THE EFFECT OF PRESSURE ON THE ELECTRICAL CONDUCTIVITY OF KTB ROCKS G. NOVER 1, S. HEIKAMP 1, A. KONTNY 2 and AL DUBA 3 1Mineralogisches lnstitut der Universitiit Bonn, Bonn, Germany; 2KTB Feldlabor, Windischeschenbach; 3Lawrence Livermore National Laboratory, Livermore, Cal., USA (Received 15 November 1993)

Abstract. Complex electrical resistivity and permeability were measured on two gneiss samples and nine amphibolites (originally located at a depth of 4150 m to 5012 m) from the main drilling of the German deep drilling project (KTB). Measurements were performed as a function of hydrostatic pressures up to 240 MPa on core samples (30 mm in diameter and 10-20 mm high). For each measurement, two samples were used, one being parallel, and one perpendicular to the borehole axis. At low pressures and again at maximum pressure the frequency dispersion (t kHz up to 1 MHz) of the complex resistivity was measured using a two electrode device. An unusual pressure effect was detected on some of the samples and was established to be due to the oriented deposition of good conducting phases in the foliation. Rock fabric and the orientation of ore mineralization was measured on thin sections and polished sections prepared from the same samples. Key words: Complex electrical resistivity, ktb, permeability, pressure

1. Introduction The K T B drill site is located in the western margin of the B o h e m i a n M a s s i f in the Zone of Erbendorf-Vohenstrauss (ZEV, N E - B a v a r i a / G e r m a n y ) . The Z E V is situated at the boundary of the Variscan units Saxothoringian and Moldanubian and is characterized by an amphibolite facies m e t a m o r p h i s m . The main lithologies of the K T B borehole are paragneisses and metabasites. These rock sequences experienced large-scale deformation and m e t a m o r p h i s m , now visible as faults and cataclastic shear zones that transect the rock samples. Two types o f cracks were found, in-situ cracks oriented according to rock texture and cracks resulting from pressure and temperature release. The last ones are perpendicular to the m i n i m u m horizontal stress, while the in-situ cracks correspond to the actual stress field. Thus lithological parameters such as permeability are anisotropic and high in horizonal directions even at in-situ pressure conditions. In vertical direction the permeability is usually m u c h lower (Huenges et al., 1990). The average porosity o f gneisses is 0.7±0.3, and metabasites are about 1 . 1 ± 0 . 9 % (B~icker et al., 1990). G o o d conducting ore minerals (sulfides and oxides) as well as graphite (Lich et al., 1992) were found in the rocks (Kontny, 1993) in the foliation. M e t a m o r p h i c and hydrothermal ore mineralization are m o s t abundant (Kontny, 1994). In the paragneisses ilmenite is m o s t l y linked to the phyllosilicate orientation (Siegesmund et al., 1991, 1993). The formation o f sulfides and graphite in shear zones, therefore m a y be related to circulating hydrothermal fluids, as they were observed in the K T B rocks beneath about 3200 m (Figgemeier, et al., 1992). Surveys in Geophysics 16:63-81, 1995. (~) 1995 Kluwer Academic Publishers. Printed in the Netherlands.

64

G. NOVER ET AL.

Rauen and Soffel (1993) reported a significant azimuth dependence and a strong anisotropy as well as inhomogeneities on cores from the KTB being measured directly after recovery. They used a moving point 4-electrode technique and detected mean resistivities of gneisses and amphibolites to be in the range (~ 103f~m). They point out that these data do not reflect in-situ anisotropies due to stress release and resulting microcrack formation. Duba et al. (1994) performed conductivity and permeability measurements on samples from the KTB at hydrostatic pressures up to 225 MPa. They report a significant anisotropy of the resistivity and permeability when measured in three perpendicular directions in regard to the foliation and lineation. In the foliation in the direction of the lineation the resistivity was low and the permeability high, while perpendicular to the foliation the resistivity was high and the permeability usually more then two orders in magnitude lower. On some of the samples an unusual pressure effect was detected, the resistivity of the sample decreases with increasing hydrostatic pressure. This observation was attributed to the reconnection of highly conducting phases when cracks were closed at elevated pressures. Most of the samples exhibited a rather high conductivity of about 10 -2 S/re. The cause for this high conductivity is most likely caused by a network of highly conducting phases interconnected by high-salinity pore fluids. This paper reports on measurement of the pressure dependence up to 240 MPa, gas-permeability and complex resistivity in the frequency range 1 kHz up to 1 MHz. Our aim was to correlate electrical and permeability data using different porosity models, and to consider rock-fluid interactions and their influence on the electrical conductivity as well as the contribution of ore and carbon mineralization on the high conductivity of the samples. Two orientations, axial and radial, of samples in regard to the borehole axis were chosen. Due to the inclination and dip of foliation its orientation in the cylindrical samples was from nearly horizontal to vertical. Thus the anisotropy of the permeability and resistivity could be determined.

2. Experimental and Results 2.1. LITHOLOGICAL DATA AND SAMPLE PREPARATION

Samples investigated originate from the metabasic sequence of the KTB-Hauptbohrung at a depth of 4000 to 5000 m, where amphibolites, garnet-amphibolite and hornblende gneisses with minor metagabbro intercalations are found. The metabasic rocks are essentially derived from an intrusive complex of basaltic composition with an E-MORB (middle-ocean-ridge-basalt) character (Schalkwijk, 1991) and by a polyphase metamorphism. High pressure metamorphism (12-14 kbar, 580-650 ° C) and amphibolite facies metamorphism (7-9 kbar, 650-700 ° C) are documented (O'Brien et al., 1992). The rocks show an inhomogeneous deformation under amphibolite facies conditions. The foliation in the investigated section is steeply inclined (50 ° to vertical) and the dip direction alternats between

65

PRESSURE AND ELECTRICAL CONDUCTIVITY OF KTB ROCKS

TABLE I C o l u m n 1 lists the identification o f the s a m p l e s , c o l u m n 2 g i v e s the orientation o f the s a m p l e (A=parallel to borehole-axis; R=radial, p e r p e n d i c u l a r to borehole-axis, b u t in the foliation plane, c o l u m n 3 g i v e s the d e p t h f r o m w h i c h the s a m p l e s were taken, c o l u m n 4 describes the rock type ( g n = g n e i s s , amp--amphibolite). C o l u m n 5 lists the permeability data at 5 M P a (value in brackets give data m e a s u r e d at elevated p r e s s u r e s in M P a ) , the sulfide a n d ilmenite c o n t e n t is s h o w n in c o l u m n s 6 a n d 7. T h e angle b e t w e e n the axis o f the s a m p l e a n d the n o r m a l o f the foliation p l a n e is g i v e n in c o l u m n 8 density data are in c o l u m n 9. Sample#

H001B20a

H001E44

H003A4

H004A5

H005B20

H006G52b

H007C25

H008B22

H009B11

H010C28

H011A8

D

Depth(m)

A

4150.41

R

tv

R o c k type

amp amp

A

4152.75

gn

R

n

gn

Permeability

Vol. %

(ttD)

sulfide

Ilmenite

(9

Density

0.01 0.03

-

1 1

90 ° 150

2.89

0.0065

-

-

450

2.76

-

<1

85 °

45.80(100)

A

4251.22

amp

0.321

-

2-3

300

R

n

amp

2.138

-

1-2

90 °

A

4341.99

amp

0.028

-

< 1

90"

R

n

amp

0.038

-

< 1

45 °

A

4448.37

amp

R

n

amp

A

4516.60

amp

R

n

amp

A

4594.64

gn

0.024

R

n

gn

0.91

1

1-2

45 °

-

-

90 °

0.025

1-2

6-7

350

0.299

<1(2)

2(6)

850

<1

<<1

35 °

<1

-

85 o

67.73(400)

71.9(200)

A

4647.47

amp

1.567

<<1

<<1

80 °

R

t!

amp

2.062

-

-

80"

0.0051

A

4685.55

amp

-

-

600

R

t!

amp

56.47(800)

-

-

800

34.9(400)

A

4822.23

amp

-

-

850

R

n

amp

0.53

-

-

-

A

5012.52

amp

-.-

-

45 °

2.86

R

n

amp

-

-

-

97.36(200)

2.93

2.66

2.92

3.00

2.78

2.95

3.03

2.96

66

G. NOVER ET AL.

Fig. 1. Ilmenite in garnet-bearing amphibolite of the KTB-Hauptbohrung (sample H006, 4516,6 m, air, parallel polars, width of field 0.42 ram) Ilmenite with a rim of titanite intergrowth with pyrrhotite in fine-grained, foliated amphibolite of the KTB-Hauptbohrung (sample H003, 4251.22 m depth, air, parallel polars, width of field 0.4mm).

PRESSURE AND ELECTRICALCONDUCTIVITYOF KTB ROCKS

67

[

Computer,} I Control Unit,

steel-adapter "-~,~,',"( samt)le " ~ - - ~

2, , , ~

iU

[N.'N]

hydrost.at ic l)rcssure up to 250 Ml'a

Fig. 2. Illustrationof the autoclave and the equipment used for the permeability measurements. The gas pressure on one side of the sample is kept constant at 5 MPa and the pressure increase in a constant volume on the other side of the sample is monitored as a function of time. The assembly is held together by a shrink tube which also acts to isolate the rock from the N2 gas, the hydrostatic pressure medium.

E and W. The rocks are partially strongly altered and show locally cataclastic overprint (Lich et al., 1992). The main rock forming minerals of these amphibolites are hornblende, plagioclase and garnet together with minor quartz and biotite. Accessories are K-feldspar, apatite, some zircon as well as titanite and ore minerals. Hornblende gneisses consist of the same minerals as amphibolites but are richer in quartz, plagioclase and biotite (Lich et al., 1992). In the metabasites and hornblende gneisses the Fe-Tioxide ilmenite predominates. As minor components rutile and the sulfide pyrrhotite, pyrite, chalcopyrite and rare sphalerite occur. Depending upon the degree of rock alteration, ilmenite is replaced by titanite and/or leucoxene. The replacement always starts from the grain rims or follows fissures in the oxides. Ilmenite mostly is oriented in the plane of foliation and shows elongated xenomorphic grain shapes. The grain size varies from a few # m to a maximum of about 400 pro. However, most grains are less than 150 #m. The ilmenite content of the samples range from < 1 to 7 vol % (Table I, Fig. 1). The oxides are related to the amphibolite facies parageneses. Mostly they were formed by recrystallization of primary oxides or by retrograde reactions from the silicates. Symplectic intergrowth with hornblende gives evidence for this origin.

68

G. NOVER ET AL.

1E+2--J

R

1E+1

~'~ 1E+0-

1E-1-

~ :

~

.....

~

H~7 eZS~

HOOl B2~IR --O--

H0O?CZSA

_ .......

IE-

"~r''t*l''''*~'"l'" 0

40

..... 80

'1';'''"''1'""'~''1 120 160

pressure (MPa)

'' 200

....

'"1 240

Fig. 3. Pressure dependenceof the permeability of some selected samples. The small sketch shows the orientation of the samples (A) and (R) in regard to the borehole axis. A total of 11 samples were chosen for this investigation. The samples were selected using different criteria, first samples from the three lithological units are included, second, the orientation of the foliation in the samples should range from horizontal to vertical and third, the content of highly conducting phases should cover the range observed in the metabasic sequence (Table I). As far as it was possible, directly neighbored samples (identical samples) were used for the electrical and permeability measurements. At least 22 samples (diameter 30 mm and about 20 mm high) were used for this study, which were oriented in two perpendicular directions with regard to the borehole axis and the foliation and lineation present in the rock (see small sketch in Fig. 3). 2.2. PERMEABILITY Permeability was measured in an autoclave using a pressure transient method and Argon gas as the permeating medium. The cylindrical sample was enclosed in shrink tubing to prevent contamination of the sample by the hydrostatic pressure medium. Hydrostatic pressures ( P ~ t l ~ ) were increased up to 240 MPa. The gas pressure (Pi~) was kept constant at 5 MPa on the left side of the sample (Fig. 2), and readings of the pressure increase (in the constant volume) on the right side of the sample (Pout), were used to calculate the permeability (Zoback & Byerlee 1975) on the basis of the Darcy equation (Equation 1):

k = Q~I/ApS

(1)

PRESSURE AND ELECTRICALCONDUCTIVITYOF KTB ROCKS

69

with Q the volume of the fluid phase flowing through the sample per unit time, r] the viscosity of the gas, t the length of the sample and S the surface area; Ap is the pressure gradient, k is the permeability. The permeability decreases by about one order in magnitude at pressures between 5 MPa up to 100 MPa. At low confining pressures as well as at high pressures a strong anisotropy (AN) of the permeability AN = V/~A/kR

(2)

was observed in the orientations A (axial) and R (radial) (kA is the permeability measured in axial direction, kR in radial orientation). Data for AN range from 0.001 to 0.87. Low permeabilities were measured in directions inclined with respect to the foliation plane. The permeability was much higher, even at high pressures, when measured in the foliation plane. Results are shown in Fig. 3 which include only data for three samples covering 5 orders in magnitude and being representative for the highest and lowest permeabilities measured. The other data are omitted in this Figure for clarity but they are included in Table I. 2.3. COMPLEX ELECTRICAL RESISTIVITY AS A FUNCTION OF FREQUENCY (1KHZ-1MHZ) AND HYDROSTATIC PRESSURES (UP TO 225 MPA). If a rock sample is subjected to an electrical field E, a current of density [ flows through the sample (Equation 3). This current I is the sum of a conduction current [ e and a displacement current Id ( I -= I~ + Id), both being a function of frequency (I¢ = a~ E and Id = v / v t ( c * d E , er* = ~ * + ¢*d/ico, c d* is the permittivity of the dielectric). Input E and response I are related by the transfer function or* which is defined in a linear system: I = cr*E

(3)

a* = a r + io N

(4)

The transfer function is used to describe the physical phenomena of the charge transport processes on the basis of equivalent circuit models. As a first approximation R C parallel circuits are used to describe the conduction process in the pore space (Ruffet et al., 1991; Lockner & Byerlee, 1985). The resistance R describes the resistivity in the bulk electrolyte solution, while the capacity C is linked to dielectric polarization between the negative surface charges of the rock forming minerals, the water dipoles and ions in the electrolyte. Consequently a network of pores (cracks) can be described by an array of R C parallel circuits. Any variations in the pore geometry (due to pressure or temperature variations) or the inner surface area are coupled with a variation of these model parameters. The frequency dependence of the complex resistivity therefore provides information on the charge transport mechanism if for example the conductivity changes from an electrolytic to a more electronic conduction process.

G. NOVERET AL.

70

to

AC Impedance

250 MPa

~ electrode ,\~--.-- filter ~N'~ sample ,~ shrink~ tubing

Analyzer

IKHz to 1MHz

-

Fig. 4. Equipmentused for the measurementof the frequencydependentcomplex electrical conductivity of KTB core samples. Hydrostaticpressure was increased stepwise up to 225 MPa. The sample is sandwichedby the electrodesand filterpaper, and evacuatedand back saturated with a 1 M NaC1solution.The side of the electrodes in contactwith the filterpaper have concentricgrooves with volumemore than sufficientto containthe pore saturantin the sample and the filterpaper.

The autoclave used for the electrical measurements (Fig. 4) allowed an increase of the hydrostatic fluid pressure up to 225 MPa. The cylindrical sample was enclosed in shrink tubing to prevent contamination of the sample with the glycerol pressure medium. Electrodes, placed on both sides of the sample, had fluid reservoirs to avoid high pore pressures and ensure that electrical measurements were made under "drained" conditions. Filter paper placed between the electrode and the sample provided good electrolytic contact to the sample. The sample was evacuated and back-saturated with a 1 molar NaC1 solution (7.8 S/m) to reduce the contribution of surface conductivity to the bulk conductivity of the sample. Measurements of surface conductivity were performed on different KTB samples saturated either with NaC1 and KC1 electrolytes different in molarity or distilled water. The surface conductivity was determined to be in the range of about 5 × 10 -5 S/m. Each measurement started at normal pressure, the pressure was increased stepwise (20 steps) up to 225 MPa. At the beginning and at 225 MPa complex electrical data were measured in the frequency range 1 kHz up to 1 MHz. Measurements at pressures in between were performed at a fixed frequency of 1 kHz. In Fig. 5a the data of those measurements are compiled that exhibit a "normal" pressure dependence of the volume resistivity. Here the normalized volume resistivity increases as a function of pressure. The resistivity increase strongly depends on the number of cracks that can be closed, and thereby depends on the orientation of the sample in regard to the foliation. The lithological investigations revealed that a major part of the in-situ porosity is linked to the foliation. Thus increasing hydrostatic pressure reduces the geometry of the interconnected pores and thereby reduces the rock conductivity. Except for sample H008, the foliation is

PRESSURE AND ELECTRICAL CONDUCTIVITY OF KTB ROCKS

71

steeply inclined, and the angle between the normal on the foliation and the sample axis is close to 90 ° (Table I). Consequently a significant anisotropy of the rock resistivity could be expected. Resistivity data for normal pressure conditions are given in Table I for both orientations of the samples together with the orientation of the foliation. Data show that the quotient of the resistivities in two perpendicular directions can be up to 6, but in general it is below 4. In Fig. 5b the data from samples are compiled that exhibit an "unusual" pressure effect (Duba et al., 1994). Some of the samples did show in the low pressure region (up to about 20 MPa) a normal pressure effect, but it was reversed for the high pressure region. In general all of these data sets reveal a lower resistivity at 225 MPa than at normal pressure conditions. The orientation of the normal of the foliation to the sample axis is around 45 ° (Table I) though 3 samples reach about 85 °. The initial volume resistivities cover the range 1000 f~m (H006G52bA) down to 56 f~m (H008B22R) and the correlated permeabilities are 0.025 #D and 2.062 #D respectively. It can be seen from Table I that the unusual pressure effect is dominant when the ore mineralization is comparably high (H006). This indicates that this effect may be linked to the presence of highly conducting phases. The pressure effect (PE) of the resistivity ranges from 0.68 up to 1.38 (lp= low pressure, hp= high pressure). 7 - _

P E = ~/eip/ehp

(5)

2.4. COMPLEX RESISTIVITY

In order to use the frequency dispersion as a tool for the detection of electronic charge transport processes, the complex response was measured at normal pressure and at 225 MPa in the frequency range 1 kHz up to 1 MHz. Lower frequencies than 1 kHz were not used for the interpretation due to electrode polarizations that dominate in the low frequency region. In Fig. 6a and 6b data for nine of the eleven samples are shown which were measured at normal pressure and at 225 MPa respectively. The data sets having high resistivities were omitted for clarity, but the complex data of these samples are included in Table II. Fig. 6 compiles the complex data shown as a Cole-Cole diagram where the real part of the resistivity is plotted versus the imaginary pare The arcs "intersect" the real axis at the low frequency end, while frequencies up to 1 MHz are found at the left side of the ares. These data were parametrized by fitting R C model data to the measured data. The simplest case to describe a rock sample by a simple equivalent circuit model uses only a resistor (R) and a capacitor (C) in parallel (B0mer, 1991, Nover et al., 1991, Ruffet et al., 1991). Such a model describes the complex response of rock samples very well as long as no contribution of electronic charge transport (graphite, ore) to the electrolytic charge transport in the pore system must be considered. The low resistivities of the rock samples imply the use of a modified equivalent circuit to model the data. As a first approximation a resistor was used which was in parallel to the R-C circuit fit the data well. The additional resistor should model any good

72

G. NOVER ET AL.

TABLE II Complex resistivity data of KTB samples measured at normal pressure and 225 MPa hydrostatic pressure. Column 1 identifies the sample, column two gives the orientation (A = axial, R = radial). Column three and four list the R and R1 data, while values for C are given in column 5, n is the angle under which the semicircles are inclined versus the real axis. The following columns list the data for the equivalent circuit measured at a pressure of 225 MPa. Normal pressure R(ohm) R l ( o h m )

C(#F)

n(grd)

225 MPa R(ohm) R l ( o h m )

C(/tF)

n(grd)

H001B20a

A 460 R 651

2735 1624

1.6 1.7

84 83

860 631

3010 1576

1.6 1.7

85 84

H001E44

A 408 R 190

2103 2908

1.7 1.7

82 79

303 170

2541 3072

1.7 1.7

82 79

H003A4

A R

141 158

1350 1410

1.6 1.7

84 84

100 287

1841 1837

1.6 1.9

84 77

H004A5

A 493 R 377

4295 3377

1.7 1.7

82 82

1127 309

5717 3549

1.7 1.8

84 82

H005B20

A 72 R 30

1786 834

1.6 1.4

82 76

66 32

3147 1492

1.6 1.5

83 80

H006G52

A 375 R 154

3469 1464

1.8 1.7

80 78

384 98

2464 1522

1.9 1.7

80 80

H007C25

A 458 R 117

3227 2451

1.6 1.5

84 82

406 118

3342 3291

1.6 1.6

84 83

H008B22

A 70 R 50

1398 2705

1.7 1.6

80 80

68 50

1489 3341

1.7 1.6

84 84

H009B11

A 550 R 88

3344 3486

1.7 2.3

83 82

429 97

3479 7603

1.8 5.1

83 85

H010C28

A 303 R 145

3218 3089

2.6 1.7

83 80

314 264

3138 3379

5.7 1.7

85 82

H011A8

A 228 R 44

3534 3586

1.7 1.6

81 76

231 54

3964 3585

1.8 1.7

82 79

Sample#

73

PRESSURE AND ELECTRICAL CONDUCTIVITY OF KTB ROCKS

24t 2.0

Hool B2OaA

1.6

0

H00 3A4 R H00 4A5 A H005 B20 R

¢-

1.2

, ~r

H007 C25 R

X

H008 B22 A

it!

H009 B l l R H010 C28 R

[]

H011 ABA H011 A8 R

0.8

I 50

100 150 pressure (MPa)

200

250

1.4

1.2

o~

.> ,-~

Hool B20a R

1.0

~i-~ ' '~ 0

N

m

H001 E44 R H001 E44 A H003 A4 A H004 A5 R

0.8

H005 B20 A H00S G52b R

o

e,,

H006 G52b A HD07 C25 A

0.6

X

H008 B22 R H009 B l l A H01D C28 A

0.4

1

50

'

I

'

1

100 150 pressure (MPa)

'

I

200

'

1

250

Fig. 5. Pressure dependence of the electrical resistivity measured at a frequency of 1 kHz. Data are shown as normalized values as a function pressure up to 225 MPa. a) Compiles the data of the samples showing an unusual pressure effect, and b) those where an unusual pressure effect was measured.

74

G. NOVER ET AL.

Z'(ohm.m) 100

200

300

400

0

A

-100 - -

E E

0 bar

J= O N

H003 A4 R

-200 - -

.....~?-

H005 BZO A

• ~;

H005 B20 R H006 G52b R

, []

H007 C25 R HOQ8 B22 A

--"

H008 B22 R H010 C28 A H011 A8 A

-300 --

Z'(ohm.m) 0

I

100

~,~

I,_

,

200

300

400

I

I

I

-100 bar HO03 A 4 R HO05 B20 A G

-200 |

~

HO0S SZ0 R

.oo~o=,R

HO07 C25 R ,

J~

HOO8 B22 A H008 B22 R H010 C28 A hOql A S A

-300

Fig. 6. Frequency dispersion in the frequency range lkHz up to 1MHz for some selected samples: a) shows the complex data of the measurements at normal pressure and b) those collected at a pressure of 200 MPa. Data are shown as Cole-Cole diagrams where the real part (pt) of the impedance is plotted versus the imaginary part (p'). The high frequency end of the arcs is always close to the origin of the coordinate system.

PRESSURE AND ELECTRICAL CONDUCTIVITY OF KTB ROCKS

75

conductor in the rock samples as it could be due to either ore or graphite, or an increase in electrolytic conduction due to high pore pressures that open microcracks (R6ckel, 1994). The parameters of this model are included in Table II. The complex response of the samples is shown in Fig. 6a,b. Electrolytic conduction dominates the total charge transport when a strong frequency dependence was detected. In this case the arcs in Fig. 6 are nearly perfectly developed semicircles. In general these samples exhibit high volume resistivities. The volume resistivity is defined as the intersection of the low frequency end (right side) of the semicircle with the real axis. A contribution of highly conducting phases (ore, carbon) to the total charge transport can be seen in a reduction of the semicircle to only a quartercircle or even less. The volume resistivity of these samples is reduced due to the two charge transport processes that now act in parallel, electronic and electrolytic one. In Fig. 6 the symbols are used for discrimination, while the line reflects the frequency range 1 kHz up to 1 MHz. 3. Discussion

In Fig. 7 we have displayed in a drawing an area marked as KTB. The height of this area reflects the resistivity data measured on KTB samples and the width of the area is defined by the mean porosity data. The porosity was measured at normal pressure, thus it must be reduced significantly for environmental conditions and should then range around 0.3 %. These data are compared with two models describing electrical charge transport in rocks. The Archie model considers only electrolytic charge transport in the fluid pore electrolyte, while the upper bound formula (UBF) (Duba & Shankland, 1982) considers electronic charge transport by interconnected highly conducting phases like carbon. Between these two extreme models we find the KTB rock resistivities measured at normal pressure and at 225 MPa. The samples were evacuated and back-saturated using an electrolyte of a conductivity of 7.8 S / m . This electrolyte conductivity is an average for the formation waters detected in the KTB. The two processes (Archie, UBF) may both contribute in parallel to the total charge transport in KTB samples. The rock fabric analysis showed that mica is oriented in the rock (Siegesmund et al., 1993) and thereby forms the microcrack surfaces. These microcracks are still open at environmental conditions and are thus responsible for the electrolytic conduction. They are linked to the foliation and consequently are oriented. But in the foliation additionally in some of the samples ilmenite was detected in form of elongated xenomorphic grains, grown on the mica surface and orientated with their long axis parallel to the lineation. The amount of ilmenite (Table I) is sufficient to produce a high conductivity if interconnected. According to the UBF only 5 × 10 - 6 % of interconnected carbon are needed to measure a rock conductivity of about 0.1 S/m. On the other hand a electrolyte conductivity of 7.8 S/m would result in a the rock resistivity as it is shown in Fig. 7 by the two curves (using the cementation

76

G, NOVER ET AL. 100000

10000 -

~

m

= 2,5

E

J= 0 •->

.~_

1000

lOO -

"

lO

KT~

upperbound '

0

I

~

formula

'

1

electrical electric; conductivity the

of lowercontinentalcrust (Sxl0"Svol%Carbon) I

2

'

I

3

porosity

Fig. 7. This diagram shows the relation between rock resistivity and porosity. The different curves were calculated using different cementation exponents in the Archie equation. The lowest curve is calculatedusing the electrolyte conductivity at a temperature of 300 °C. The area marked (KBT) gives the range of data detected on core samples.

exponent m = 2.5 and m = 2.0 respectively) when the porosity is varied from 0.5 to 3%. Considering the temperature effect (300 °C) on the electrolyte conductivity and a pressure of 400 MPa increases the electrolyte conductivity by about a factor of 5 (environmental conditions at about 10 km in depth) and thus shifts the lower curve as it is marked by the arrow. This general feature shows that it is not possible to use only a simple model to interpret the measured pressure dependence of rock resistivities and their anisotropy. The stress relaxed cores revealed a porosity usually less than 1% and rarely exceeding 2%, and mercury intrusion methods showed a maximum in the pore radii distribution in the range of 20-50 #m representing the in-situ cracks. This finding correlates with rock fabric analysis where mainly grain boundary cracks with a low aspect ratio (Siegesmund et al., 1993) were detected. As Table I and Fig. 3 show, the permeability is high when measured in the foliation and up to three orders of magnitude lower when measured perpendicular to it. In general the axial orientated plugs (A) exhibit a lower permeability then the radial (R) ones, thus indicating the anisotropy of the permeability and confirm the findings of Huenges et al., (1990). Some of this difference in permeability is undoubtedly related to core disking. As a consequence simple geometric aspects like tortuosity, aspect ratio and the degree of pore-interconnection are of critical importance for the bulk resistivity of rocks. Additionally these parameters control the electrolytic charge- and/or fluid- transport process. The low aspect ratio of

PRESSURE AND ELECTRICAL CONDUCTIVITY OF KTB ROCKS

77

the microcracks definitely influences the electrolytic charge transport by surface conductivity. Surface conductivity is due to electrochemical interactions between the pore electrolyte and the rock matrix, and may increase the rock conductivity. It is linked to the inner surface area as it was shown by Ruffet et al., (1991). Surface area measurements were performed on KTB samples by Bticker (1990) using BET techniques. He reports Spo~ values ranging from 2 to 78 #m -1 (Spo~ is the ratio surface to porosity, Spo~ = S~ dm( 1 - O) / 0; S,~ is the inner surface area, d~ is the matrix density, O is the porosity). Cataclastic zones that are intensively fractured in microscopic scale are correlated with high Spot readings. The conductivity (o0) of rocks now was described by a modified Archie equation o-~ = ( 1IF)oral + a~l where the additional parameter crdl considers the contribution of surface conductivity to the bulk conductivity. Johnsen and Sen (1988) gave an extended equation for surface conductivity crdz that includes the influence of diffusion and double layer conductivity as well. They point out that surface conductivity must be multiplied by a geometric factor that ranges as a function of porosity from 1.5 to more than 5 for low porosities, thus indicating that surface conductivity must be considered in low porous rocks. On the other hand it strongly depends on the mobility of the ions in the double layer and the ionic strength (concentration). Thus surface conductivity is negligible for high salt concentrations in the electrolyte. (6) ~rdl is the surface conductivity, #dl the mobility of ions in the double layer, Q is the pore volume, Vp/S is the pore volume to surface area ratio. To have rough estimate for surface conductivity we have measured on different KTB gneisses and amphibolites the volume resistivity as a function of the electrolyte concentration using KC1 and NaCt solutions different in molarity and distilled water. Amphibolites as well as gneisses revealed a surface conductivity in the range 5 x 10 -5 S/m (Nover et al., 1991). Using this estimate, neither surface conductivity nor the anisotropy of the crack orientation can explain alone the measured high conductivities of the KTB samples. Brace et al. (1965) showed that an increasing external confining pressure on crystalline samples results first in a sharp increase in resistivity and at higher pressures it is followed by a linear increase. At low pressures the crack porosity having a high aspect ratio is closed, while the pore porosity (tubes) is still open even at higher pressures. This picture was verified by others (Lockner et al., 1985; Waft et al., 1974, Johnson & Mannig, 1986). However, our investigations showed quite the opposite for some KTB samples: increasing hydrostatic pressure closes fractures (as it was seen by permeability measurements), but the electrical conductivity is increased too (Fig. 5a, 5b). Thus we must assume a third effect contributing to the electrical conduction process: electronic charge transport in interconnected ore phases or in thin graphite films. As mentioned before, high concentrations of carbon were detected especially in

78

G. NOVER ET AL.

cataclastic fracture zones. In these areas high electrical conductivity was measured by borehole geoelectrics as well as in the laboratory. The findings were further confirmed by laboratory measurement of the frequency dependence of the complex electrical resistivity where a significant frequency dispersion was detected at low frequencies (10 - 2 - 1 0 2 Hz), the region where IP effects can be detected. The samples were located originally in cataclastic zones and revealed high amounts of carbon that was detected by optical inspection, thin section and polished section analysis. A high chargeability as well as a significant IP effect were measured. But the samples used for this study did not show high amounts of graphite on microcracks. So we can assume that the high conductivity is mainly due to the presence of ore phases like ilmenite. On average, ilmenite was detected to range between 1-2 vol %. This can produce theoretically a rock conductivity of 3.5 X 10 -2 S/m (1 vol % ilmenite) if the ore particles are interconnected. We did not observe such good conductivities, so we can assume that the degree of interconnection is rather low. To answer this question on the basis of Cole-Cole equivalent models, we have fitted model parameters to the measured data. These model parameters describe the conductivity of the sample using an electrolytic charge transport as well as polarizations. Second we have selected a KTB sample exhibiting a low conductivity if saturated with a highly conductive electrolyte. This sample acted as a model for an ore rich sample by placing artificially a good conductor (tiny wire) in the sample. We measured the complex response of this modified sample again, and knowing the initial RC equivalent circuit parameters, we now introduced the highly conducting phase by considering an extended RCmodel for the least-squares fit. We used this model to fit all data sets. Results are given in Table II. The frequency dependent complex resistivity data were interpreted using the Cole-Cole model described above. This model showed two general features: The complex response of the sample includes always the polarizations due to fluid solid interactions. This can be seen in Table II where an electrolytic effect is described by a nearly constant capacitor (C) and a resistor (R 1) describing the rock resistivity without any contribution of a good conducting phase (e.g. H001B20a(R), H001E44(A), H009Bll(A)). Values for R1 range from 800 ohm to 4300 ohm at normal pressure conditions. The capacitor is constant at about 1.7 #F, but R1 increases in general when pressure is applied (1500 ohm to 5700 ohm). This reflects the decrease in interconnected porosity due to the closing of cracks. A more detailed inspection of the results exhibits the observation that some samples show a decrease in R 1 though the pressure was increased (e.g. H006G52b(A)). These samples show an unusual pressure effect that can be seen additionally and more significantly in the third model parameter (R), describing a highly conducting and interconnected phase in the rock sample. All samples shown in Fig. 5b exhibit a significant decrease in R when pressure is applied. This can physically be understood by a reconnection (or higher connectivity of a fluid phase) of ore phases when microcracks are closed

PRESSURE AND ELECTRICAL CONDUCTIVITY OF KTB ROCKS

79

at elevated pressures. Sample H006G52b(A) is characterized by a strong unusual pressure effect, it has the highest amount of ilmenite and it has a low permeability. Obvious is the significant change of R1 as a function of pressure while R is nearly unchanged. Even when a high degree of interconnection of ore phases is realized (our artificial model), we still measure a frequency dispersion due to the fluid solid interactions. From this observation we conclude that a metallic conductor such as ilmenite plays a significant role in the unusual pressure effect, although there is still a contribution from the fluid phase. On this basis either the closing of cracks between metallic conductors and fluid removed from cracks are a string close as proposed by Duba et al. (1994) or the injection of fluid from core disking cracks into microcracks interconnecting ilmenite grains as proposed by Rauen et al. (1994) can be responsible for the pressure effect. The anisotropy of the conductivity can be explained considering the anisotropy of the foliation. Duba and Shankland (1982) used the upper bound formula to estimate the influence of small amounts of interconnected carbon in rock samples on the electrical conductivity in order to explain observed electrical anomalies in the upper mantle. If we extrapolate their data on the KTB pressure and temperature conditions for the final depth, then only a volume fraction of carbon of less than 5 × 10 -6 could produce a conductivity of 0.1 S/re. Magnetotelluric measurements indicated an electrically conductive (0.1 S/m) continental crust beyond the KTB in about 10 km depth. Possible models to explain this high conductive layer (HCL) are based on the assumption of highly conductive brine, good conducting ore phases or carbon (Duba et al., 1988, 1994; Glover et al., 1992, J6dicke., 1992). But to model the HCL on the basis of interconnected pores needs at least porosities > 5% if the pore saturant has the conductivity of sea water (Hyndman et al., 1989). In the KTB formation waters of a conductivity of cr~l 6.5 S/m were found (Na: 6300 ppm, Ca: 15400 ppm, CI: 37000 ppm;), but the porosity of the KTB rocks is much too low to explain the HCL only on the basis of highly conductive brine if no pore pressure is assumed. Pore pressures significant enough to maintain the equivalent of 5% porosity over a significant period of time at a depth of 10 to 12 km are difficult to comprehend however.

Acknowledgements We would like to thank E. Hinze and H. J6dicke whose comments have contributed to the improved manuscript. A1 Duba's participation was partially funded by the Alexander yon Humboldt Stiftung and by the Geoscience program office of the Basic Energy Sciences of the DOE under contract W7405-ENG-48. The research was financially supported by the Deutsche Forschungsgemeinschafl (DFG) which is gratefully acknowledged.

80

G. NOVER ET AL.

References Bemabe, Y.: 1986, 'Pore Volume and Transport Properties Changes During Pressure Cycling of Several Crystalline Rocks', Mechanics of Materials 5,235-249. Bernabe, Y.: 1988, 'Comparison of the Effective Pressure Law for Permeability and Resistivity Formation Factor in Chelmsford Granite', Pageoph 127, No.4: 607~525. B6mer, E, Jung, E and Scht~n, J.H.: 1991, 'Complex Electrical Conductivity of KTB Samples in the Frequency Range Between 10 -3 and 104 Hz - First Results', Scientific Drilling 2, 102-104. Brace, W. F., Orange, A. S. and Madden, T. R.: 1965, 'The Effect of Pressure on the Electrical Resistivity of Water Saturated Crystalline Rocks', J. Geophsy. Res. 70, 5669-5678. Bficker, C., Huenges, E., Lippmann, E., Rauen, A., Streit, K. M., Wienand, J. and Soffel, H. C.: 1990, 'KTB Pilot Hole. Results Obtained in the KTB Field Laboratory' KTB Report 90-8, D1-D29. Duba, A. and Shankland, T. J.: 1982, 'Free Carbon and Electrical Conductivity in the Earth's Mantle', Geophys. Res. Let. 9, NO.11, 1271-1274. Duba, A., Huenges E., Nover, G., Will, G. and J6dicke, H.: 1988, 'Impedance of Blackshale from Mtinsterland 1 Borehole: An Anomalous Good Conductor?' Geophysical Journal 94, 413-419. Duba, A., Heikamp, S., W. Meurer, Nover, G. and Will, G.: 1994, 'Evidence from Borehole Samples for the Role of Accessory Minerals in Lower-Crustal Conductivity', Nature 367, 6, January 1994. Figgemeier, C., Machon, L., Kamm, H., Hansmann, J., GleiB, N., Umsonst,T. and Erzinger, J.: 1992, 'KTB Hauptbohrung, Geoscientific Investigations in the KTB-Field-Laboratory, Depth interval 0-6000 m'. KTB Report, 92-2, C1-C45. Glover, E W. J. and Vine, E J.: 1992, 'Electrical Conductivity of Carbon Bearing Granulite at Raised Temperatures and Pressures', Nature, 360,723-725. Haak, V.: 1989, 'Electrical Resistivity Studies in the Vicinity of the KTB Drill Site', Oberpfalz. In. Emmermann, R. and Wohlenberg, J. (eds.). The German Continental Deep Drilling program (KTB ) - Site selection Studies in the Oberpfalz and Schwarzwald. Springer; Berlin, Heidelberg, London; Paris, Tokyo, Hong Kong; 527-553. Haak, V., Stoll, J. and Winter, H.: 1991, 'Why is the Electrical Resistivity around the KTB Hole so Low?', Phys. Earth Planet. Int. 66, 12-23. Huenges, E., Wienand, J. and Nover, G.: 1990, 'KTB Pilot Hole: Permeability Profile', KTB Report 90-8, II-I6. Hyndmann, R. D. and Shearer, P. M.: 1989, 'Water in the Lower Continental Crust: Modelling Magnetotelluric and Seismic Reflection Results', Geophys. J. Int. 98, 343-365. Jt~dicke, H.: 1992, 'Water and Graphite in the Earth's Crust - An Approach to Interpretation of Conductivity Models', Surveys in Geophysics 13, 381-407. Johnson, D. L. and Sen, P. N.: 1988, 'Dependence of the Conductivity of a Porous Medium on Electrolyte Conductivity', Physical Review B, 37, 3502-3510. Johnson, D. L. and Manning, H. J.: 1986, 'Theorie of Pressure Dependent Resistivity in Crystalline Rocks', Journal ofGeophys. Res. 91, No. B 11, 11611-11617. Konmy, A., Keyssner, S. and Friedrich, G.: 1993, 'Ore Petrology in the KTB Hauptbohrung (60007220 m)', KTB Report 93-2, 511-514. Kontny, A.: 1994, 'Bildungsbedingungen von Erzmineralen in metamorphen Krustengesteinen der KTB-Bohrungen / Oberpfalz', Dissertation RWTH Aachen, 155p. Lich, S., Duyster, J., Godizart, G., Keyssner, S. and de Wall, H.: 1992, 'German Continental Deep Drilling Program (KTB) - Geological Survey of the Hauptbohrung 0-6000 m', KTB-Report 92-2, B 1-B42, ed. R. Emmermann, Schweizerbart'sche Verlagsbuchhandlung Stuttgart. Lockner, D. A. and Byerlee, J. D.: 1985. 'Complex Resistivity Measurements of Confined Rock', J. Geophys. Res., 90, B9, 7837-7847. Nover, G. and Will, G.: 1991, 'Laboratory Measurements on KTB Core Samples: Complex Resistivity, Zeta Potential, Permeability and Density as a Tool for the Detection of Flow Phenomena', Scientific Drilling 2, 90-100. O'Brian, P. C., Rt~hr, C., Okrusch, M. and Patzak, M.: 1992, 'Eclogite Facies and a Multistage Breakdown in Metabasites of the KTB Pilot Hole, North-East Bavaria: Implications for the Variscian Tectonometamorphic Evolution of the NW Bohemian', Massif. Contr. Mineral. Petrol. 112, 261-278.

PRESSURE AND ELECTRICAL CONDUCTIVITY OF KTB ROCKS

81

Rauen, A.: 1991, 'Untersuchungen des komplexen elektrischen Widerstandes, insbesondere dessen Anisotropie und Frequenzabh~ingigkeit, von Proben des Kontinentalen Tiefbohrprogramms der Bundesrepublik Deutschland (KTB)', Thesis, Fakult~it fur Geowisenschaften der Universitat Mtinchen. Rauen, A. and Soffel, H.C.: 1993, 'Determination of Electrical Resistivity, Its Anisotropy and Heterogenity on Drill Cores - A New Method', submitted to Geophysical Prospecting. Ruffet, C., Gueguen, Y. and Darot, M.: 1991, 'Complex Conductivity Measurements and Fractal Nature of Porosity', Geophysics 56, 758-768. Schalkwijk, G.: 1991, 'Matebasites in the Pilot Borehole of the German Continental Deep Drilling Project, Windischeschenbach, Eastern Bavaria. Structures and Fabrics as Documents of Crustal Evolution'. Dissertation Universitat Bochum, 203 p. Siegesmund, S., Vollbrecht, A. and Nover, G.: 1991, 'Anisotropy of Compres sional Wave Velocities, Complex Electrical Resistivity and Magnetic Suszeptibility of Mylonites from the Deeper Crust and their Relation to the Rock Fabric', Earth Planet Science Letters 105,247-259. Siegesmund, S., Vollbrecht, A., Chlupac, T., Nover, G., Dtirrast, H., Mtiller, J. and Weber, K.: 1994, 'Fabric-Controlled Anisotropy of Petrophysical Properties Observed in KTB Core Samples', Scientific Drilling 4, 31-54. Shankland, T. J. and Waft, H. S.: 1974, 'Conductivity in Fluid Bearing Rocks', J. Geophys. Res. 79, 4863-4868. Soffel, H. C., Bticker, C., Gebrande, H., Huenges, E., Lippmann, E., Pohl, J., Rauen, A., Schult, A., Streit, K. M. and Wienand, J.: 1992, 'Physical Parameters Measured on Cores and Cuttings from the Pilot Hole Well (0 m - 4000.1 m) of the German Continental Deep Drilling Program (KTB) in the Oberpfalz area, Bavaria, Federal Republic of Germany', Surveys in Geophysics 13, 1-34. Stoll, J.: 1993, 'A Mise-a-la-masse Experiment for Detecting an Electrical Network in Cataclastic Zones around the KTB-Site', KTB Report 93-2, 361-364. Stoll, J.: 1990, 'Messung der Eigenpotentialanoamalie im KTB Umfeld und deren Interpretation', In: K. Brain (ed), KTB Report 90-3, 173-194. Trimmer, D., Bonnet, B., Heard, H.C. and Duba, A.: 1980, 'Effect of Pressure and Stress on Water Transport in Intact and Fractured Gabbro and Granite', J. Geophys. Res. 85, 7059-7071. Zoback, M. D., and Byerlee, J. D.: 1975, 'The Effect of Microcrack Dilatancy on the Permeability of Westerly Granite', J. Geophys. Res. 80 (5), 752-755.