EVIDENCE FOR ELECTROMAGNETIC EMISSIONS DURING ROCK … · 2017-04-19 · We monitored the electromagnetic emission in the acoustic band (20 Hz-20 kHz) with a low frequency recorder

GNGTS – Atti del 22° Convegno Nazionale / 06.22

A. Nardi (1), M. Caputo (2), G. De Natale (3) e G. Scarascia Mugnozza (1)

(1) Dipartimento di Scienze della Terra, Università La Sapienza, Roma (2) Dipartimento di Fisica, Università La Sapienza, Roma (3) Istituto Nazionale di Geofisica e Vulcanologia, Napoli

EVIDENCE FOR ELECTROMAGNETIC EMISSIONS DURING ROCK LOADING AND FRACTURE:

A WAY TOWARDS EARTHQUAKE FORECAST Abstract. We present the results of experiments made on rock samples in the laboratory and in the field which show the very low frequency (ELF-VLF) radio emission associated to mechanical deformation and fracture of the rocks. The phenomenon is of great interest because it suggests the possibility to observe it as a systematic precursor of fracture and therefore of earthquake occurrence. We monitored the electromagnetic emission in the acoustic band (20 Hz-20 kHz) with a low frequency recorder called radio-geophone. The phenomenology at the origin of these emission has not yet been determined with certainty. Here we will refer to two hypotheses both based on the micro-fracturing implied by dilatancy. The field experiments have been made in a quarry of Calcare Massiccio during and after the explosions used to produce the rock material. The laboratory experiments has been made on samples of different rocks (Calcare Massiccio, Calcare a Rudiste, Scaglia Variegata, Porphyry, Corniola) subject to uniaxial compression. Besides the radio-geophone we used acoustic sensors and two conventional radio apparatus to monitor the acoustic emission, the pressure wave, and the HF and UHF electromagnetic emissions. The premonitory signals of the fracture have been systematically observed in the band ELF-VLF. These emissions have a maximum intensity in the radio-acoustic band, with amplitude decreasing with increasing frequency and therefore very low in the radio band of major use. The emission mechanism seems independent from the type of rock and from the type of perturbations used which lead to the fracture. Concerning the type of rock the signals seem more correlated to the structural homogeneity rather than to the mineralogical composition.

EVIDENZE DI EMISSIONI ELETTROMAGNETICHE DURANTE IL CARICAMENTO E LA FRATTURAZIONE DELLE ROCCE: UNA VIA VERSO LA PREVISIONE SISMICA

Riassunto. Presentiamo I risultati di esperimenti di laboratorio su campioni di rocce che, sottoposte a deformazione meccanica e fratturazione, danno origine a radio emissioni a bassa frequenza. Questo fenomeno risulta interessante perché potrebbe essere sistematicamente osservato prima della fratturazione delle rocce e, perciò, quale precursore sismico.

INTRODUCTION

The forecast of earthquakes is one of the major issues of science. It is extremely important for the number of lives it may save it is also a major scientific challenge for the difficulties it has encountered theoretically and experimentally. Many important books have been written on the state of art of earthquake prediction in various times beginning with the classics of Rikitake (1976), Asada (1982), Mogi (1985), where the different observational approaches are presented.

Today there are two major tentative solutions of the problem: the observation of precursors of the earthquakes in the field and in the catalogues. The experimental method is based on the observations of physical, chemical, electric, magnetic and tectonic phenomena which may precede the earthquakes; a typical example is the well known VAN method. Another method is based on precursors detected with sophisticated analyses of earthquake patterns which may precede the earthquakes,


a typical example being that of the M8 and CN algorithms (e.g. Keilis Borok et al., 1988).

The topic has also raised questions on the possibility of the solution of the problem mostly concentrated on criticisms of the various methods used to observe the precursors and on the alleged meaning of the precursors (Wyss, 1997; Kagan, 1997; Geller et al., 1997; Geller, 1999; Marzocchi et al., 2003).

The debate was very useful but did not discourage the community of those who pursue the goal of predicting the earthquakes. In the most authoritative scientific journals now and then appear articles with suggested methods of observation of precursors or of actual intermediate or long term forecasts.

It was perhaps De Rossi (1879) the first who gave a list of phenomena, occurring prior or during earthquakes, in his book where the earthquakes are considered as phenomena of an endogenous meteorology. Concerning the electric phenomena it was Galli (1910) to give the first list of 148 luminous phenomena observed in the atmosphere prior or during earthquakes.

The observation of phenomena precursory to earthquakes had several encouraging successes, as for instance the VAN method, but no systematic uncontested result was so far obtained. Often the laboratory confirmation of the observed phenomena was lacking.

Two cases of variation of the Earth’s electric potential were observed along the San Andreas fault that seemed related to incoming earthquakes (Corwin and Morrison 1977). Numerous variations of the Earth’s electric potential preceding earthquakes have been reported also in Kamchatka (Myachkin et al. 1972) and in China (Noritomi 1978, Zhu Fung-ming 1976). Rikitake and Yamazaky (1970) observed variations of electric resistivity at Aburatsubo (Japan) at the time of the Tokachi (1968) magnitude 7.8 earthquake; a variation occurred several hours before the earthquake and, at the time of the earthquake, a stepwise variation was also observed which did not occur instantly but seemed to take several minutes to complete.

In an experiment made in California (Mazzella and Morrison, 1974) an electric current of 200 A and 10 s period was transmitted through the ground and the ratio between the voltage and the current was observed between two electrodes 1.5 km apart. Beginning 60 days prior to an earthquake of magnitude 3.3. that occurred at 80 km distance from the electrodes. a variation of resistivity was observed which indicates that these variations may be observed at relatively large distance from the earthquake.

Kamogawa and Ohtsuki have observed radio emissions in the range ELF to HF before the strong earthquake of Sugadaira in 1980.

Langford et al. (1987) observed simultaneous emissions of protons and electrons during the fracture of crystals of MgO.

Bella et al. (1993) made a series of measures in the natural caves of the Central Apennines observing electric, magnetic, and seismo-acoustic signals in relation to a presumed pre-seismic stress. Few days before the earthquake of June 4th 1993 (Bella et al., 1998) an attenuation of 20 dB in the recordings of the long waves, about 216 kHz, of Radio Montacarlo has been observed.

Mognaschi (1998) made a study of the signals observed by a radio-amateur before the earthquake of May 12th 1997 in Central Italy; and called the attention to a precise dynamic of the phenomenon which interests a wide frequency band. From these remarks an experimental study followed (Mognaschi and Zezza, 2000), made with a radio receiver in the medium frequency band, which showed EM impulses


during, and sometimes before, the fracture following the uniaxial compression of a gneiss sample.

A comprehensive list of the precursory phenomena may be seen in Caputo (1987).

It is then clear that earthquakes may produce electromagnetic waves although we know that there is no generally accepted theory which relates the tectonic stress and fractures to the radio waves.

THE EXPERIMENTS

In order to observe the electromagnetic emissions (EME) of the rocks, two

types of experiments have been made: one in a quarry during the extraction of rock and one in the laboratory stressing several samples of different types of rocks.

In order to observe the EME we use a very low frequency radio receiver, called radio-geophone, which operates in the band 20 Hz-20 kHz of the ELF and VLF bands which correspond to the acoustic spectrum. The radio-geophone transforms the EME in acoustic signals in one of a 4 tracks magnetic recorder. The other 3 tracks are used to record simultaneously the acoustic signals of an acoustic microphone (measuring the velocity of the signal), of a pressure microphone (measuring the pressure of the signal) and the radio receivers in the bands LF, MF, HF one of them at the time in the field experiments. In the laboratory experiments, the UHF was recorded in the channel formerly used for the pressure microphone.

All the radio frequencies have been recorded in amplitude modulations while the radio-geophone, which worked directly in the acoustic band, gives the direct transduction, in the electric field, of the characteristics of the signal. The analogue records have been digitised through a sound card (which samples at a frequency of 44.1 kHz) and processed with normal audio software to obtain the waveforms and the spectra.

IN SITU EXPERIMENTS

These experiments have been performed in the cave of Colle Grosso of the firm BUZZI-UNICEM in Guidonia (Rome) which produces cements using rocks extracted from the cave. The production of the rocks is made working alternatively on the large steps, with 10 m height, located on the sides of a hill (see Fig. 1). The dimensions of the block extracted are 30m front, 10m height and 4 m depth. Every day are extracted 3500-4000 tons of material and, while the material is removed, one proceeds to the procedure for the extraction from another step.

The extraction is made by means of the explosion, in rapid succession, on 8 charges of explosive in wells 11 m deep filled from the bottom to a height of 6 m. The wells are aligned to the border of the step at a distance of 4 m from each other and from the border as shown in Fig. 1. In each experiment are used a radio-geophone and a radio-receiver to observe the EME emitted during the mechanical stress generated by the explosion and the variation of stress following the removal of the 4 m thick wall of rock.; an acoustic microphone and a pressure microphone allow to observe the instant of the explosion.

Initially we made 9 experiments with a rudimental recording system in order to show the existence of the EME. Six months later, with the more sophisticated


instrumentation described in this paper, the experiments have been repeated 8 times. The rock at the site of the experiments is Calcare Massiccio (Lias inferior) differently fragmented and sometimes with inserted Pyroclastyte. The explosive used is pulvirulento (based Diesel oil), with 145 kg one may extract 1200 m3 of rock. In order to obtain reliable data the antenna of the radio was mechanically isolated, the background noise was recorded. We controlled the possible effect of the electric trigger of the explosions and also the possible interferences of a power line located at 1.5 km distance from the cave.

During the experiments are monitored the acoustic signal, the pressure wave and the amplitude of the EME signal on the 750 kHz and in the band ELF-VLF. The recording instruments have been placed at a distance of 15 m from the wells with the explosive (see Fig. 1); the recordings began few seconds before the explosions.

V

S

P

Fig. 1 - Schematic representation of the steps in the Calcare Massiccio cave. P are the holes for the explosive. The parallelepiped V is the mass which will be removed by the explosion and S is the location where are instrumentation was placed during the experiment.

THE RESULTS

The analogical recordings are digitised associating the radio-geophone signal to all the other signals; the discussions of each case is made in the following.

The acoustic signals were clear but somewhat weak since the explosive was set in a way to cause minimal noise to the near villages. The impulse signalling the explosion coincides with that of the pressure microphone which was then used to record the initial time of the explosions to be correlated with the radio signals.

The pressure channel shows low frequency signals which define the time of the explosion but also a subsequent event probably due to the fall of the material ejected by the explosion.

The E channel of the radio-geophone has shown in all experiments a sequence of EME signals (precisely impulses) beginning at the instant of the explosion.

The radio channel recording higher frequencies did not show any significant signal above the environmental noise.


DISCUSSION OF THE ELF-VLF SIGNALS

In Figs. 2a and 2b the characteristic aspects of the signals resulting from the pressure microphone (P) and the ELF-VLF (E) are shown. In the same figures is shown a sequence of spectra which exhibit their evolution in time; each spectrum results from the analysis 512 samples separated by 1/44 103 s and therefore covering a time span of 0.01161 s and is represented by a segment parallel to the frequency axis and its amplitudes are proportional to the intensity of the grey in the segment.

Fig. 2a - Recording of experiment labelled UC2 04 GP in which a step of Calcare Massiccio is extracted from the cave. Above is the spectrogram recorder in the pressure microphone (EC), below is that of the E signal. The explosion and collapse appear in the EC recording with a continuous sequence followed by a succession of impulses caused by the fall of the ejecta from the wells. In the E channel instead one may observe a more complex sequence: A) an impulse at the time of the explosion, B) a brief sequence [another to be almost undistinguishable, D) a clear sequence of strong impulses well separated in time, E) random impulses occurring after the previous ones. In this case the event B is more evident than normally, on the contrary the event C is unusually reduced and almost undistinguishable from the sequence D. The abscissa is time in seconds. The ordinate is kHz; the intensity of the spectrum of the signal, in Volt, has the amplitude described with 256 different tones of grey. The horizontal lines are monochromatic noise arising from a power line (50 Hz) and from apparatus near the Radio-geophone (7.5 kHz); the impulsive events in the high frequency are caused by statics of the atmosphere known as spherics.

In the Fig. 2a and 2b it is seen that first signal recorded is an E impulse,

followed by an impulse P with some delay due to the slower propagation of sound in the air (circa 0.05 s). The P signal appears as the signal of a seismogram while E signal, after the initial impulse shows a set of other impulses. The first impulse is very probably due to the collapse of the wall of rock, while the others, arriving after more than a second, are probably due to the variation of the stress field in the layer of rock which is now the new front of the cave.

However in the oscillogram (Fig. 3 bottom) one may not distinguish the true nature of the E signals which, on the contrary, is evident in the spectrogram (Fig. 3 top) where we may distinguish three types of noise; the 3 top horizontal gray lines


(the highest is barely visible) are due to the amplifier of the microphone in the frequencies 7407 Hz, 10078 Hz, 13351 Hz. The lower horizontal bands correspond to the emission spectrum (50 Hz and associated harmonics) of the electric line which runs parallel to the cave. In the spectrogram appear also few vertical lines extending from 4 to 12 kHz and are due to the frequent atmospheric electrostatic phenomena which are well distinguishable from the radio-seismic signals which instead have the maximum energy in the lowest parts of the spectrum.

Fig. 2b - Recording of experiment labelled UC2 05 GP analogous to that presented in Fig. 2a, in which however results the variability of the emitted signals. In this case, contrary to that of Fig. 2a, the event C is clearly evident, but the sequence D is not well developed. The abscissa is time in seconds. The ordinate is kHz; the intensity of the spectrum of the signal, in Volt, has the amplitude described with 256 different tones of grey.

In the first approximation we may indicate as real radio-seismic signals all those

impulses which appear in the E channel and with wide band spectrum, with energy maximum in the low frequencies decreasing with increasing frequency. These characteristic properties are not appearing in the recording of the background noise, the only phenomenon which causes a somewhat similar spectrum would be a distant lightning.

THE TYPICAL SEQUENCE OF THE E SIGNAL

The pressure signal, used as a time reference to identify the events, is formed

by a sequence of oscillations, with decreasing spectral amplitude and extension of the spectrum, associated to the explosion and the collapse (see Fig. 2a and 2b). Let us call this sequence EC. After a short time follow some short sequences, probably caused by the fall of the material ejected from the explosion, in particular the excavation residue used to fill the top portion of the wells. Let us call these episodes EJ.


Fig. 3 - Radio-seismic successions of the type C and D in the band E of the experiment labelled UC 7 in the Calcare Massiccio cave. In abscissa is time in seconds. In the top figure is the spectrogram, where the ordinate is in kHz, with the intensity in Volt, expressed in 256 different tones of grey in the spectrum, represents its relative amplitde; in the bottom figure is the oscillogramme where the time is in a more detailed scale than in Fig. 2 and the ordinate, in Volt, is in percent of the maximum recorded.

The radio signal E is characterised by 5 distinct episodes which we classify in

the following 5 ways.

A. A single impulse, with weak or very weak intensity accompanying the explosion.

B. A very brief set of impulses with medium intensities which appears during the collapse but not systematically in coincidence with it, probably caused by the further fractures of the larger blocks right after the collapse.

C. A brief series of pulses with strong intensity, which follows at short time separation, appears right after the collapse probably due to the relaxation of the rocks which did not collapse and remained in their position.

D. A set of impulses with medium intensity which appear clearly separated, perhaps caused by the fracture of the rock still in situ.

E. Single clear impulses separated by relatively large time, as shown in the Fig. 2a and 2b, probably caused by single fractures in the rock still in situ.


Rarely appears a set of impulses probably caused by the interaction of clouds of charges with the atmosphere.

THE DYNAMIC OF THE EMISSION

Considering the dynamic of the explosion and the typical sequences of the

impulses E we may now hypothesise a relation between the mechanical event and the E emissions shown schematically in Fig. 4.

Fig. 4 - Dynamic of the collapse of a sped of limestone in the cave and correlation with the associated radio-seismic emission. 1) The explosive charges set in the Calcare Massiccio step. 2) Successive explosion of the charges aligned in the step. 3) Detachment of the front of the Calcare Massiccio step. 4) Collapse of the step. Fractures in the wall side of the former step.

The explosions in the 8 wells are not occurring simultaneously but in rapid

succession from the bottom to the top, from the first well through the 8th which favours the collapse of a uniform front. It is this uniform collapse which generates a unique large impulsive signal which we called A.

Follows the collapse of the mass already totally fragmented which, guided by the orientation of the wells, slides down on its side thus limiting the dispersion of the material and causing an orderly collapse which at times causes a signal B due to the fracture of the possible larger blocks still unfractured. After the fall of this layer of rock, the rock in situ forming the new wall of the cave is lacking the support of the fallen layer and is elastically unstable seeking a new equilibrium through a relaxation of the most exterior part of the volume and causing new micro-fractures emitting the signal C.


After the relaxation one may note on the top of the unfractured step a set of fractures parallel to the new front. To this phase we may associate the clear impulses of the type D. In Fig. 3 we may note with more detail the typical oscillograms and spectrograms of the sets C and D; we will later compare these with the corresponding ones recorded in the laboratory experiments.

LABORATORY EXPERIMENTS WITH UNIAXIAL COMPRESSION

These experiments have been made in the Experimental Laboratory of Italian

National Railways with a 500 tons heavily structured press, whose size allowed to use large rock samples generally with dimensions of 8 x 8 x 10 cm, which were increasingly compressed until they were fractured. The metal structure of the press and its location screened the environmental radiation. The preparation of the parallelepiped samples consisted of three phases: selection, cut and compaction test with special attention to the parallelism of the sided to be compressed to an accuracy of 0.06 cm while the accuracy requested by the press configuration was of 0.1 cm.

The laboratory analysis allowed to compare the data resulting from different rocks subject to the same pressure. We used samples of Calcare Massiccio taken from the cave in Guidonia and also from Palombara Sabina several kilometres fro Guidonia, in order to compare the results of the field tests with those of the explosion; but also other samples were used coming from rocks located in zones which are subject to earthquakes such as Scaglia Variegata and Corniola from the area of Norcia (North of Rome); moreover we used samples of Porphyry from the region of Trento region and Rudist Limestone from the geologic region of Lazio-Abruzzi. The rock samples used in the experiments are then rather heterogeneous in structure and mineralogical composition which allows to test the efficiency of the radio-seismic radiation.

To test the compaction of the different samples we used the Ultrasonic Tester CONTROLS – E46 of the Laboratory. An advantage of this experiment is the very reduced distance of the sensors from the samples, which reduced practically to zero the uncertainty of the velocity of the signals. Since the sample, at fracture, explodes laterally it was necessary to build a structure protecting the surrounding sensors.

The lower plate of the press is on rails set on the ground; the metallic mass of the press and the surrounding metal closets were a sufficient screen of the operation area from the outside disturbances. All the samples were positioned in the same place on the lower plate of the press.

Unfortunately in these experiments we could not use the pressure microphone due to a malfunction of its amplifier; however its function was completely substituted with the acoustic microphone. The recording channel left free from the pressure microphone was used to record the UHF radiation. The 4 channels recorded: the radiation E, the intensity of the electromagnetic signals in the frequencies 3.9 MHz (HF) and 460 MHz (UHF) and the acoustic signal.

The E signals recorded in these experiments are very similar to those recorded in the cave. The only differences are in the temporal distribution and the background noise; the signals are more clear since the atmospheric disturbances are not present in the laboratory . We made many check on environmental and instrumental noise to ensure the reliability of the results. For instance we verified that the press itself, under stress, does not produce any disturbance; also, with a short circuit of the samples with the plates and the of the plates themselves we verified that they are not


causing any influence on the phenomena recorded. The antenna receiving the E impulses was mechanically insulated from the press and from the surrounding structure sheltering the instruments from the splinters generated by the fracture.

THE LABORATORY RESULTS

Also in the laboratory tests the acoustic channel was used as a time marker of the mechanical events, but its impulses are non necessarily associated to the radio-seismic ones.

Fig. 5 - Final sequences of the precursory signals observed few seconds before the fracture indicated with R. Top and median couples are obtained with two different samples of Porphyry, the bottom couple is obtained with Scaglia Variegata. In each of the 3 couples (Mic) - (E) the top figure represents the acoustic signal (MIC), the bottom represents the electric signal. The abscissa is seconds. The ordinate is relative and in percent of the maximum recorded for the acoustic signal and in Volt for the electric signal.


As in the cave experiments, the analogical recordings have been digitised combining them with the radio-acoustic signal and analysed with the same methodology. The E initial emissions are clear and precede the instant of the fracture with a set of impulses which is rather regular for the same type of rock; the impulses are evident also in the other frequency bands.

From the rock samples examined resulted radio-seismic sources independently from the lithology and the place of origin of the rock. The intensity of the impulses varied depending on the lithology and the type of fracture. The impulses begin sporadically before the fracture and culminate with an outburst when the sample is fractured. Applying the pressure at a constant rate the emission of impulses occurs in a time interval which is variable but generally half the total duration of the increasing pressure. The E impulses seem associated with the acoustic signal (see Fig. 5) but are not synchronized with them, they show no systematic order of appearance and the time difference separating them is within 0.1 sec measured with an instrumental time resolution of 2.27 10-5 sec.

In the HF channel appear emissions shortly before and during the fracture, but not systematically, with intensity much lower than that of the E impulses of the ELF and not necessarily synchronized with them. The UHF channel showed non systematic very weak emissions before the fracture and weak emission at the time of fracture, sometime confused with the background noise; their waveform, as shown in Fig. 6, is similar to that of the HF signals probably because both are due to an amplitude demodulation.

In general the intensity of the emissions decreases with increasing frequency. Again the efficiency of the radio-seismic emission seems related to the mechanical behaviour of the rocks and to the different type of fractures; however the maximum intensity and the maximum time separation between the first signals and the fracture are recorded in the E channel. In this band one may distinguish long term, medium term and near precursors of the fracture.

THE PRECURSORS OF THE FRACTURE

The precursor signals are single impulses or sets of them compressed in a

relatively short time (about 0.2 s) interval, which appear preferably in the E channel probably because of the low intensity. Their amplitude spectrum reveals the typical distribution of a set of harmonics.

The time distribution of the impulses reminds that observed in the experiments on the cry of tin; when the pressure increases at a constant rate the phenomenon begins at a time before the fracture which is about 2/3 of that necessary to reach the fracture; in our case about 4 minutes of the 6 needed to reach the fracture. The impulses first appear sporadically, then at increasing rate in more numerous sets until they reach the culminating fracture outburst.

The E precursory signals are mostly associated to the acoustic signals as seen in Fig. 5. But observing the impulses with a higher time resolution it emerges that the corresponding maximum intensity peaks are not perfectly synchronous, they precede randomly one another with a time separation which varies between 0.01 sec and less than 0.001 sec. However there is a recurrent coincidence of the time of the initial impulses in the two sets.

There is a systematic relation between the maximum impulses E and the acoustic impulses which occur at the initial time of the micro-fracturing. This


phenomenon has been studied in all its evolution with a time resolution is 2.27 10-5 sec, which, is sufficient for the identification and separation of the impulses.

The fracture is a gradual process formed through the initial formation of micro-fracturing which leads to the sudden formation of some macro-fractures and to the final collapse of the rock samples. The culminating fracture sequence is not impulsive, it is a combination of macro and micro structural events which condition the details of the final collapse. In fact we did not find two broken samples of the same rock which show the same structural characteristics as we did not find two samples originally structurally identical except for the lithological classification. This resulted clearly from the ultrasonic analysis besides the direct observation in which the homogeneous rock, as limestone, presented stylolitic surfaces.

The E signals and the acoustic one reflect the dynamics of the fracture of rocks since we found more precursors in the non homogenous rocks.

Fig. 6 - Examples of culminating sequences in the laboratory fracture (indicated with F in the figure) of two samples of Porphyry. In each of the 2 couples (Mic) - (E) the top figure represents the acoustic signal, the bottom represents the electric signal. The abscissa is seconds. The ordinate is in percent of the maximum recorded for the acoustic signal and in Volt for the electric signal.

THE CULMINATING SEQUENCE OF FRACTURES

The culminating sequence of fractures coincides with the maximum intensity of emission and the maximum concentration of impulses; the latter are sometime so near one another that in few cases it is difficult to separate them. Comparing the results of the experiments on the different types of rock, as seen in Figg. 5 a, b, c, we may observe that this phenomenology does not present a significant variability,


rather it is similar to a succession, sometimes a fusion, of events of type C and D observed in the experiments made in the cave. Apparently there is no precise imprinting in the time and frequency domain which allows to distinguish one rock type from another.

The most evident differences are in reality conditioned by type of fracture, that is by the structural homogeneity of the sample. In a case the plates of the press were not perfectly parallel which caused the fracture of the sample in parallel splinters instead of the classic double wedge; in this case however we recorder E emissions with sequences similar to the others.

Fig. 7 - Example of precursory (Pre), culminating sequence (Par) and post fracture (Post) emissions in the laboratory fracture of a sample of Porphyry. The top figure represents the spectrum and the bottom represents the oscillogram of the same sequence where one may note the posthumous signals recorded in E-VLF channel. The ordinate of the spectrum is kHz and the of the grey tone represents its relative amplitude. On the oscillogramma the ordinate is relative intensity in percent. The abscissa is always seconds.

Also in the culminating fracture sequence, as we have already seen for the

precursors, the maximum peaks of the radio-seismic signal are not exactly coincident with those of the acoustic emission. However in the culminating fracture sequence the separation of the two maximum peaks is of the order of 1 ms while in the case of the precursors it is of the order of 0.1 s as seen in Fig. 4. In the case of the culminating fracture sequence we may find that the fracture occurs in an interval of


the order of 0.1 s. The Fig. 7 shows the radio-seismic signals compared with the acoustic ones during the culminating sequence.

Also in the laboratory experiments are present post fracture emissions which are illustrated in Fig. 7. This confirms the observations in cave about the relaxation of the material.

Fig. 8 - Comparison of the acoustic (Mic) and the 3 electric channels emissions in the laboratory fracture of a Porphyry sample. The maximum emission is observed in the Mic and the radio-geophone (E). In the signals at higher frequency, 3.9 MHz and 360 MHZ, the similarity between the latter two channels is due to the amplitude modulation. The abscissa is seconds. The ordinate is relative intensity in percent.

THE EFFICIENCY OF THE RADIO-SEISMIC SIGNAL

We have seen that rocks which are different from the petrographic and mineralogical points of view, because of different structural homogeneity, react differently during the fracture. In particular Scaglia Variegata, which is more fragile than porphyry, has manifested an emission of a smaller number of less intense electromagnetic signals, as if it had a smaller number of smaller micro-fractures, before and during the fracture. This phenomenon is in agreement with the theory of Griffith (1924) according to which the fracture begins in the points of weakness which one may identify in the discontinuities of mineralogical and petrographic nature. It is also in agreement with the experimental results of Mogi (1962) which showed that samples less homogeneous structurally have a larger number of micro-fractures.


Concerning the lithologic variability of our samples, Scaglia Variegata represents a rock structurally more homogeneous while Porphyry, which showed maximum reactivity in the signal emission is more heterogeneous and granular. The Calcare Massiccio which reacted with large and frequent impulses has a uniform micro-crystalline structure but has numerous stylolite surfaces and also several spar surfaces.

With any type of rock the three radio channels recording the signal, as seen in Fig. 8, show that the amplitude of the signal decreases with increasing frequency, which could explain why these signals, normally, are not interfering with radio communication and also why the anomalous signals reported by radio amateurs before earthquakes, are not systematic.

Fig. 9 - Typical waveform of the ELF band’s pulses.

GENERAL COMMENTS: SPECTRAL DISTRIBUTION AND WAVEFORM

The radio-seismic emission recorded in the E band are characterised by sequences of impulses. These sequences, as we have seen, have a characteristic temporal distribution correlated with the phenomenon originating them: explosion and collapse or uniaxial compression. The impulses of the sequences have common properties which are then independent from the phenomenon originating them. In other words the sequences observed in the quarry and in the laboratory differ only for the peculiar combination of impulses, which however have the same spectral properties.

The spectral analysis reveals that the spectrum of a single impulse (recorded with a sampling of 44100 Hz and a resolution of 81 Hz) has a typical maximum intensity in the low frequencies of the band between 200 and 500 Hz, it decreases with increasing frequency up to 14000 Hz and, at times, up to 20000 Hz in the strongest signals. The waveforms of the impulses recorded in the E band have some variability but are typical and similar to that shown in the Fig. 9.

In the Fig. 10 and 11 we show two typical sequences of radio-seismic and acoustic signals. The sequence, seen in Fig. 10, concerns a recording in the quarry, that shown in the Fig. 11 concerns the fracture of a sample of porphyry; both figures exhibit the radio-seismic emission in the E band and the acoustic emission. On the side of each spectrogram are indicated the presumed causes of the emission.


HYPOTHESIS BASED ON THE ELECTRIC SPARK

We have already possible explanations for the electromagnetic emissions associated to the micro-fracturing preceding the fracture. These explanations are based on the generation of free electric charges or of esoelectrons. Of particular interest is the hypothesis suggested by Mognaschi (1987, 1998), which does not require piezoelectric phenomena but is based directly on the micro-fractures, which break the chemical bonds and cause an unbalance in the distribution of electric charges. When the material is not a conductor, the fracture would separate the free charges generating an electric field which in turn would create an electric micro-discharge between the two sides with a consequent electromagnetic emission. This effect may be verified by simply unwinding scotch tape and observing that little sparks are formed which may be easily recorded with a radio.

Fig. 10 - A typical field experiments. Acoustic channel (Ac) and E channel. The top figure is the continuation in time of the bottom one. Note that the acoustic signal is present only at the explosion and collapse and later in the fall of the by products of the explosion. The abscissa seconds. The ordinate is kHz and the intensity of the grey tone in the spectrum represents its relative amplitude with 256 different tones.


Fig. 11 - A typical laboratory experiments on a Porphyry sample for which we show the acoustic channel (Ac) and E channel. The top figure is the continuation in time of the bottom one. Note that the acoustic signal is well present before and after the fracture. First we have the compression with the formation of isolated micro-fractures which, with increasing time and pressure, increase in number per unit time causing emission of acoustic and electric signals. Then the micro-fractures coalesce and originate the fracture. The abscissa is seconds. The ordinate is kHz and the intensity of the grey tone in the spectrum represents its relative amplitude with 256 different tones.

HYPOTHESIS BASED ON THE MINERAL CIRCUIT

An attractive hypothesis is that of the mineral circuit which may be considered similar to a “transductor” of the cry of rock. In a compact rock the water does not permeate uniformly the volume but circulates only inside the fractures. One may then assume that it has an intrinsic conductivity that is function of the minerals forming it and is independent from the degree of saturation of the total volume.


The greatest majority of rocks, independently from their nature and origin, may be considered a mixture formed in great majority with non conducting minerals and only in a minor part with minerals that have piezoelectric properties. We have thus a basically insulating mass where are distributed conductive elements which may form electric circuits with direct contact or without contact for capacity connection.

Dielectric crystals

Conductive crystals

C

A E

F

D

B

F

E

D

C

A

B

Piezoelectric crystals

Fig. 12 - Symbolic representation of the radio-seismic mechanism: a rock is subject to compression (left) and the conductive minerals imbedded in a non conductive background form chaotic circuits which connects the piezoelectric crystals; these in turn create the electric tension. The phenomenon is represented by an electronic circuit (right) in which every cycle of compression and dilatation will invert the polarity of the elements.

The circuits, that are closed paths, are chaotically present in the three dimensions and are longer than the minimal distance which could connect the active elements. They will be as much longer as is rarefied the distribution of the conductive minerals to the limit that the paths will not be able to close and form a circuit.

With these hypotheses, we may understand that a pressure wave penetrating the rock will excite the piezoelectric crystals and generate free charges which in turn will run along the mineral circuits connecting, in series or parallel, some active crystals (see Fig. 12).

The pressure wave, that consists of compressions and dilatations periodically alternating, will generate an alternating polarity. The mineral circuits will nullify the differences of electric potential and will have an alternate current with frequency equal to that of the pressure wave.

Considering that the wavelength of the seismic waves generating the currents are kilometres long one may believe that the efficiency of the system of loops, as generators of electromagnetic signals, be not high. However the great number of active elements, the enormous energy released by tectonic phenomena and the vast extension of the volume involved, which covers all the focal area and all the volume of lithosphere which participates to the genesis of the earthquake, the alternate currents (generated by the compression and the decompression of the wave) are possibly capable of generating the signals recorded also with a low efficiency of the process.


We may also put forward the hypothesis that through this mechanism the earthquake be related to electromagnetic waves with extremely low frequency in two ways: acoustic and ultrasonic frequencies emitted before the earthquake, related to the size of the focal volume, and also successively acoustic emissions mostly infrasonic during the shock.

CONCLUSIONS

We have shown that in the field experiments, during and after the explosions in a quarry used to produce the rock material of Calcare Massiccio, and in the laboratory experiments with samples of different rocks (Calcare Massiccio, Calcare a Rudiste, Scaglia Variegata, Porphyry, Corniola) subject to uniaxial compression, have been systematically observed emissions of electromagnetic signals in the band ELF-VLF.

These emissions have a maximum intensity in the radio-acoustic band, decreasing with increasing frequency and therefore very low in the radio band of major use. The emission mechanism seems independent from the type of rock and from the type of perturbations used which lead to the fracture.

Concerning the type of rock the signals seem more correlated to the structural homogeneity rather than to the mineralogical composition. Acoustic sensors and two conventional radio apparatus monitored the acoustic emission and the pressure wave. An emblematic example of precursory (Pre), culminating fracture outburst (Par) and post fracture (Post) emissions in the laboratory fracture of a sample of Porphyry shown in Fig. 7 where the top figure represents the spectrum and the bottom represents the oscillogram of the same sequence where one may note the posthumous signals recorded in E-VLF channel.

Acknowledgement. We are grateful to the BUZZI-UNICEM firm in Guidonia (Rome) for allowing and cooperating in the series of experiments, which required the explosion of charges, in their cave. We also grateful to the Experimental Laboratory of the Ferrovie dello Stato in Rome, in particular to Dr. Misano, for allowing to use the apparatus of the laboratory and to have the help of the personnel of the laboratory.

REFERENCES Asada, T. (Ed.), Earthquake Prediction Techniques, University of Tokyo Press, 1982. Bella, F., Biagi, P.F., Caputo, M., Della Monica, G., Ermini, A., Pettinelli, V. and Sgrigna, V.,1993.

Ground tilt variations detected in the Central Apennines in the period 1986-1989 and their correlation with seismicity, Nuovo Cimento C, 16, 303-311.

Bella, F., Biagi, P.F., Caputo, M., Cozzi, E., Della Monica, G., Ermini, A., Plastino, W., Sgrigna, V., 1998. Field strength variations of LF radio waves prior to earthquakes in central Italy, Physics of the Planetary Interiors, 105, 279-286.

Caputo M., Introduzione alla sismologia e segnali precursori dei terremoti, Calderini Editore, Bologna, 1987

Casati, P.,1996. Scienze della Terra, Città Studi Edizioni. Corwin, R.F. and Morrison, H.F., 1977. Self-potential variations preceding earthquakes in Central

California, Geophys. Res. Letters., 4, 171-174. De Rossi, M.S., 1879. Meteorologia endogena, Fratelli Dumolard Editori, Milano. Galli, I., 1910. Raccolta e classificazione di fenomeni luminosi osservati nei terremoti, Bollettino della

Società Italiana di Sismologia, 14.


t

Geller, R.J., Jackson, D.D., Kagan Y.Y. and Mulargia, F., 1997. Earthquakes cannot be predicted. Science, 275, 1616-1617.

Geller, R.J., 1999. Earthquake prediction: is this debate necessary? In: Is the reliable prediction of individual earthquakes a reliable scientific goal? Nature Debate, http://helix.nature.com/debate/earthqwuake/.

Griffith, A., 1924. Theory of rupture, Proc. 1DT Int. Cong. App. Mech., Delft. Kagan, Y.Y., 1997. Are earthquakes predictable?, Geophys, J. International, 131, 505-525. Langford S. C., Dickinson J. T. and Jensen L. C., 1987. Simultaneous measurements of the electron

and proton emission accompanying fracture of single-crys al MgO. J. Appl. Phys., 62, 1437-1449.

Marzocchi, W., Sandri, L. and Boschi, E., 2003. On the validation of Earthquake forecasting models: the case of pattern recognition algorithms, Bull. Seis. Soc. Am., 95, 5, 1994-2004.

Mazzella, A. and Morrison, H.F., 1974. Electrical resistivity variations associate with earthquakes on the San Andreas fault, Science, 185, 855-857.

Mogi, K.,1962. Study of the elastic shocks caused by the fracture of heterogeneous materials and its relations to earthquake phenomena, Bull. Earthquake Res. Inst. n.40, Univ. Tokyo, 1962.

Mogi, K., 1985. Earthquake prediction, Academic Press, New York. Mognaschi E.R., 1997. Segnali radio che precedono eventi sismici, Radioonde 8, ottobre, 1997. Mognaschi E.R., 1998. Precursori elettromagnetici dei sismi, XVI Meeting dell'Associazione Italiana

Radioascolto a Sirolo (AN) 1-3 maggio 1998; Radioonde, 12, maggio 1998; Radiorama, inserto n.10 ottobre 1998; Internet: pagina web http://web.tiscali.it/G2000/R_Mog.html

Mognaschi, E.R. and Zezza, U., 2000. Detection of electromagnetic emission from fracture of rocks and building stones under stress, Atti del “5th International Congress on Restoration of Architectural Heritage”, Firenze, 17 settembre, 553-562.

Myatchin, V.I., Sobolev, G.A., Morozov, V.N. and Preobrazensky, V.B., 1972. The study of variation in geophysical fields near focal zones in Kamchatka, Tectonophysics.

Rikitake, T., 1976. Earthquake prediction, Elsevier, Amsterdam. Rikitake, T.,1986. Previsione dei terremoti, Libreria Dario Flaccovio Printer. Rikitake, T. and Yamazaki, 1070. Strain steps as observed by resistivity variometer,

Tectonophysics, 9,197-203. Scholz, C.H., Sykes, L.R. and Aggarwal, Y.P. 1973. Earthquake prediction: a physical basis,

Science, 1. Wyss, M., 1997. Cannot earthquakes be predicted?, Science, 278, 487-488. Zhu, F.M., 1976. Prediction warning and disaster prevention related to the Haicheng earthquake

of magnitude 7.3, Prod. Lectures by the Seismological Delegation of the People’s Republic of China, Seism. Soc. Japan Special Publication, 15-26.

Documents

EVIDENCE FOR ELECTROMAGNETIC EMISSIONS DURING ROCK … · 2017-04-19 · We monitored the electromagnetic emission in the acoustic band (20 Hz-20 kHz) with a low frequency recorder