compilation [Ouellet, 1990] contains 52 reports, of which 46 were judged to be of
sufficient quality to warrant further study. Several of these reports were selected for follow-up
contacts with the individuals who had made the observations and for site visits [St-Laurent, 2000].
We shall focus on one particularly well-documented observation made 19 km north of the
epicenter, almost coseismic with the main shock. It was reported by Joseph A. Dallaire, a trapper
who lives at Laterrière, close to the town of Chicoutimi. On Nov. 25, 1988, Dallaire returned from
the nearby forest, where he had inspected his traps. The forest consisted mostly of conifers with a
few birches, which had shed their leaves. The wind stood at 5 km/hr and the air temperature
hovered around –8°C with a relative humidity of 65%. A few patches of a thin layer of icy snow
that had been put down four days earlier remained on the ground. It was before 7 o’clock at night,
about 2 h 45 min after sunset, about 1 hr after moonrise at phase 0.9, with mostly clear sky. Just as
Dallaire emerged from the forest, looking NW in the direction of his house across an open field 700
m away, he was startled by a crackling sound approaching fast from behind him.
As soon as the crackling noise had reached Dallaire, he saw a curtain or sheet of bluish light
emerging from the forest to his left and to his right. The light was hugging the ground as it moved
past him, passed into the open field, and disappeared in a general North-West direction. Dallaire
estimated that the sheet of light traveled the distance from the edge of the woods to his house, 700
m, in about 2 seconds. He reported that the light was bright enough to illuminate his house. As it
passed his house, it exceeded the height of its roof, 6 m, and may have been as high as 15 m. At the
moment the sheet of light faded and disappeared, he felt the earthquake.
The description contains important details. (i) The event started with a bristling or crackling
noise approaching from the direction of the epicenter 19 km to the left as shown in Figure 1. The
noise appears to have been caused by electric discharges off the conifer branches, indicating the
buildup of a strong electric field. (ii) The curtain or sheet of light suggests an even stronger electric
field that led to a discharge at the ground-to-air interface. (iii) The discharge was traveling from the
direction of the epicenter. (iv) The speed of propagation was in the order of 100-300 m/sec. (v) The
electric discharge and EQL were not co-seismic but clearly preceded the seismic wave train.
Tsukuda [Tsukuda, 1997] reported luminous phenomena observed before and during the M=7.2
Jan. 17, 1995 Hyogo-ken (Kobe) earthquake, which occurred in early morning darkness at 5:46
local time. The light spread rapidly to several kilometers in width and was estimated to reach up to
200 m in height with intensities estimated at 103 candela/m2: "According to most eyewitnesses the
luminosity started from ground level on land, suggesting that discharge processes… in near-surface
rocks may be the primary driving force”.
While there are many observations from many independent sources, a great uncertainty remains
about the physics that underlies the emission of light from the Earth’s surface before or during
major earthquakes or during aftershocks. Most efforts to provide a physical explanation of the
phenomenon centered on piezoelectricity, a property of quartz to generate electric fields on opposite
sides of a single crystal when stressed in certain crystallographic directions [Bishop, 1981;
Finkelstein, et al., 1973]. However, when a rock containing quartz is stressed, the electric fields
generated by a large number of individual crystals cancel and the resulting field becomes zero. This
is true for random orientations of the quartz crystals as well as for preferentially aligned quartz

St-Laurent et al.: Earthquake Lights
4
crystals1. Other processes like tribo- or fracture electrification [Yamada, et al., 1989] or exoelectron
emission [Enomoto, et al., 1993] seem inadequate to produce the large electric fields needed on
large scales to account for at least the more powerful and sustained EQLs that have been reported.
Dormant Electronic Charge Carriers in Rocks
Igneous rocks contain charge carriers, the very existence of which has been overlooked in the
past. As described in more detail elsewhere in this Special Issue, the charge carriers are electronic in
nature and seem to be ubiquitous in igneous rocks. Most importantly in the context of EQLs they
can be activated by stress.
Figure 2 shows schematically what happens when we load one end of a 1.2 m long granite slab
fitted with Cu electrodes at both ends. Two types of charge carriers appear inside the stressed rock
volume on the left: holes and electrons. As indicated by the arrow inside the rock the holes flow
from left to right through the unstressed portion of the granite slab. The electrons are unable to flow
through the rock. Instead they flow from the source S into the Cu electrode on the left and thence to
ground. The electric circuit closes by the electrons flowing through the external wire as indicated by
the thin arrow and meeting the holes at the right-hand end of the granite slab.
A
A
+ + + + + + + +
+ + + + + + + +
+ + + + + + + +
+ + + + + + + +
+ + + + + + + +
+ + + + + + + +
V
h•
e'
e'
S
Figure 2: Two self-generated currents flow out of the stressed rock volume, the source S, a hole
current (h•) and an electron current (e’), while a positive surface potential builds up. The boundary
between stressed and unstressed rock is indicated by the symbol for a diode.
The outflow of two currents from the stressed rock volume in opposite directions indicates that
the boundary between stressed and unstressed rock acts as a diode. The diode lets the holes pass but
blocks the electrons. Note that the two currents flow without externally applied voltage. They are
self-generated. The driving force is provided by the stress gradient. As the holes flow out from the
stressed rock volume into the unstressed rock, they set up an electric field. This electric field pulls
the electrons out of the stressed rock volume and causes them to flow through the external circuit to
meet the holes at the right end of the slab. This electric field tightly couples the two currents and
forces them to synchronize [Freund, et al., 2005].
Next we discuss the nature of the charge carriers involved in the generation of these currents.
1 Though in texturally aligned rocks the polar axes of the quartz crystals may show a preferred orientation,
there is no mechanism in nature that could select their individual +– directions over their –+ directions and
thereby create conditions to produce, upon stressing, a net piezoelectric effect.

St-Laurent et al.: Earthquake Lights
5
The holes are defect electrons in the O 2p-dominated valence band of the silicate minerals, i.e.
that consist of electronic states that can best be described as a change of the valence of oxygen
anions from their usual 2- to the 1- state. They are equivalent to O- in an O2-matrix. We call them
“positive holes” or p-holes for short, symbolized by h•. The presence of these charge carriers in
(otherwise insulating) oxide and silicate materials has been established through studies involving
single crystals and rocks summarized in [Freund, 2003]. Normally the p-holes are inactive, i.e.
dormant in the form of positive hole pairs, PHPs, chemically equivalent to peroxy links of the type
O3Si–OO–SiO3 [Freund, 2002]. When rocks are placed under stress, mineral grains begin to
plastically deform. This deformation causes dislocations to move and new dislocations to be
generated. The moving dislocations intersect the peroxy links and cause them to momentarily break.
The higher the rate of deformation, the more dislocations appear and the more p-holes are activated
per unit rock volume. We can represent this generation process as a two-step process:
O3Si–OO–SiO3 = O3Si–O••O–SiO3 = [O3Si–OO–SiO3]’ + h•
PHP broken PHP loosely bound electron mobile p-hole
where the dots • signify O- states of the broken peroxy link, the superscript ’ an electron that has
moved in, and h• a p-hole that has moved out.
The electrons are co-activated alongside with the p-holes. They are loosely bound and,
hence, mobile. A stressed rock volume becomes a “source” which can release both p-holes and
electrons.
The p-holes have the unusual property that, being electronic states in the valence band, they
are able to spread out of the source into the surrounding unstressed rock. To travel they use the O
2p-dominated upper edge of the valence band, which provides for a small but finite p-type
conductivity. The p-holes can therefore propagate through unstressed rock and even cross different
rock types [Freund, 2002].
The electrons are unable to flow from the stressed rock into the unstressed rock. The reason
is that they need an n-type conducting pathway. In the experiment sketched in Figure 2 we
artificially provided this n-type connectivity by applying one of the Cu electrodes directly to the
source volume. In nature, in the Earth’s crust, n-type conductivity becomes available at high
temperatures, viz. at greater depth along the geotherm.
Geophysical Scenario
Figure 3a-c translates the geometry of the laboratory experiment shown in Figure 2 into a
geophysical scenario. (a) Assume a slab of brittle rocks spanning the thickness of the crust from the
surface to the hot and ductile mid- to lower crust. The slab is pushed by tectonic forces from the left
to the right. (b) As stresses build up, p-holes and electrons are activated. The p-holes can flow out
from the source volume laterally through the p-type conductive cool upper portions of the crust. The
electrons can leave the source volume only by connecting downward to the n-type conductive,
hotter regions of the mid- to lower crust. The two outflow currents are balanced as indicated by the
solid arrows. (c) A situation ensues where the two outflow currents can no longer keep up with the
production rate of p-holes and electrons in the source volume, symbolized by dashed arrows.
By comparing the laboratory set-up with the geophysical scenario we can elaborate on some
details of the charge generation process. The outflow of p-holes produces an electric field, which
couples to the outflow of the electrons in the opposite direction. This means that the two outflow
currents must be tightly coupled. The system can go out of balance if the coupling breaks down.

St-Laurent et al.: Earthquake Lights
6
One possible cause for imbalance could be that, when the outflow currents cannot keep up with
the production rate of the p-holes and electrons. This will happen when, due to rapid stress increase,
the concentration of charge carriers in a confined rock volume shoots up rapidly. It is conceivable
that the p-holes reach a concentration at which they begin to form a degenerate charge cloud. Such a
charge cloud can become unstable and break out of its confined volume as a solid state plasma
propagating outward explosively. When the plasma reaches the surface, it causes instant ionization
of the air, a corona discharge and, hence, light emission. This scenario is sketched in Figure 3c.
Increasing temperature
Increasing ductility
Increasing electrical conductivity
Brittle and weak
Brittle and strong
Surface
km
0
10
20
30
40
50
60
Push
Figure 3(a)
Simplified model of a crustal block
25 km thick, being pushed by
tectonic forces.
h•
e'/h•
e'
Push
Figure 3(b)
Electronic charge carriers, p-holes
and electrons are activated in the
stressed rock volume. The boundary
between stressed and unstressed
rocks act as a diode, allowing p-holes
to flow out laterally but blocking
electrons. The electrons have to flow
downward into the n-type conducting
lower crust.
h•
e'
e'/h•
Push
Figure 3(c)
If, due to very rapid increase of the
stresses, more of the charge carriers
are activated than can be dissipated
(dashed arrows), a solid state plasma
can form that breaks out of its
confinement and bursts through the
Earth’s surface, causing an electric
discharge and attendant luminous
phenomena.

St-Laurent et al.: Earthquake Lights
7
Electrical Discharges in Laboratory Experiments
In fact there are already laboratory observations that suggest some kind of solid state plasma
formed in highly stressed rock volumes by p-hole charge carriers and their burst-like expansion,
though these observations had not analyzed in this physical context. An example is shown in
Figure 4a. We took a granite plate, approximately 30 x 20 x 2 cm3, and loaded one inner portion of
it at a constant stress rate up to failure. During loading but before failure several cracks formed
inside the stressed rock volume. We recorded their acoustic signals with a microphone placed about
5 cm from the edge of the piston. In addition, as sketched in the inset in Figure 4a, we had installed
a capacitive sensor placed about 20 cm from the edge of the piston to record the surface potential.
Several cracks spread over several minutes before failure of the rock triggered the data
acquisition system. During each crack we recorded at the location of the capacitive sensor a burst of
positive voltage, lasting less than 50 µsec and followed by a longer lasting negative voltage.
However, as shown in Figure 4a for crack #11, the burst-like positive signal arrived about 1 msec
before the microphone recorded the acoustic signal of the crack itself. The amplitude of the voltage
burst, +3 V but occasionally up to +12 V, was significantly higher than the steady state surface
potentials, less than +100 mV, observed during loading [Freund, et al., 2004]. Takeuchi and
Nagahama [Takeuchi and Nagahama, 2001] observed similar positive voltage signals during strike-
slip faulting experiments and Enomoto reported pulses up to +17 V [Enomoto, et al., 1993]. The
positive sign, high amplitude and shortness of the signal are all consistent with a p-hole charge
cloud that expanded rapidly from inside the severely stressed rock volume, presumably from the
microvolume which was about to crack. From this we can infer that the rate of generation of the
charge carriers in the stressed rock volume must reach its maximum shortly before fracture.
Why this is so has to do with the mechanism of deformation of mineral grains. Deformation
occurs via the generation of dislocations. As illustrated schematically from left to right in Figure 4b
the density of dislocations increases with increasing stress. Eventually, the number of dislocations
per unit volume becomes so large, on the order of 1010-1012 cm/cm3, that saturation is reached.
From this point onward few new dislocations are formed and most existing ones begin to entangle.
They coalesce to form microcracks, which rapidly evolve into larger cracks and eventually into
fractures. If dislocation movement is the mechanism by which charge carriers are activated as
suggested above, the rate at which p-holes are generated must reach a maximum before fracture.
We can take it one step further and look at the timing of the positive voltage burst relative to
the fracture event. From impact experiments we know that p-hole charge clouds propagate at speeds
between 100-300 m/sec [Freund, 2002]. These speeds are consistent with the concept of p-holes
propagating via electrons hopping in the opposite directions from O2- to O2- sites at the frequency of
the lattice phonons. The jump distance is on the order of 3 Å (3 x 10-10 m). The phonon frequency is
on the order of 1012 sec-1. Hence, 3 x 10-10 x 1012 = 300 m/sec. Acoustic signals propagate through
granite at 6 km/sec for compressional (P) waves and at 3.4 km/sec for transverse (S) waves. Hence,
the acoustic signal of the crack event would reach the microphone in about 10 µsec, while the
voltage pulse takes 100 times longer, about 1 msec, to reach the capacitive sensor at the location
shown in Figure 4a. Since the voltage pulse arrived 1 msec before the acoustic signal, it must have
been generated at its starting point, in the aforementioned microvolume, 2 msec before the crack2.
2 If the voltage pulse were not due to a propagating charge cloud but to the electric field generated by
piezoelectric quartz crystals, its speed of propagation would be the speed of light divided by the dielectric
constant, ~50,000 km/sec. In this case, given the limited time resolution of the experiment, the arrival time of
the voltage pulse would be coincident with the acoustic signal.

St-Laurent et al.: Earthquake Lights
8
-1
0
1
2
3
4
-0.050
0.0
0.050
0.10
0.15
-0.002 0 0.002 0.004 0.006 0.008
Surface Potential [V]
Acoustic [V]
Time [sec]
V
V
Capacitor
Microphone
Piston
Ground
Insulation
Insulation
Piston
Granite
Run #24
Crack #11
Dislocations
Dislocation
Entanglement
Dislocation
Coalescence
Figure 4a: Surface potential and acoustic signal
recorded during crack formation in a granite
plate placed under load. Note the early arrival
of the voltage pulse [Freund, et al., 2004].
Figure 4b: Schematic representation of the
number of dislocations generated during plastic
deformation, their entanglement and eventual
coalescence into microcracks, initiating fracture.
During the experiment depicted in Figure 4a we did not simultaneously measure the light
emission or the emission of kHz electromagnetic (EM) radiation that is characteristic of electric
discharges. However, the shape of the positive voltage burst and its subsequent broader negative
voltage are very similar to the voltage pulses recorded during low-speed (100 m/sec) impact
experiments. In these experiments a positive charge cloud was generated by the sudden stress to
rock volume around the impact point. The charge cloud spread through the rock samples and
arrived at the surface, causing the voltage signal to increase. The increase of the positive surface
potential was interrupted above 400 mV by a sudden emission of light. The light came from the
sharp edges of the rock where the electric field is highest. At the same time a burst of 10-20 kHz
EM radiation was recorded. Light emission and a burst of radiofrequency radiation are both clear
indicators of a luminous corona discharge.
Taking these various observations into account we can confidently say that the +3 V voltage
burst recorded about 1 msec before cracking and shown in Figure 4a must have been accompanied
by a corona discharge due to the high electric field that built up at the rock surface upon arrival of
the p-hole charge cloud. Hence, the experiment depicted in Figure 4a must have produced a small,
artificial EQL.

St-Laurent et al.: Earthquake Lights
9
Saguenay Earthquake Light Observations
We now return to the field observations, specifically to the reports on EQLs associated with
the Saguenay earthquake [St-Laurent, 2000]. The wide distribution of reported EQL sightings
during the time leading up to the Saguenay event and the fact that the earliest report came 25 days
before the main shock indicates that the entire region was building up stress to a critical level.
Intersecting faults, buried plutons, rift pillow and other deep structures are capable of localizing
stresses [Zoback and Richardson, 1996] and of acting as “stress concentrators” [Gangopadhyay and
Talwani, 2003]. According to Du Berger [Du Berger, et al., 1991] “plutons in the region span an
area of few hundred square kilometers and extend to middle and lower crustal depths and are mostly
composed of granite, mangerite, mafic dikes and gabbro”. Small volumes within this crystalline
basement presumably became critical in the sense that, by accumulating localized stresses, they
produced p-hole concentrations high enough to initiate an outburst of a charge cloud. The outbursts
led to electric discharges at the Earth’s surface but were not necessarily accompanied by foreshocks.
In some ways these local stress concentrators behaved like the microvolumes in our rock
deformation experiment depicted in Figure 4a. These microvolumes became critical one after the
other and cracked, while the overall stress increased. In the case of the Saguenay event, each of the
local stress concentrators released stresses at depth or transferred them laterally onto the remaining
asperity, which eventually broke catastrophically during the main shock.
Figure 5 (top) shows a vertically exaggerated cross section of the Saguenay graben, ~40 km
wide. Indicated from right to left are the North Wall, the Saguenay River, the location of the
observer (J.A. Dallaire), the South Wall, and the epicenter. The Saguenay earthquake was initiated
at a depth of 29 km near the south rim of the Graben [Roy, et al., 1993]. Figure 5 (bottom) depicts a
section through the crust with the hypocenter marked by a star where t1 and t2 are, respectively the
epicenter and the location of the observer J.A. Dallaire at a distance of 19 km. The hypocenter is
shown as a source of a charge cloud expanding at ~300 m/sec.
Figure 5 top: Schematic, vertically exaggerated cross section through the Saguenay graben;
bottom: Hypocenter at 29 km depth and outburst of a charge cloud assumed to travel at ~300 m/sec.

St-Laurent et al.: Earthquake Lights
10
We assume that the hypocenter was the last asperity where the stresses accumulated and
reached their highest values shortly before the main rupture. The volume of this asperity can be
considered analogous to the microvolume in the stressed laboratory rock sample shortly before a
crack forms. As in the case of the laboratory experiment, when stresses build up, dislocations are
generated in increasing numbers on the scale of the individual mineral grains, thereby activating an
ever larger number of p-holes and electrons. Assuming a constant rate of deformation, the stresses
increase very rapidly as the system moves closer to catastrophic failure. The consequences will be
that the rate at which p-holes and electrons are generated will eventually exceed the rate at which
the charge carriers can be dissipated. It is therefore conceivable that, shortly before rupture, a cloud
of charge carriers, presumably p-holes, will burst out of the confined volume of the asperity and
expand outward as depicted in Figure 5 (bottom). On intersecting the surface above, this massive
charge cloud will induce a large electric discharge at the air-rock interface and a luminous corona
discharge similar to the corona discharges produced during the impact experiments [Freund, 2002].
Assuming a speed of propagation of the charge cloud in the range of 100-300 m/sec and
knowing from Dallaire’s report that the rapidly moving curtain of light arrived at his location before
the first seismic wave train, we can estimate the time at which this outburst must have started at the
asperity which we identify with the hypocenter. The distance from the hypocenter to Dallaire’s
location is ~36 km. The P waves take ~6 sec to travel this far at 6 km/sec. A p-hole charge cloud
travelling at 300 m/sec would cover the same distance in 120 sec. Hence, in order to arrive at
Dallaire’s location before the fastest seismic wave train, the charge cloud must have burst out of its
confinement near the hypocenter 130 sec before the rupture. This implies that the maximum of the
charge carrier generation process in the source volume was reached about 130 sec before rupture.
The estimated 2 sec which the light curtain needed to traverse the 700 m wide open field in
Dallaire’s view agree with a speed of p-hole propagation on the order of 300 m/sec. The relatively
low speed of propagation of the charge cloud is echoed in many other reports that the luminous
flashes associated with earthquakes are short-lived but longer-lasting than lightning strikes
[Enomoto and Zheng, 1998; Tsukuda, 1997; Yoshida, et al., 1995].
Conclusions
On the basis of the foregoing discussion we propose that the luminous phenomena
associated with earthquakes, often called earthquake lights, EQL, are caused by electric discharges.
The source of these discharges lies in the Earth’s crust, in confined rock volumes that represent
asperities and build up high and rapidly increasing stresses as part of the earthquake preparation
process. Such stresses activate electronic charge carriers that lie dormant in the rocks. These charge
carriers are p-holes and electrons, of which the p-holes have the unique property that they can
propagate through otherwise insulating rocks.
If the rate at which the p-holes and electrons are activated exceeds the rate at which they can
be dissipated, a situation may arise where the p-holes form a degenerate solid state plasma that can
burst out of its confined rock volume and propagate at relatively high speed through the overlying
rocks. When this charge cloud intersects the Earth’s surface, it causes ionization of the air and,
hence, corona discharges, which are accompanied by the emission of light. The many different
forms and shapes of EQLs that have been reported suggest that the conditions of the solid state
plasma and its discharge through the Earth’s surface can be highly variable.
Our conclusions seem to be consistent with not only the observed luminous phenomena but
also with the reported emission of radio-frequency electromagnetic radiation and other effects.

St-Laurent et al.: Earthquake Lights
11
References
Araiza-Quijano, M. R., and G. Hernández-del-Valle (1996), Some observations of atmospheric
luminosity as a possible earthquake precursor Geofisica Internacional, 35, 403-408.
Bishop, J. R. (1981), Piezoelectric effects in quartz-rich rocks, Tectonophys., 77, 297-321.
Derr, J. S. (1973), Earthquake lights: a review of observations and present theories, Bull. Seismol.
Soc. Amer., 63, 2177-21287.
Du Berger, R., et al. (1991), The Saguenay (Quebec) earthquake of November 25, 1988:
Seismological data and geological setting, Tectonophys., 186, 59-74.
Enomoto, Y., et al. (1993), Exoelectron emission: Possible relation to seismic geo-electromagnetic
activities as a microscopic aspect in geotribology, Wear, 168, 135-142.
Enomoto, Y., and Z. Zheng (1998), Possible evidences of earthquake lightning accompanying the
1995 Kobe earthquake inferred from the Nojima fault gouge, Geophysical Research Letters, 25,
2721-.
Finkelstein, D., et al. (1973), The piezoelectric theory of earthquake lightning, J. Geophys. Res., 78,
992-993.
Freund, F. (2002), Charge generation and propagation in rocks, J. Geodynamics, 33, 545-572.
Freund, F. T. (2003), On the electrical conductivity structure of the stable continental crust, J.
Geodynamics, 35, 353-388.
Freund, F. T., et al. (2004), Stress-induced changes in the electrical conductivity of igneous rocks
and the generation of ground currents, Terrestrial, Atmospheric and Oceanic Sciences (TAO), 15,
437-468.
Freund, F. T., et al. (2005), Self-Generated Electric Currents from Igneous Rocks under Stress,
Science, in review.
Gangopadhyay, A., and P. Talwani (2003), Symptomatic features of intraplate earthquakes,
Seismological Research Letters, 74, 863-883.
Lomnitz, C. (1994), Fundamentals of Earthquake Prediction, 326 pp., New York, NY.
Musya, K. (1931), On the luminous phenomena that attended the Idu earthquake, November 6th,
1930, Bull. Earthquake Res. Inst., 9, 214-215.
Ouellet, M. (1990), Earthquake light and seismicity, Nature, 348, 492.
Roy, D. W., et al. (1993), Lineaments from airborne SAR images and the 1988 Saguenay
earthquake, Québec, Canada, Photogrammetric Engineering & Remote Sensing, 59, 1299-1305.
St-Laurent, F. (2000), The Saguenay, Québec, earthquake lights of November 1988-January 1989,
Seismolog. Res. Lett., 71, 160-174.
Takeuchi, A., and H. Nagahama (2001), Voltage changes induced by stick-slip of granites,
Geophys. Res. Lett., 28, 3365-3368.
Tributsch, H. (1983), When Snakes Awake: Animals and Earthquake Prediction, MIT Press,
Cambridge, Mass.
Tsukuda, T. (1997), Sizes and some features of luminous sources associated with the 1995 Hyogo-
ken Nanbu earthquake, J. Phys. Earth, 45, 73-82.
Yamada, I., et al. (1989), Electromagnetic and acoustic emission associated with rock fractures,
Phys. Earth Planet. Interiors., 57, 157 - 168.

St-Laurent et al.: Earthquake Lights
12
Yasui, Y. (1973), A summary of studies on luminous phenomena accompanied with earthquakes,
Memoirs Kakioka Magnetic Observatory, 15, 127-138.
Yoshida, S., et al. (1995), Observations of electromagnetic radiation associated with aftershocks of
the Hyogoken-nanbu earthquake, Bull. Earthquake Res. Inst., 70, 1-8.
Zoback, M. L., and R. M. Richardson (1996), Stress perturbation associated with the Amazonas and
other ancient continental rifts, J. Geophys, Res. B, 101, 5459-5475.
•Araiza-Quijano, M. R., and G. Hernández-del-Valle (1996), Some observations of
atmospheric luminosity as a possible earthquake precursor?