Eur. J. Mineral.

2004, 16, 585–595

A microstructural study of a “crack-seal” type serpentine vein

using SEM and TEM techniques
MURIEL ANDREANI*1, ALAIN BARONNET**, ANNE-MARIE BOULLIER* and JEAN-PIERRE GRATIER*

* LGIT, Université Joseph Fourier, BP 53, 38041 Grenoble cedex 9, France

1 Corresponding author, e-mail: Muriel.Andreani@obs.ujf-grenoble.fr

** CRMC2-CNRS, Campus Luminy, case 913, 13288 Marseille Cedex 9, France

Abstract: Serpentine banded veins are frequently observed in massive serpentinized peridotites. They form by extension or
extensional shearing during hydrothermal alteration of peridotites. Serpentine minerals display different structural varieties, the
occurrences of which are not well defined in terms of temperature, pressure, and chemistry, but may be controlled by departure from
equilibrium and by the local water/rock ratio. Serpentines are therefore potential markers of environmental conditions during vein
formation. However, they have never been used to assess the mechanism of banded vein formation. Using multi-scale microscopy
techniques, and comparing detailed observations of natural samples from cm to nm scale with available experimental results, we
attempt to deduce constraints on growth mechanisms of serpentines in banded veins. The banded internal structure and the filling
along the vein-wall contact suggest a crack-seal mechanism of formation. Each crack is homogeneously filled with chrysotile and
some rare polygonal serpentines (tubular serpentine varieties) and disseminated patches of gel-type protoserpentine. The tubes are
not parallel to each other, but clearly show a preferred orientation perpendicular to the crack wall. Recent synthesis experiments
describe a temporal succession of occurrence of these three serpentine microstructures. The observations suggest that such an
evolution can occur in natural samples. The geometric peculiarities of macroscopic growth mechanisms in microscopic interstices
may account for capillary effects. Based on this consideration, a simple qualitative model of serpentine banded vein formation is
proposed. This model provides a possible origin for the enhancement and maintenance of a diffusional mass transfer from the matrix
to the crack. This model also predicts the very good tracking of vein opening directions in such veins.
Key-words: serpentine, microstructures, crack-seal, growth, TEM.

1. Introduction (Wiltschko & Morse, 2001). However, this mechanism is

not linked to a specific transport mechanism or cause for
The study of veins in deformed rocks is of particular interest precipitation. Transport of elements may occur by fluid ad-
because they are markers of rock deformation history, mass vection or/and by local diffusion within the pore-fluid (Fi-
transfer processes, and also a site of mineral crystallization. scher & Brantley, 1992). Crystal shape within the veins can
They can provide reliable information if their formation be stretched or elongate-blocky (Fisher & Brantley, 1992)
mechanism is well understood. Factors concerning vein for- and authors still discuss whether or not the fibrous texture
mation that need to be addressed include i) cause of opening can be attributed to the crack-seal mechanism (Cox & Ethe-
and propagation, ii) transport mechanism and iii) precipita- ridge, 1983; Urai et al., 1991; Bons & Jessel, 1997; Means
tion conditions of the vein filling material. They can all vary & Li, 2001). The cause of material precipitation in veins is
widely and the internal microstructure of veins carries an in- due to a variation of the chemical activity of the solute that
print of formation conditions. These processes are still be- can be attributed to changes in temperature, fluid pressure,
ing discussed, in particular for banded veins, and are usually differential stress...
explained by the crack-seal mechanism. Crack-seal is a The majority of the previous studies concerned quartz or
mechanism proposed to explain the formation of veins calcite veins in low-grade metamorphism environments
showing banding, made for example by wall or fluid inclu- with high fluid pressure. Although serpentine banded veins
sions. This model consists in an incremental opening fol- are frequently observed in massive serpentinized peridotites
lowed by complete sealing (Ramsay, 1980) of small succes- (Wicks & Whittaker, 1977; Dilek et al., 1997), no detailed
sive cracks. Hundreds of cracking episodes are commonly observations of their internal microstructure have been
interpreted this way. Episodic crack openings have been ex- made. Serpentine minerals display four main structural vari-
plained by oscillations in fluid pressure or bulk stress (Ram- eties: lizardite (plate-like structure), chrysotile (cylindrical
say, 1980) and more recently by a crystallization pressure structure), polygonal serpentine (tubular type with a poly-

0935-1221/04/0016-0585 $ 4.95
DOI: 10.1127/0935-1221/2004/0016-0585 ˇ 2004 E. Schweizerbart’sche Verlagsbuchhandlung, D-70176 Stuttgart
586 M. Andreani, A. Baronnet, A.-M. Boullier, J.-P. Gratier

gonized section) and antigorite (modulated structure). For 10 to 30 km depth resulted in metamorphism under HP-LT
all but antigorite, recent experiments showed that their oc- conditions before they were brought back to the surface
currence may also be controlled by departure from equilibri- (Ernst, 1971; Page, 1981).
um and by the local water/rock ratio (Grauby et al., 1998;
Normand et al., 2002). Therefore, serpentines are potential
markers of specific environmental conditions during vein 2.2. Serpentinite samples
formation. This possibility was therefore explored by com-
paring detailed observations of natural samples with avail- Partially to totally serpentinized peridotite bodies are distrib-
able experimental results in order to indicate the constraints uted throughout the Franciscan melange (Page, 1972; Page et
on growth mechanisms of serpentines in veins and to obtain al., 1998; Coleman, 2000). Samples of serpentinites have
a better understanding of the formation processes of banded been collected north of Santa Barbara, California (Fig. 1a),
veins in general. along the Santa Ynez Fault, in the Blue Canyon, where this
130 km long fault cuts across an elongated serpentinite body
(1 x 0.1 km). The structures related to the Santa Ynez Fault are
2. Geological setting and sample location strictly limited to a narrow fault gouge. Surrounding massive
serpentinites are organized in decimetric to metric blocks that
2.1. Geological setting are not affected by the recent San Andreas faulting. Thus, we
can access to markers of the serpentinite history before active
The subduction of the Pacific plate under the North America faulting. Block surfaces show striations and blue-green sli-
margin from late Jurassic to late Cenozoic resulted in the ken-fibres which suggest that they have undergone rigid rota-
formation of the Franciscan accretionary complex along the tion by sliding past one another on their surfaces during their
western coast of North America (Bailey et al., 1964; Atwa- complex history (Twiss & Gefell, 1990).
ter, 1989). In California, this complex is a mixing of de- Serpentinized peridotites show the pseudomorphic mesh
formed and variably metamorphosed sedimentary rocks and and bastite replacement textures of olivine and pyroxene re-
oceanic crustal and mantle components. Their subduction to spectively (Wicks & Whittaker, 1977) with some rare relicts
and these textures are typical of oceanic hydration. Several
generations of veins are also observed, comparable to those
a described by Dilek et al. (1997) or Stamoudi (2002) in serpen-
SA

tinized peridotites exposed on the seafloor of the Mid-Atlantic

F

Ridge (MARK area) and the East-Pacific Ridge (Hess-Deep).

Thus, they could form during oceanic hydrothermal history of
the past Pacific oceanic crust. A 3D-network of planar blue-
SYF 34°30'
Pacific
Ocean SB Blue
50 km Canyon
119°35'
Quaternary
Plio-Pleistocene SB = Santa Barbara
SAF = San Andreas Fault
Fig. 1. Location and structural sketch of
Neogene the studied vein. a: Geological map of
SYF = Santa Ynez Fault
Jurassic to Paleogene California, north of Santa Barbara (S.B.)
Franciscan Complex Major faults and the major faults of the San Andreas
Serpentinites and ophiolites
system. Massive serpentinized peridotites
are adjacent to the Santa Ynez Fault
(S.Y.F.), along the Blue Canyon. b: Pic-
veins: v1 v0 v1 ture of a massive serpentinized peridotite
v3 core v0 block at the outcrop. A schematic close-
v1 v0
v2 up of the different generations of veins is
v2
5

v0 rim drawn on the upper-left corner. The chro-

cm
0

nological order of the veins is as follows:

v0 & v1: primary serpentine veins, v1 is a
v3 fibrous vein; v2: banded serpentine vein;
Fig.2 v3: later isotropic serpentine vein. c:
Structural diagram of the banded vein (v2)
developing perpendicular to the block sur-
Core face. The double cross pattern represents
the serpentinized peridotite matrix. Dot-
ted lines accross v2 follow optical “trails”.
d: Geometry of the vein tip. This indicates
Rim 1 mm that increments are preferentially younger
b d c
from the right to the left.
 Crack-seal type serpentine vein 587

banded veins is developed both along and normal (in two

perpendicular directions) to the block limits (Fig. 1b). The
diagram in Fig. 1c presents the geometry and structure of a
vein in a plane perpendicular to the vein walls. This vein
crosscuts a massive serpentinite block and has been studied
in detail. Vein width varies from few µm to several mm. This
type of vein has been interpreted as the result of incremental
stress release during unroofing of serpentinized peridotite in
the case of the MARK area (Dilek et al., 1997). This is in
good agreement with our observations which indicate a syn-
tectonic vein formation in extension and/or extensional
shearing, that accommodates progressive separation and ro-
tation of relatively undeformed neighbouring serpentinite
blocks (Fig. 1b). Nevertheless, the relative 3D-radial geom-
etry of massive serpentinized blocks and vein arrays sug-
gests that these fractures could also accommodate a volume
increase of the peridotite during its hydration to serpentini-
te, as suggested by O’Hanley (1992).

3. Microscale observations of the banded

internal fabric of the vein
3.1. Optical microscopy: vein texture and morphology

A segment of the vein is pictured on Fig. 2a and 2b. The in-

ternal structure is revealed only under crossed polars. This
type of vein has a variable birefringence that can reach ab-
normal 2nd order colours for a thin section of 30 µm thick-
ness. Two types of structure, similar to those described as
bands and trails by Ramsay (1980), can be observed. They
are drawn on Fig. 1c. First, a finely spaced banding is ob- Fig. 2. Serpentine banded veins observed under crossed polars. a:
served parallel to vein margins. The bands are separated by This low magnification image of a vein shows that the bands are par-
allel to the walls and perfectly mimics their morphology. The mar-
lines of low birefringence and their irregular shape perfectly
gins can be superimposed on each other and linked by the trails. b:
mimics the shape of the vein margins that can be superim-
Enlargement of Fig. 1c. Bands are parallel to the NE-SW axis of the
posed on each other (Fig. 2a). The geometry of the margins picture and trails are underlined by the dotted line. i: wall rock inclu-
are thus preserved across the vein width, except when wall sions.
fragments have been included in the vein filling (Fig. 2b).
Secondly, the simultaneous extinction of zones perpendicu-
lar to the banding defines the trails which link edge irregu-
larities across the vein width (Fig. 2a). No inclusions or sec- 3.2. SEM: organisation of vein filling in space
ond phases are observed within these trails so they are only
extinction figures. The minimal width of one band is close Scanning Electron Microscopy (SEM) gives the geometry
to 1µm and the maximum observed is less than 5 µm. Statis- of vein filling at an intermediate scale between micrometric
tical analysis reveals a characteristic width of 3 µm and a (thin section) and nanometric (TEM) scales.
random succession of band widths (Renard et al., in press). The massive sample containing the banded vein described
Crack propagation is not guided by matrix texture except above was broken perpendicularly to the vein, cleaned in an
when cutting through bastites where it preferentially fol- ultrasonic bath, then glued on a support and covered with a
lows the primary {110} cleavage planes of pyroxene. At 10-15 nm thick carbon coating. SEM imaging was per-
grain-scale (hundreds of nm), lizardite being the main mesh formed on a JEOL 6320FEG scanning electron microscope
component, the opening path might be locally guided or at at the CRMC2 (CNRS, Marseille, France).
least helped by the weak (001) plane of lizardite. The inter- At low magnification (Fig. 3a), individual bands are al-
nal structure of the vein is often asymmetric close to the tip, ready visible. Smallest units range from 1 to 4 µm in thick-
as drawn on the simplified scheme of the tip on Fig. 1c. ness. Each band shows the same internal structure: a homo-
Bands tend to be discordant on the right vein margin and geneous filling of tubular-shaped serpentine crystals with a
concordant with the left one. preferred orientation perpendicular to banding (Fig. 3b and
The grain size of serpentine crystals is usually submicronic 3c). Successive bands are separated by a plane marked by
and cannot be observed with a petrographic microscope, so the termination of serpentine tubes, along which fracturing
electron microscopy is needed to identify the serpentine type due to sample preparation is localized. Locally, they appear
and to understand the features seen in optical microscopy. to be embedded in a poorly crystallized phase (Fig. 3c). De-
588 M. Andreani, A. Baronnet, A.-M. Boullier, J.-P. Gratier

Fig. 3. a: Low magnification SEM image of

vein filling (area indicated in Fig. 2b). Note the
irregular thickness of bands ranging from 5 µm
to less than 1 µm. b, c: Enlargement on bands.
Tube diameters are around 100 nm and a residu-
al porosity exists between them. Within one
band, the tubes seem to be locally glued in a
shapeless, poorly crystallized phase (P1: cir-
cled). d: Image of the vein-wall contact. It is
marked by a layer filled with a tubular serpen-
tine and a poorly crystallized phase.

spite the high density of tubes, a residual porosity remains be- of tubes (Fig. 4b). It was also found that, on the same milled
tween them. Unfortunately, owing to the sample geometry, it sample, areas with a majority of longitudinal sections
is not possible to measure the tube length, nor to check wheth- (Fig. 4b) tend to alternate with zones containing longitudi-
er one tube necessarily spans a whole band. The vein-wall in- nal, oblique and cross-sections (Fig. 4a).
terface has also been investigated (Fig. 3d). Along the left Two varieties of tubular serpentine minerals are identified
wall of the vein on Fig. 1c and 1d (the youngest increment), a (Fig. 5). Cylindrical tubes with the smallest diameter
µm-size interstice is filled with the “gel-type” (poorly crystal- (< 100 nm) predominate and are chrysotile (P2 phase) (Baron-
lized) phase in which very small tubular-shaped crystals can net & Devouard, 1996). The second type of tube is shorter and
be distinguished. Round and smooth pores within this materi- shows polygonized sections and a larger diameter (>100 nm).
al may be interpreted as primary porosity. This layer at the This corresponds to polygonal serpentines (P3 phase) (Zuss-
vein-wall interface could represent the last addition of a new
band, thus defining an antitaxial vein filling.
Only TEM observations can allow a rigorous identifica-
tion of serpentine microstructures in this vein and specify
their mutual arrangements.

3.3. TEM: identification of serpentine microstructures

and their intergrowths
TEM images were taken with a JEOL 2000fx high-resolution
transmission electron microscope at the CRMC2 (CNRS,
Marseille, France), with a 200kV accelerating voltage. A thin
section, attached with Lakeside® resin to the glass substrate,
was cut perpendicular to the banded vein. A single-hole cop-
per TEM grid was glued on a selected area of the section. The
specimen was extracted from the thin section by heating the
resin after drilling around the grid. Prior to observation, speci-
200 nm a b 200 nm
mens were thinned by ion-beam milling (Precision Ion Polish-
ing System – Gatan 690) and carbon-coated. Fig. 4. TEM images of the vein-filling.An alternance of the two fol-
The banding direction was determined under a petro- lowing microtextures is observed. a: TEM image of a zone with longi-
graphic microscope before TEM observation. Banding is tudinal, oblique and cross sections of chrysotile and polygonal serpen-
not as evident as on SEM images, probably because of the tine. The lack of a preferred orientation of tubes should result in a low
milling which can result in a slightly inclined section com- birefringence under crossed polars that can correspond to the limit be-
pared to banding, and to an extremely reduced surface of ob- tween bands. b: TEM image of a zone with a majority of longitudinal
servation (few µm) around the hole made by ion-beam mill- sections. This should correspond to the part of the bands displaying a
ing. Nevertheless, images confirm the preferred orientation higher birefringence.
 Crack-seal type serpentine vein 589

per formula unit (a.p.f.u.) calculated on the basis of seven

P1 oxygens (Table 1).
The chrysotile-polygonal serpentine vein cuts across a
massive serpentinized peridotite matrix showing a pseudo-
morphic mesh and bastite texture. Mesh and bastite textures
P3 in serpentinized peridotites have been intensively investi-
gated and lizardite is the dominant serpentine variety in the
matrix, plus chrysotile and oxides (e.g. Coleman & Keith,
1971; Dungan, 1979; Viti & Mellini, 1998). The concentra-
tion of major cations (Mg, Si, total Fe, Al) is relatively con-
stant in all samples with a general low amount of Al. Within
the matrix, small variations in Ni and Cr contents are noted.
They can be explained by the various nature of the precursor
P2 mineral. Indeed, preferential replacement of Mg by Ni and
Cr may occur to a small extent in serpentine when it replaces
olivine (mesh texture) and pyroxene (bastite), respectively
P1 (Dungan, 1979; Viti & Mellini, 1998). Cl and F are seldom
detected but can appear locally, in very small amounts.
At EMPA scale, vein composition is highly homoge-
neous for all elements and is independent of distance from
the edge. Ni and Cr contents are very low, close to the detec-
tion limit, as are Cl and F that appear locally.
200 nm

Fig. 5. TEM image of the vein-filling minerals. Three serpentine 4.2. Microanalysis of serpentine structural types
types can be identified: P1: protoserpentine – gel-type, poorly crys-
tallized phase; P2: chrysotile – cylindrical structure with a diameter
The sample prepared for TEM in section 3.3 (Fig. 5) was
< 100 nm; P3: polygonal serpentine – tubular structure with a poly-
used again. The details of its preparation are explained in
gonized section and a diameter usually > 100 nm. section 3.3. It contains the three different types of serpentine
(protoserpentine, chrysotile and polygonal serpentine) en-
man et al., 1957; Devouard et al., 1997). A third serpentine countered in the vein. Each serpentine type was analysed
phase is poorly crystallized (P1 phase): the protoserpentine. It separately with a transmission electron microscope JEOL
has seldom been found in natural samples but has been ob- 2000fx equipped with a Tracor Northern 5520 X-ray energy
served in synthesis experiments (Grauby et al., 1998). Fig. 5 dispersive system, working under a 200 kV accelerating
illustrates a patch of protoserpentine from which small tubes voltage. Diameters of analysed surfaces are around 30 nm.
seem to grow radially. Indeed, protoserpentine contains few Cliff-Lorimer K-factors ki,Si with i=Mg, Fe, Al, Cr, Ni, were
nm-sized particles that seem to roll up. calibrated against natural and synthetic standards. All the
analyses were run on the same day. Serpentines are not very
resistant under the electron beam, especially the less orga-
4. Chemical composition nized types of protoserpentine. This constraint has limited
the number of analyses. Compositions in atom per formula
Serpentine minerals, except antigorite, are trioctahedral 1:1 unit (a.p.f.u.) are reported in Table 2.
sheet silicates represented by the ideal formula Results are comparable to those obtained by electron mi-
Mg3Si2O5(OH)4. The principal substitutions are FeMg-1 and croprobe. Small variations in composition can be noted
(Al, Fe3+ )VI (Al, Fe3+ )IV Mg-1VI Si-1IV. Measurements were within one type of serpentine but the mean compositions of
done in order to test for chemical variations: (i) between the each serpentine type are very similar. As usual in natural
vein and the host rock, (ii) across the vein (e.g. during vein samples, the structural formulae of serpentines differ slight-
formation) and (iii) within one band (between the different ly from the ideal one. 15 to 20 % of Mg is substituted by Fe
structural type). which is commonly distributed among the octahedral sites.
In this case, the fairly high Fe content for all species can to-
4.1. Electron microprobe analysis (EMPA) tally compensate for the lack of Mg. Al content is very low
(< 0.1 apfu) as previously observed in EMPA. No chemical
Analyses within the matrix and the vein were performed differences between serpentine types are observed.
with a Cameca SX100 electron microprobe equipped with
four wavelength-dispersive spectrometers (WDS) at the
DSTU- Montpellier (France). Spot size (~3µm) is similar to 5. Interpretation and discussion
the mean band thickness, which does not allow comparison
5.1. Formation of serpentine banded veins: a crack-seal
between successive bands, but by taking several measure-
mechanism
ments scattered across the vein it is possible to detect any
large scale variations. Compositions are reported in oxide- Formation of veins can be split into three steps as follows: 1)
weight % (wt%) along with the structural formulae in atom crack opening, driven by fluid pressure, stress or crystalliza-
590 M. Andreani, A. Baronnet, A.-M. Boullier, J.-P. Gratier

Table 1. Electron microprobe analyses of serpentine.

Analysis #14 #16 #18 #54 #55 #57 #61 #56 #34 #42 #50
Location mtx mtx mtx mtx mtx CS-15µm CS-15µm CS-15µm CS-45µm CS-105µm CS-220µm
Compositions in wt% of oxides
SiO2 40.9 40.9 42.8 41.8 41.3 40.3 40.3 40.6 40.1 40.1 40.9
MgO 37.7 38.1 38.6 37.0 37.9 37.5 37.4 37.2 37.2 37.5 36.2
FeO 3.4 3.9 3.6 4.7 4.6 5.0 5.3 5.0 5.3 5.1 5.3
Al2O3 0.3 0.4 0.1 0.2 0.2 0.2 0.3 0.3 0.3 0.3 0.3
NiO 0.4 0.2 0.0 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1
Cr2O3 0.1 0.1 0.0 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2
SO2 0.3 0.2 0.1 0.1 0.1 0.2 0.2 0.2 0.2 0.1 0.1
Cl 0.3 0.3 0.0 0.1 0.1 0.2 0.1 0.2 0.1 0.2 0.2
F 0.0 0.1 0.0 0.1 0.0 0.2 0.0 0.0 0.3 0.1 0.3
Total 83.5 84.5 85.4 84.3 84.5 83.9 83.9 83.8 83.9 83.7 83.5
Stuctural formulae in atom per formula unit (apfu)
Si 2.01 1.99 2.04 2.04 2.01 1.99 1.99 2.00 1.98 1.98 2.02
Mg 2.76 2.77 2.75 2.68 2.75 2.75 2.75 2.73 2.73 2.76 2.67
Fe 0.14 0.16 0.14 0.19 0.19 0.21 0.22 0.21 0.22 0.21 0.22
Al 0.02 0.02 0.01 0.01 0.01 0.01 0.02 0.02 0.02 0.02 0.02
Ni 0.01 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.00 0.00
Cr 0.00 0.00 0.00 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01
S 0.01 0.01 0.00 0.00 0.00 0.01 0.01 0.01 0.01 0.00 0.00
Cl 0.02 0.02 0.00 0.00 0.00 0.01 0.01 0.02 0.01 0.02 0.01
F 0.00 0.01 0.00 0.02 0.00 0.02 0.00 0.00 0.05 0.01 0.04
Total 4.98 5.00 4.95 4.96 4.98 5.01 5.00 4.99 5.03 5.01 4.99
Analysis within the matrix: mtx – Analysis within the banded vein: CS-distance to the matrix

Table 2. EDS-TEM microanalyses of serpentine types.

Analysis #1 #2 #3 #4 #5 #6 #7 #8 #9
Serp. type proto.s proto.s proto.s ctl ctl ctl polyg.s polyg.s polyg.s
Stuctural formulae in atom per formula unit (apfu)
Si 1.95 1.86 1.90 1.84 2.00 1.95 2.05 1.96 1.79
Mg 2.80 3.02 2.93 2.97 2.59 2.79 2.65 2.80 3.21
Fe 0.24 0.21 0.16 0.24 0.27 0.22 0.22 0.19 0.17
Al 0.02 0.00 0.06 0.06 0.09 0.05 0.02 0.04 0.02
Cr 0.01 0.00 0.00 0.01 0.00 0.01 0.00 0.01 0.00
Ni 0.00 0.00 0.00 0.00 0.01 0.00 0.00 0.00 0.00
Total 5.02 5.10 5.07 5.12 4.95 5.02 4.94 5.01 5.20
Mean value:
Serp. type proto.s ctl polyg.s
Si 1.91 1.93 1.93
Mg 2.92 2.78 2.89
Fe 0.20 0.24 0.19
Al 0.03 0.07 0.03
Cr 0.00 0.01 0.00
Ni 0.00 0.00 0.00
Total 5.06 5.03 5.05
Proto.s: protoserpentine or P1 phase; ctl: chrysotile or P2 phase; polyg.s: polygonal serpentine or P3 phase.

tion forces; 2) transport of vein forming elements by advec- 5.1.1. Crack opening and propagation
tion and/or diffusion; 3) crystallization of the vein minerals.
Interesting questions concern the origin and preservation of Banding can have different origins: mineralogical alter-
supersaturation within the crack and the capability of inter- nances, chemical variations, fluid or rock inclusions... Thus,
nal microstructures to track the displacement history. These banding can be formed by different mechanisms. In the stud-
points are discussed in the following chapters for the partic- ied serpentine banded vein, no chemical variations have been
ular case of the serpentine banded vein. observed across the vein width. SEM images clearly show
 Crack-seal type serpentine vein 591

that a physical and morphological discontinuity exists be- interconnectivity between cracks required for fluid circula-
tween bands. The limit between successive bands is only tion and advection is not always observed. There are no sig-
marked by a discontinuity in the growth of minerals. This nificant variations in major elements either within the vein,
suggests that, in this case, banding can be interpreted as a re- or between the host rock and vein mineral compositions.
sult of individual crack openings, as proposed by the crack- This indicates that there is no chemical change in space and
seal process of Ramsay (1980). Oscillations in fluid pres- time during vein formation, thus suggesting a constant
sure and/or local stress are usually thought to open the inter- source of matter and no chemical segregation. Others analy-
stice corresponding to each band (Ramsay, 1980). In the ses realized on this type of vein confirm that the composi-
particular case of the serpentine banded veins, two possible tion of the vein is sensitive to the composition of the adja-
causes of cracking were proposed, as pointed out in section cent wall rock (Andreani, 2003). Therefore, the nutrient
2.2: i) incremental stress release due to progressive unroof- source may have been the immediately adjacent matrix.
ing of serpentinized blocks (Dilek et al., 1997), or ii) accom- Chemical microanalysis shows that vein filling and trans-
modation of volume increase during serpentinization of the formations during the potential protoserpentine-chrysotile-
peridotite (O’Hanley, 1992). The field evidence and the ana- polygonal serpentine sequence occur isochemically. Taken
lytical results are, at this stage, insufficient to discuss this together, these results suggest a closed-system behaviour
point. Thermal expansion has not been considered since it implying small-scale transport of elements in a fluid by a
should not be significant at this range of temperatures diffusional process that may have taken place through the
(< 500˚C). walls of the vein thanks to their porosity. A chemical poten-
The direction of serpentine fibre growth varies in differ- tial gradient due to a pressure, temperature, stress or chemis-
ent directions around a preferred orientation. This internal try gradient is required to establish diffusion and needs to be
microstructure can only occur in a fluid-filled open-crack maintained until total filling of the crack. Supersaturation
where geometrical selection occurs during growth. The should also steadily occur in the interstice to allow mineral
preservation of the wall geometry (Fig. 2a) during vein for- growth. These points are examined in the following section.
mation is possible only if sealing of a crack is achieved Macroscopic crystal growth theory commonly uses a
before the opening of the following one. This is consistent capillary model based on the analogy between the liquid/va-
with the model proposed by Hilgers et al. (2001). pour and crystal/solute systems (Mullins & Sekerka, 1963).
The presence of a poorly crystallized layer at the vein- Crystal growth is only possible when it results in a decrease
wall contact (Fig. 3d) from which tubes protrude suggests in Gibbs free energy G of the system. Hence the favourable
that addition of a new band occurs at the vein-wall limit, variation of G may be written as:
thus defining an antitaxial vein filling. According to the asy-
metric geometry of the vein tip (Fig. 1c and 1d) and SEM 2 G = 2 Gvol + 2 Gsurf < 0 (1)
observations, it seems that cracking preferentially occurs at
one vein-wall interface and propagates in mode I. The shape where 2 Gvol = –n. 2 µ is the volume free energy difference,
of the vein margins is irregular and the morphology of the with n the number of growth (or dissolution) units forming
first crack is preserved during vein formation since each the crystal and having crossed the crystal/medium interface,
new crack follows the surface of the preceding one (Fig. 1c and 2 µ the chemical potential difference undergone by one
and 2a). There is no evidence of margin dissolution or catac- of those growth units.
lastic deformation resulting from crack opening. This has 2 Gsurf = + * A is the crystal surface contribution to this
been proposed by Robert & Boullier (1994) to be in favour Gibbs free energy change equal to the product of the crystal
of a slow or sub-critical propagation mode. Different mech- surface tension * (assumed to be isotropic here) multiplied
anisms are considered to accommodate such a steady-state by the crystal surface area A.
mode of propagation (Atkinson, 1984) but too few indices Two different cases for nucleation are possible: homoge-
remain in this final deformation state to allow conclusions. neous versus heterogeneous. In nature, nucleation is predomi-
Complete filling of the crack by precipitating minerals al- nantly heterogeneous because of the availability of preferen-
lows stress transfer through the vein and the opening of an- tial nucleation sites which lower the nucleation activation en-
other crack at the vein-wall interface. Indeed, the different ergy barrier. In this way, the convex nucleus decreases its area
texture across the vein-host rock interface makes this site a in contact with the solution lowering, thereby, the surface en-
weak area. ergy. This benefit is all the greater if the crystal-chemical af-
finity between the crystal and support is large. However,
2 Gsurf remains positive so that we should still have 2 Gvol < 0,
5.1.2. Crack filling and capillary effect i.e., 2 µ = kTLn(a/aeq) > 0 for crystal growth, where k is Bolt-
When the µm wide crack is opened, it is filled by a fluid to zman’s constant, T the absolute temperature, a the actual sol-
allow mineral precipitation. Then a mass transfer and a lo- ute activity, and aeq the equilibrium activity or solubility at T.
calized precipitation in the interstice is needed to explain the Accordingly, crystal growth is only possible if a/aeq > 1 that is,
vein filling. if the solution is supersaturated.
Given the very low solubility of phyllosilicates, transport Baronnet & Saül (2003) examined theoretically the con-
of nutrients by advection through the fracture network ditions of crystallization in very tight interstices (around µm
would require huge amounts of circulating fluid to fill each and below). They showed that tight interstitial opening fa-
interstice totally with minerals, as Fisher & Brantley (1992) vours crystallization in two ways: i) the simultaneous con-
already concluded for quartz-filled veins. Furthermore, the tact of the nucleus with both matrix walls decreases twice
592 M. Andreani, A. Baronnet, A.-M. Boullier, J.-P. Gratier

the surface energy for the supported nucleus compared to tine and should precipitate first in the interstice as soon as it
the free nucleus, ii) good capillary wetting induces a con- opens and fills with a fluid (Fig. 6, stage 2, Phase P1). Ob-
cave crystal/solution meniscus. Effect ii) is important in that servations and chemical analyses do not support a large-
it depresses the crystal solubility just as a sub-saturated scale fluid circulation, so a variation in serpentine solubility
pressure exists within a liquid in case of capillary condensa- in the interstice can hardly be related to a variation in fluid
tion. The main point is that surface tension, while pulling the temperature. A pressure drop linked to a micrometre crack
condensed phase, may now act to promote crystallization as may briefly help nucleation but would not significantly de-
2 Gsurf < 0 is now possible. In the latter case, 2 G < 0 may be crease solubility to establish and maintain such a departure
obeyed over a certain range of 2 Gvol > 0, or for 2 µ < 0, i.e., from equilibrium. The interstice capillary effect described
for nominally undersaturated solutions (a/aeq < 1). The above would allow the system to diverge considerably from
range of undersaturation concerned increases with the inter- equilibrium as soon as the crack opens, without the need to
stice narrowness and with the wetting capability of the crys- have a highly supersaturated fluid within the crack. Low
tal/support pair. This situation of possible crystal growth in surface tension between nucleus and nucleation site, a pre-
undersaturated solutions may be seen as analogous, but op- requisite for the model, is expected because precipitated
posite, to that of Oswald ripening where the capillary effect minerals and substrates are both serpentines (high structural
makes particle dissolution possible in a nominally supersat- and chemical affinity). This preliminary form, protoserpen-
urated solution. tine, can then evolve into chrysotile tubes with residual po-
The driving force for diffusion of the serpentine material rosity due to loss of water by this preliminary “gel-type” ser-
from the wall to the crack is related to the difference be- pentine (Fig. 6, stage 3). Chrysotile would preferentially nu-
tween the convex grain surface in the matrix and the con- cleate against the most similar material, i.e., the one it has
cave solid/solution meniscus in the interstice. Indeed, the the most affinity with, and thus against the surface of the
Gibbs-Thomson relationship shows that the solubility of a crack opposite the vein wall. In this way, the arrival of nutri-
crystal is inversely proportional to its size. So the activity of ents from the matrix is not stopped by crack filling and
the solute in the matrix would be enhanced by the small should be able to fill each narrow space completely (Fig. 6,
grain-size of the matrix material (normal capillary effect on stage 4).
solubility), and possibly also by the supersolubility induced
by stresses on matrix grains (pressure solution, Weyl, 1959).
Contrasting with this, the equilibrium activity inside the
vein interstice would be depressed by the above-mentioned
reversed solid/solution meniscus effect. The activity gradi- 1 3
ent responsible for solute diffusion from the matrix to the in- aqueous fluid
terstice could result from such physical contrast between the
matrix, acting as the solute source, and the vein interstice, Crack opening Chrysotile crystallization
acting as the sink. During progressive filling, such activity
gradient and diffusional process would even increase owing
to the narrower and narrower opening until complete clo- 2 4
sure. This process depends on the presence of an interstice
and should thus stop when the crack is filled. However, its
action might be resumed if the seeping capability of the wall Protoserpentine formation Complete filling of the crack
for solute inflow were to be refreshed by successive cracks. Serpentinized peridotite Chrysotile nanotubes
Observations show that three different structural types of Protoserpentine
Solutes
serpentine coexist in each crack: protoserpentine, chrysotile
and polygonal serpentines. Hydrothermal synthesis of ser- Fig. 6. The different steps of the inferred model of serpentine banded
pentines at 300°C and 200°C by Grauby et al. (1998), from vein formation. It includes four main steps: 1) The µm sized crack
a stoechiometric (excess H2O) gel revealed a temporal suc- opens and is filled with an aqueous fluid. It is accompanied by a
cession of serpentine types: protoserpentine-chrysotile-po- small pressure and solubility drop. 2) Protoserpentine (or quasi iso-
lygonal serpentine and much later lizardite (platy structure). tropic gel-type serpentine) forms locally because it is the kinetically
This isochemical evolution has been attributed to the de- favoured metastable phase under such conditions. The crystal forms
creasing degree of supersaturation of the solution with re- an inverse solid/solution meniscus with the two walls of the inter-
spect to serpentine. Normand et al. (2002) reached the same stice. This depresses the solubility within the crack. The difference
conclusion for the relative appearance of chrysotile and li- between the high activity around the small matrix grains and the
lower activity in the interstice creates a diffusional transport of ele-
zardite. Moreover, it has been suggested that polygonal ser-
ments that feeds the crack and helps the protoserpentine to propagate
pentine, which has a larger diameter than chrysotile, is a
laterally. 3) Diffusion can be maintained until total filling of the
more evolved tubular form of serpentine (Baronnet et al., crack. The chrysotile initially nucleating on the last crack can there-
1994). fore progressively replace the protoserpentine. Chrysotile can then
Our observations may be indicative for the occurrence of evolve to a more stable phase, the next one being polygonal serpen-
the same succession also in natural samples. A system out of tine. 4) The crack is filled with a majority of chrysotile and some po-
equilibrium would classically favour the precipitation of ki- lygonal serpentines. Some patches of protoserpentine may remain.
netically controlled (metastable) phases first. The most The diffusional process stops until the system is refreshed by a new
metastable phase here (the most soluble one) is protoserpen- crack opening.
 Crack-seal type serpentine vein 593

Wall Vein wall

Vein
5
4
Incremental
3 opening
2 directions
i 1
Vein
Local extinction of tubes
under cross-nicols.
a b Mean orientation of tubes

Fig. 7. a: Vein-wall contact under crossed polars. White arrows show margin irregularities. They correspond to the starting points of crystallo-
graphic trails parallel to the dotted line across the vein. A wall fragment (i) locally modifies the orientation of the trails. b: Diagram explaining
the formation of trails during banded vein formation. The trails are crystallographic since they correspond in each band to the local extinction
of the tubes. By following these extinctions, it is possible to join the morphological irregularities of the two walls and to track the opening di-
rection.

Chrysotile grows along the [100] fibre axis and under- polars, an inhomogeneous extinction across the vein links
goes a progressive geometric segregation leading to the pre- edge irregularities regardless of banding (Fig. 2a and 2b) as
ferred orientation of tubes. Tube zones of higher disorder described in part 3.1. The extinction position of serpentine
(Fig. 4a) should therefore appear first and mark the begin- fibres is parallel to their long axis. No deformation has
ning of filling. Growth competition between tubes would been observed, so banding and extinction are only due to
then lead to ordered zones (Fig. 4b). If the solute concentra- crystallographic orientation during growth. Within one
tion remains high enough, the chrysotile growth, limited by band, fibres are not parallel to each other but they show a
a critical radius of lattice curvature (energy storage con- preferred orientation that is perpendicular to the limit of
straint), would progressively evolve to polygonal serpentine the last increment. Extinction occurs when the band limit is
(Baronnet et al., 1994; Devouard & Baronnet, 1995; Baron- parallel to the N-S or E-W axis of the microscope under
net & Devouard, 1996). crossed polars (Fig. 7a) and, therefore, fibre orientation is
not displacement-controlled. The separation line between
bands is always close to extinction to a certain degree since
5.2. Origin and efficiency of the tracking capability of it represents the limit between the end of a band where the
the vein opening directions tubes are well oriented (Fig. 4b) and the beginning of a new
one marked by remaining protoserpentine patches and dis-
Syntectonic veins are potential structural markers of strain oriented tubes (Fig. 4a) (low birefringence material). Nu-
directions (e.g. Durney & Ramsay, 1973; Passchier & Tro- cleation of the first chrysotile tubes would be easier on the
uw, 1996). Only fibrous and crack-seal veins assumed to margin of the preceding crack since it is filled with the
form progressively or by opening increments have this po- same serpentine material and some seeds may remain. So
tential. The direction of fibre growth or crystal elongation the tubes will more probably form perpendicular to the pre-
has long been considered to be a direct indication of stretch- ceding crack than to the wall-rock surface. If the crack sur-
ing direction during vein formation. However, it has also face has asperities, fibre orientation will follow them and
been proved that they do not necessarily indicate opening the extinction will not occur simultaneously in one band
direction (Van Der Pluijm,1984; Cox, 1987; Williams & Ur- (Fig. 7a and b). Consequently, an oblique opening incre-
ai, 1989; Urai et al., 1991). These particular internal micro- ment will be recorded by the displacement of the local ex-
structures have been intensively studied in order to deter- tinction in the new band. The dotted line in Fig. 2b follows
mine the conditions of a real tracking capability. The inter- these local extinctions in successive bands across the vein
nal structure of crack-seal veins commonly shows elongate and have a similar signification as inclusion trails (Ram-
blocky or stretched crystals. The opened crack is assumed to say, 1980) since they draw the opening vector of the vein.
be filled unitaxially (in one direction) by overgrowth of an- Wall rock inclusions are rare but, when observed, they con-
cient vein crystals exposed to the fluid-filled fracture. The firm the displacement paths determined by the crystallo-
relative influence of different parameters on the tracking ca- graphic trails.
pability of antitaxial veins has been investigated through nu-
merical modelling of crystal growth kinetics (Bons, 2001;
Hilgers et al., 2001). They lead to the conclusion that open- 6. Conclusions
ing directions are best recorded by the fibrous crystal long-
axis if the fracture is tight, its surface is rough and if it is to- The characteristics of serpentine banded veins at different
tally filled before the next opening increment. scales suggest a crack-seal type process of formation. Vein
In the present crack-seal serpentine veins, fibrous min- morphology and chemical analysis favour diffusional trans-
erals are not distinguishable optically but, under crossed port of nutrients from the adjacent matrix to the vein.
594 M. Andreani, A. Baronnet, A.-M. Boullier, J.-P. Gratier

The three different serpentine types observed in each dersaturated solutions: a model and geological applications.
crack: protoserpentine-chrysotile-polygonal serpentine, are EGS-AGU-EUG, Nice, Abstract P0308.
described together for the first time in natural samples, and Baronnet, A., Mellini, M., Devouard, B. (1994): Sectors in polygo-
may be compared to the succession described in synthesis nal serpentine. A model based on dislocations. Phys. Chem. Min-
experiments during a decreasing supersaturation of an iso- erals, 21, 330-343.
chemical system. Bons, P.D. (2001): Development of crystal morphology during unit-
Based on these observations, we propose a model of for- axial growth in a progressive widening vein: I. The numerical
mation for serpentine banded veins that could explain: 1) the model. J. Struct. Geol., 23, 865-872.
Bons, P.D. & Jessell, M.W. (1997): Experimental simulation of the
high departure from the equilibrium required for the precipi-
formation of fibrous veins by localised dissolution-precipitation
tation of protoserpentine and 2) the enhancement and the
creep. Min. Mag., 61, 53-63.
maintaining of a steady-state diffusional process for vein Coleman, R.G. (2000): Ultramafic rock (serpentine) in California
filling. This model takes advantage of capillary effects on and Oregon. Standford Geological Survey.
crystallization that can occur in micrometre-size interstices. Coleman, R.G. & Keith, T.E. (1971): A chemical study of serpentini-
Further observations are needed but this study supports the zation – Burro Mountain, California. J. Petrol., 12, 311-328.
view that serpentine microstructures may be able to record Cox, S.F. (1987): Antitaxial crack-seal vein microstructures and
their environmental growth conditions. Thus they can act as their relationship to displacement paths. J. Struct. Geol., 9, 779-
a new tool for understanding crystal growth conditions un- 787.
der dynamic regimes. Cox, S.F. & Etheridge, M.A. (1983): Crack-seal fibre growth mecha-
Trails, resulting from an inhomogeneous extinction of nisms and their significance in the development of oriented layer
bands under crossed polars, track the opening direction silicate microstructures. Tectonophysics, 92, 147-170.
across the vein. Serpentine banded veins have a very good Devouard, B. & Baronnet, A. (1995): Axial diffraction of curved lat-
tracking capability if the surface of the initial fracture is suf- tices: geometrical and numerical modeling. Application to chrys-
ficiently rough compared to fibre size. otile. Eur. J. Mineral., 7, 835-846.
Development of such veins with discontinuities creates Devouard, B., Baronnet, A., Van Tendeloo, G., Amelinckx, S.
localized weak planes within massive serpentinite blocks (1997): First evidence of synthetic polygonal serpentines. Eur. J.
that may preferentially accommodate further brittle defor- Mineral., 9, 539-546.
Dilek, Y., Coulton, A., Hurst, S.D. (1997): Serpentinization and hy-
mation. This example also shows that, under wet and rela-
drothermal veining in peridotites at site 920 in the Mark area. in
tively low temperature conditions, serpentines can accom-
“Proceedings of the Ocean Drilling Program, Scientific Results”,
modate progressive deformation by a dissolution-diffusion- J.A. Karson, M. Cannat, D.J. Miller, D. Elthon, eds., 153, 35-59.
crystallization process. This can also have major implica- Dungan, M.A. (1979): A microprobe study of antigorite and some
tions in other tectonic settings such as the active San Andre- serpentine polymorphs. Can. Mineral., 17, 771-784.
as fault, in which serpentinites are abundant especially in Durney, D.W. & Ramsay, J.G. (1973): Incremental strains measured
seismic creeping segments. by syntectonic crystal growth. in “Gravity and Tectonics”, K.A.
De Jong & R. Scholten, eds., John Wiley, New York, 67-96.
Acknowledgments: We thank C. Nevado, C. Merlet, S. Nit- Ernst, W.G. (1971): Do mineral parageneses reflect unusually high
sche and D. Chaudanson for their technical assistance, S.F. pressure conditions of Franciscan metamorphism? Am. J. Sci.,
Cox and O. Grauby for fruitful discussions, J. Urai and B. 271, 81-108.
Grobéty for their helpful reviews. This work was supported Fisher, D.M. & Brantley, S.L. (1992): Models of quartz overgrowth
by the Observatoire de Grenoble and the IT02 program of and vein formation: deformation and episodic fluid flow in an an-
the INSU-CNRS. cient subduction zone. J. Geophys. Res., 97, 20043-20061.
Grauby, O., Baronnet, A., Devouard, B., Schoumacker, K., Demird-
jian, L. (1998): The chrysotile-polygonal serpentine-lizardite
References suite synthesized from a 3MgO-2SiO2-excess H2O gel. The 7th
International Symposium on Experimental Mineralogy, Petrolo-
Andreani, M. (2003): Les microstructures de déformation des ser- gy, and Geochemistry, Orléans, Abstracts. Terra Nova, supple-
pentines et la partition sismique-asismique: exemple de la Cali- ment 1, 24.
fornie. PhD Thesis, Joseph Fourier University, Grenoble, France. Hilgers, C., Köhn, D., Bons, P.D., Urai, J.L. (2001): Development of
Atkinson, B.K. (1984): Subcritical crack-propagation in geological crystal morphology during unitaxial growth in a progressive wid-
materials. J. Geophys. Research, 89, 4077-4114. ening vein: II. Numerical simulations of the evolution of antitaxi-
Atwater, T. (1989): Plate tectonic history of the northeast Pacific and al fibrous veins. J. Struct. Geol., 23, 873-885.
western North America. in “The Eastern Pacific Ocean and Ha- Means, W.D. & Li, T. (2001): A laboratory simulation of fibrous
waii. Geology of North America”, E.L. Winterer, D.M. Hussong, veins: some first observations. J. Struct. Geol., 23, 857-863.
R.W. Decker, eds., Geological Society of America, Boulder, Col- Mullins, W.W. & Sekerka, R.F. (1963): Morphological stability of a
orado, N, 21-71. particle growing by diffusion or heat flow. J. Appl. Phys., 35,
Bailey, E.H., Irwin, W.P., Jones, D.L. (1964): Franciscan and related 323-329.
rocks and their significance in the geology of western California. Normand, C., Williams-Jones, A.E., Martin, R.F., Vali, H. (2002):
California Division of Mines and Geology Bull., 183, 177. Hydrothermal alteration of olivine in a flow-through autoclave:
Baronnet, A. & Devouard, B. (1996): Topology and crystal growth nucleation and growth of serpentine phases. Am. Mineral., 87,
of natural chrysotile and polygonal serpentine. J. Crystal Growth, 1699-1709.
166, 952-960. O’Hanley, D.S. (1992): Solution to the volume problem in serpenti-
Baronnet, A. & Saül, A. (2003): Interstitial crystal growth from un- nization. Geology, 20, 705-708.
 Crack-seal type serpentine vein 595

Page, B.M. (1972): Oceanic crust and mantle fragment in subduction Twiss, R.J. & Gefell, M.J. (1990): Curved slicken fibers: a new brit-
complex near San Luis Obispo, California. Geol. Soc. Amer. tle shear sense indicator with application to a sheared serpentini-
Bull., 83, 957-972. te. J. Struct. Geol., 12, 471-481.
– (1981): The Southern Coast Range. in “The geotectonic develop- Urai, J.L., Williams, P.F., Van Roermund, H.L.M. (1991): Kinemat-
ment of California”, G.W. Ernst, ed., Prentice-Hall, Englewood ics of crystal growth in syntectonic fibrous veins. J. Struct. Geol.,
Cliffs, New Jersey, 330-417. 13, 823-836.
Page, B.M., Thompson, G.A., Coleman, R.G. (1998): Late Cenozoic Van der Pluijm, B.A. (1984): An unusual “crack-seal” vein geome-
tectonics of the central and southern Coast Ranges of California. try. J. Struct. Geol., 6, 593-597.
Geol. Soc. Amer. Bull., 110, 846-876. Viti, C. & Mellini, M. (1998): Mesh textures and bastites in the Elba
Passchier, C. & Trouw, R. (1996). Microtectonics. Springer, Berlin, retrograde serpentinites. Eur. J. Mineral., 10, 1341-1359.
289 p. Weyl, P.K. (1959): Pressure solution and the force of crystallization:
Ramsay, J.G. (1980): The crack-seal mechanism of rock deforma- a phenomenological theory. J. Geophys. Research, 64, 2001-
tion. Nature, 284, 135-139. 2025.
Renard, F., Andreani, M., Boullier, A.M., Labaume, P. (in press): Wicks, F.G. & Wittaker, E.J.W. (1977): Serpentine texture and ser-
Crack-seal patterns: records of uncorrelated stress release varia- pentinisation. Can. Mineral., 15, 459-488.
tions in crustal rocks. in “Deformation, Mechanisms, Rheology Williams, P.F. & Urai, J.L. (1989): Curved vein fibres: an alternative
and Tectonics: from Minerals to the Lithosphere”, D. Gapais, J.P. explanation. Tectonophysics, 158, 311-333.
Brun, P. Cobbold, eds., Special Publication of the Geological So- Wiltschko, D.V. & Morse, J.W. (2001): Crystallization pressure ver-
ciety of London. sus “crack seal” as the mechanism for banded veins. Geology, 29,
Robert, F. & Boullier, A.M. (1994): Mesothermal gold-quartz veins 79-82.
and earthquakes. in “The mechanical involvement of fluids in Zussman, J., Brindley, G.W., Comer, J.J. (1957): Electron diffraction
faulting”, S.H. Hickman, R.H. Sibson, R.L. Bruhn, eds., U.S. studies of serpentine minerals. Am. Mineral., 42, 666-670.
Geological Survey, Open-file report 94-228, 18-30.
Stamoudi, C. (2002): Processus de serpentinisation des péridotites Received 16 October 2003
de Hess-Deep et de la zone de MARK. Approche chimique et mi- Modified version received 12 February 2004
néralogique. PhD Thesis, Paris VI University, France. Accepted 18 March 2004