cryogenic vitrification and 3d serial sectioning using high ... · study of the 3d morphology of a...
TRANSCRIPT
Cryogenic vitrification and 3D serial sectioning using highresolution cryo-FIB SEM technology for brine-filled grainboundaries in halite: first results
G. DESBOIS1 , J . L. URAI1, C. BURKHARDT2, M. R. DRURY3, M. HAYLES4 AND B. HUMBEL5
1Structural Geology, Tectonics and Geomechanics, Geological Institute, RWTH Aachen University, Aachen, Germany;2Natural and Medical Sciences Institute (NMI) at the University of Tubingen, Reutlingen, Germany; 3Department of Earth
Sciences, Utrecht University, Utrecht, The Netherlands; 4FEI Company, Nanoport, Eindhoven, The Netherlands; 5Department
of Biology, Utrecht University, Utrecht, The Netherlands
ABSTRACT
The structure of brine films in grain boundaries of halite has been the subject of much controversy over the past
20 years; although a number of innovative methods have been developed to study these structures, much is still
unknown and fundamental information is missing. In this study, we investigated different methods of plunge-
freezing to vitrify the brine fill of grain boundaries for natural salt polycrystal. This was followed by a preliminary
study of the 3D morphology of a vitrified grain boundary in a natural rock salt sample with a focused ion beam
(FIB) excavation system. We have shown that brine-filled grain boundaries in rock salt can be efficiently well fro-
zen when dimensions are less than about 1 mm. Coupled with an ion beam tool, cryo-SEM allows 3D observa-
tion of the well-frozen grain boundaries in large volumes and high resolution. Initial results of brine-filled natural
halite grain boundaries show non-faceted crystal–brine interfaces and unexpectedly low dihedral angles at room
temperature and pressure.
Key words: brine, cryo-SEM, FIB, grain boundary, halite, serial sectioning, vitrification
Received 3 September 2007; accepted 13 December 2007
Corresponding author: Dr Guillaume Desbois, Structural Geology, Tectonics and Geomechanics, Geological Institute,
RWTH Aachen University, Lochnerstrasse 4-20, 52056 Aachen, Germany.
Email: [email protected]. Tel: +49 241 80 95780. Fax: +49 241 80 92358.
Geofluids (2008) 8, 60–72
INTRODUCTION
The transport and mechanical properties of rock salt (poly-
crystalline NaCl) are of importance to the geological evolu-
tion of sedimentary basins, and many applications such as
subsurface storage of oil, gas and radioactive waste. Water
can greatly increase the mobility of grain boundaries in
evaporite minerals so that dynamic re-crystallization and
solution–precipitation processes can become rapid, even at
room temperature (Urai et al. 1986a,b; Drury & Urai
1990; Spiers et al. 1990; Peach et al. 2001; Watanabe &
Peach 2002). This is caused by the presence of brine in
the grain boundaries. The structure of this brine film has
been the subject of much controversy over the past
20 years (Urai et al. 1986a; Den Brok et al. 2002; Urai &
Spiers 2007). Although a number of innovative methods
have been developed to study these structures, such as
infrared spectroscopy (De Meer et al. 2005), interference
microscopy (Lohkamper et al. 2003), deformation experi-
ments in transmitted light (Schenk & Urai 2004), electron
backscattered (BSE) diffraction (Pennock et al. 2006) and
cryogenic scanning electron microscopy (SEM) (Schenk
et al. 2006), much is still unknown and fundamental infor-
mation is missing.
One of the recent developments to study wet grain
boundaries is with high-resolution cryogenic SEM after
stabilization of the water phase in the grain boundaries at
cryo-temperatures. Cryo-SEM is widely used in Life Sci-
ences (Echlin 1978; Sargent1988; Erlandsen et al. 1997;
Fauchadour et al. 1999; Walther & Muller 1999; Frederik
& Sommerdijk 2005; Marko et al. 2007) but applications
in Geosciences are as yet limited to studies of fluid
Geofluids (2008) 8, 60–72 doi: 10.1111/j.1468-8123.2007.00205.x
� 2008 The Authors
Journal compilation � 2008 Blackwell Publishing Ltd
inclusions and wettability (Fassi-Fihri et al. 1992; Mann
et al. 1994; Chenu & Tessier 1995; Robin et al. 1995;
Tritlla & Cardellach 1997; Durand & Rosenberg 1998;
Shepherd et al., 1998; Vizika et al. 1998; Barnes et al.
2003; Negre et al. 2004; Kovalevych et al. 2005, 2006;
Schenk et al. 2006).
One of the critical issues of cryo-SEM (Karlsson 2002)
is to ensure that plunge-freezing vitrifies the pore fluids to
avoid the formation of ice crystals which tend to damage
the microstructure. In geoscience studies to date this issue
has not yet received much attention. If plunge-freezing is
not efficient enough, the brine contained in pores may
crystallize or become segregated in two distinct phases (ice
crystals and hydrohalite; Schenk et al. 2006).
In this study, we investigated different methods of
plunge-freezing (liquid nitrogen, slushy nitrogen and
slushy ethane) to find the best cryo-coolant to well freeze
the brine fill of grain boundaries for polycrystalline natural
salt. This was followed by a preliminary study of the 3D
morphology of a well-frozen grain boundary in a natural
rock salt sample with a focused ion beam (FIB) excavation
system and observation of the frozen fluid at high reso-
lution.
MATERIALS AND METHODS
Samples and experimental procedure
We used three different kinds of samples: (i) 1 mm thick
copper rivets with a drop of saturated brine (called ‘stan-
dards’ in what follows); (ii) thin films of brine sandwiched
between plates of NaCl crystals; and (iii) thin slabs of nat-
ural polycrystalline salt. The synthetic ‘salt sandwiches’
were made from two 0.5 mm thick plates of NaCl single
crystal clamped together with a drop of saturated brine in
between, prepared at least 3 weeks before the cryo-SEM
experiments and stored in a container with saturated
brine. The rivets with brine and the salt sandwiches
were easy to prepare and were used to compare the effi-
ciency of different plunge-freezing procedures for brine
vitrification.
The natural sample is a salt mylonite from the Qum
Kuh salt glacier in Iran (Schleder 2007). The sample of
5 · 5 · 1 mm was gently cut with a 0.3 mm diamond saw
with hexane as lubricant and kept in a container with satu-
rated NaCl solution until the cryo-SEM experiments. For
the three kinds of samples, the basic procedure of prepara-
tion and imaging is the same (Fig. 1): (i) the sample is
plunged in a bath of cryo-coolant with the help of the
clamp for the ‘salt sandwiches’ and with tweezers for the
salt sandwiches and the natural salt polycrystals, and stirred
in the cryo-coolant to improve the cooling rate. (ii)
Within the bath of cryo-coolant, the samples are fixed on
the pre-cooled sample holder and freeze-fractured with a
cold-knife. (iii) The samples are then SEM-imaged at
cryo-temperatures, with ion milling if required. The trans-
fer from the bath of cryo-coolant to the cryo-preparation
chamber and from the cryo-preparation chamber to the
SEM chamber is performed by a cryo-shuttle under
vacuum.
Three kinds of cryo-coolants have been investigated:
liquid nitrogen, slushy nitrogen and slushy ethane. For
nitrogen and ethane, the boiling and melting temperatures
are: TBN2 = )195.8�C, TMN2 = )210�C, and TBE =
)88.6�C and TME = )182.2�C. This provides a range of
cryo-coolants including liquid and slushy phases with
different boiling and melting temperatures, and different
gaps between TM and TB.
For 3D sectioning of the samples, we used an FIB to
cut and remove thin slices in the SEM chamber to image
the fresh surface, in ‘slice and view’ imaging mode. Initially
the top surface of the sample is sputter-coated with Pt but
the freshly cut surfaces are not coated. Before the ‘slice
and view’ imaging, we prepared the area of interest by
milling ‘channels’ around this area to allow removal of the
milling waste (see Fig. 7C).
Analytical techniques and instruments used
The first cryo-SEM experiments were performed at NMI
(Reutlingen, Germany), using a Zeiss cryo-SEM (1540XB
CrossBeam, Zeiss) with a high-resolution GEMINI field
emission column operating at very low acceleration volt-
age with a thermally assisted Schottky field emission gun
that yields a very low energy spread, and a Canion FIB
(FIB) from a gallium metal ion source installed at an
angle of 54� to the electron beam. The instrument also
has an EDX system. Transport and preparation of the
samples are done using a Bal-Tec system which allows
keeping the samples under cryogenic conditions at each
stage of the procedure, in a sample shuttle (VCT 100)
which can be attached via a vacuum gate to the freezing
bath, a cryo-sputter coater (SCD500) and the micro-
scope.
The study of different methods of plunge-freezing and
the ‘slice and view’ imaging of a wet grain boundary were
carried out in the Department of Earth Sciences of
Utrecht University (the Netherlands). The cryo-SEM
available here is a Nova Nanolab 200 Small Dual Beam
system from FEI with FIB and EDX facilities. The specifi-
cations of these features are comparable with the facilities
available in NMI-Reutlingen; but the cryo-transport and
the cryo-preparation of the samples were performed in a
PP2000T cryo station (Quorum Technologies). This sys-
tem provides facilities to import pre-frozen samples in the
cryo-preparation chamber which provides flexibility to use
different plunge-freezing methods and the cryo-prepara-
tion chamber includes a sputter coater facility and a cold
Cryo-FIB-SEM of grain boundary in halite 61
� 2008 The Authors
Journal compilation � 2008 Blackwell Publishing Ltd, Geofluids, 8, 60–72
knife which is useful for the preparation of clean
fractures.
RESULTS
Vitrification of the grain boundary
One of the critical aims of this study was to validate a
method to freeze thin brine films between grains of NaCl.
If plunge-freezing is not efficient enough, the brine con-
tained in grain boundaries is crystallized and segregated
into two distinct phases (ice crystals and hydrohalite;
Schenk et al. 2006). This makes high-resolution study of
grain boundary microstructures difficult.
Standards
The results of these experiments are summarized in Fig. 2.
For each method of plunge-freezing, a photograph at low
magnification shows the fractured surface of the frozen
drop of brine in the rivet, and a high-magnification image
presents the detail of the frozen brine in BSE mode which
enables the identification of the solid phases at constant
cryo-temperature ()150�C).
In liquid nitrogen, the frozen brine contains ice crystals
surrounded by the eutectic phase, which is evidence of
phase segregation during plunge-freezing. In slushy nitro-
gen, the frozen brine exhibits a phase with absence of vis-
ible segregation close to the external rim (40–50 lm) and
a dendritic organized structure in the centre. This struc-
ture is different from the eutectic phase formed in liquid
nitrogen, and is interpreted as incipient phase segregation
during plunge-freezing. In slushy ethane, the BSE image
does not show organized structures through the whole
brine drop. We interpret this as evidence that the
brine filling the rivet was well frozen during the
plunge-freezing. We prefer here to use the term ‘well fro-
zen’ rather than vitrified because the vitrification (sensu
stricto) can only be verified by transmission electron
microscopy (TEM). In summary, the order of the grain
boundary vitrification efficiency for the three investigated
Fig. 1. Schematic drawing of the cryo-sample preparation procedure from the sample mounting stage to SEM imaging.
62 G. DESBOIS et al.
� 2008 The Authors
Journal compilation � 2008 Blackwell Publishing Ltd, Geofluids, 8, 60–72
Fig. 2. SEM photographs (SE and BSE modes) of the standard samples after plunge-freezing either in liquid nitrogen, slushy nitrogen or slushy ethane, show-
ing the different stage of vitrification as a function of the cryo-coolant. The samples are not coated.
Cryo-FIB-SEM of grain boundary in halite 63
� 2008 The Authors
Journal compilation � 2008 Blackwell Publishing Ltd, Geofluids, 8, 60–72
plunge-freezing methods is: liquid nitrogen < slushy nitro-
gen < slushy ethane.
Synthetic salt sandwiches
Based on the results shown above we abandoned the liquid
nitrogen coolant because of its low efficiency, and investi-
gated 10 synthetic ‘salt sandwich’ samples and with two
methods of plunge-freezing (slushy nitrogen and slushy
ethane). For both coolants, the samples show similar mor-
phologies. Figure 3 shows typical results for the two meth-
ods after freeze-fracturing. At larger scale, the frozen brine
film between the crystals indicates a homogeneously solidi-
fied material in both cases. At higher magnification, in BSE
the segregation patterns are absent, indicating fully well-
frozen grain boundaries.
Natural polycrystalline halite
We also performed a test with a natural polycrystalline salt
plunge-frozen in slushy nitrogen and freeze-fractured
(Fig. 4). This sample suggests a well-frozen grain boundary
but some of the images could also be interpreted as show-
ing segregation structures. In BSE there are clear contrasts
Fig. 3. SEM photographs (SE and BSE modes) of the synthetic ‘salt sandwich’ samples after plunge-freezing either in slushy nitrogen or slushy ethane. It is
clear that the brine film between the crystals is well frozen in both coolants. The samples are not coated. The hackles on the sample surface are formed dur-
ing cutting with the cold knife.
64 G. DESBOIS et al.
� 2008 The Authors
Journal compilation � 2008 Blackwell Publishing Ltd, Geofluids, 8, 60–72
between domains in the frozen fluid phase; however this
geometry does not resemble the segregation patterns
observed in the standard experiment (Fig. 2) as formed by
well-defined ice crystals surrounding by the eutectic phase.
Because to get contrast in BSE there must some chemical
or density difference, this geometry could also be due to
two fluids.
Freeze-drying
We have also performed a freeze-drying experiment
(Fig. 5) where a grain boundary initially well frozen in
slushy nitrogen was heated from )150 to )90�C for
60 min. This sample has undergone dramatic changes
showing a frozen brine surface very similar to the one
described by Schenk et al. (2006), with a dense pattern of
holes and 100 nm size cavities. However, we did not find
evidence that these changes caused damage to the sur-
rounding in the form of cracks or fractures.
Application: investigation of wet grain boundaries in
natural rock salt
We present some preliminary results of a study of grain
boundaries from a rock salt mylonite sample from the
Qum Kuh salt glacier (Iran). A detailed description of
these samples in outcrop and thin section, using various
techniques is presented by Schleder (2007).
SEM investigation of dry natural polycrystalline salt
samples
We first studied our samples ‘dry’ in a conventional SEM
to compare observations with the cryo-SEM results. Before
Fig. 4. SEM photographs [(A) SE and (B) BSE modes] of a grain boundary
in a natural polycrystalline salt sample plunge-frozen in slushy nitrogen
(NaCl is the phase on the left, frozen brine on the right). Here, it is unclear
if segregation of phases occurred. The sample is non-coated.
Fig. 5. SEM photographs of the grain boundary of a ‘salt sandwich’ sample
after freeze-drying from )150�C to )90�C for 60 min. Before the freeze-
drying stage, the frozen grain boundary was initially well frozen. The
sample is non-coated. (A) Large view of the freeze-dried grain boundary;
(B) Focus on the NaCl crystal-freeze dried grain boundary interface.
Cryo-FIB-SEM of grain boundary in halite 65
� 2008 The Authors
Journal compilation � 2008 Blackwell Publishing Ltd, Geofluids, 8, 60–72
SEM imaging, the samples were gently broken, dried in air
and coated with Pt sputtering. Observation of grain
boundaries in fractured and dried polycrystalline salt sam-
ples, can provide information on the grain boundary struc-
ture after removal of the fluid and precipitation of an
unknown amount of NaCl (Urai 1985, 1987; Urai et al.
1986a,b; Schenk & Urai 2004).
Figure 6 shows SEM images obtained from the salt gla-
cier samples. Figure 6A,B gives the typical indirect evi-
dence of the presence of fluid in the grain boundaries,
(A) (B)
(C) (D)
(E) (F)
Fig. 6. SEM pictures (SE mode) of grain boundaries in a natural polycrystalline salt sample under dry conditions. (A) Efflorescence at grain boundary; (B) a
grain boundary with a complicated island-channel(?) structure; (C) pores localized along the grain boundary; (D) wavy efflorescence structures along a grain
boundary; (E) triple junction and pores located along the grain boundary; and (F) triple junction showing grain boundary grooving. The sample is Pt-coated.
66 G. DESBOIS et al.
� 2008 The Authors
Journal compilation � 2008 Blackwell Publishing Ltd, Geofluids, 8, 60–72
such as efflorescence structures similar to those reported by
Schenk & Urai (2004), formed by evaporation in combina-
tion with capillary forces and a mobile and connected fluid,
and grain surface which likely formed by solution–crystalli-
zation processes in a brine-filled grain boundary.
Interpreting the efflorescence process along the lines of
argument of Schenk & Urai (2004) allows study of some
aspects of the grain boundaries in detail, and indicates a
rich assembly of grain boundary structures in these sam-
ples: fluid and gas inclusion arrays, grain boundaries con-
taining thin fluid films, and large grain boundary pores
with complex morphologies (Figs 6C–F). However, there
are two major problems with this method: first, at the
nanometre scale it is not possible to meaningfully interpret
the structure of grain boundaries containing thin films or
perhaps island-channel structures, because even a very small
amount of NaCl precipitation will change this structure in
an unpredictable way, and secondly it is very difficult to
know what the other (removed) half of a grain boundary
looks like (cf. Olgaard & FitzGerald, 1993, for a detailed
discussion) and whether it contained a fluid or a gas phase
at this location. Therefore, the possibility to quench the
fluid-filled boundaries and study them in 3D is a major
step forward in this field.
3D investigation of a wet triple junction under cryo-SEM
and FIB milling
In a thin slab of the same sample as studied dry, we inves-
tigated the 3D structure of a halite grain boundary
(Fig. 7) under cryogenic conditions. Our area of interest
was a triple junction surrounded by three grains with dif-
ferent crystallographic orientation in the plane of observa-
tion. The freeze-fracture morphology shows the {0 0 1}
cleavage planes, which allow estimation of the crystallo-
graphic orientation of the surrounding grains (Fig. 7B).
The mis-orientations of all three surrounding grains are
large, and all three boundaries are high-angle grain bound-
aries.
We performed 54 ‘slice and view’ steps of this volume,
with the thickness of each slice around 500 nm. In Fig. 8,
we show a BSE image corresponding to every third slice,
i.e. with a spacing of 1.5 lm. In Fig. 9, the last slice is
shown to give an overview of the morphology of the vol-
ume surrounding the slices. All the cross-sections are non-
coated.
The images (Figs 8 and 9) of the uncoated plunge-fro-
zen grain boundaries show features similar to those
(A)
(B)
(C)
Fig. 7. SEM photographs (BSE mode) of a triple junction located in a natu-
ral polycrystalline salt sample. (A) Overview at larger scale; (B) the overview
at smaller scale reveals the cleavage planes which indicate the different
crystallographic orientation of the three adjacent grains; (C) the dashed line
shows the location of the channels milled to drain the milling waste during
the excavation. The sample is Pt-coated.
Cryo-FIB-SEM of grain boundary in halite 67
� 2008 The Authors
Journal compilation � 2008 Blackwell Publishing Ltd, Geofluids, 8, 60–72
(A) (B) (C)
(D) (E) (F)
(G) (H) (I)
(J) (K) (L)
(M) (N) (O)
(P) (Q) (R)
Fig. 8. SEM images (BSE mode) of progressive slices of brine-filled high-angle grain boundaries in a polycrystalline glacier salt sample. The grain boundary
was well frozen by plunge-freezing in slushy nitrogen. The cross-sections are not coated.
68 G. DESBOIS et al.
� 2008 The Authors
Journal compilation � 2008 Blackwell Publishing Ltd, Geofluids, 8, 60–72
described above: areas which are homogeneous and exhi-
bit no visible segregation with holes in the frozen brine
grain boundaries which appear black on the BSE images,
or become bright because of electron charging. Some of
these structures appear correlated from slice to slice. We
interpret these images to indicate a well-frozen grain
boundary system. The charging effect is probably due to
a very long dwell time of the electron beam during imag-
ing; this effect could be avoided in the future by fine tun-
ing the beam conditions of the ‘slice and view’
procedure.
Figure 8 shows clearly that the brine is not only located
in the triple junction tubes but also in an interconnected
network of fluid films. The thickness of this fluid film
slowly varies along the grain boundary. In Fig. 8A (dashed
arrow), the grain boundary contains a thin fluid film of a
few nanometres which progressively dilates until in Fig. 8r
it has a thickness of a few micrometres. Another interesting
feature (black arrow (Fig. 8K) is the lobate morphology of
the grain surface, whereas on the other side of the grain
boundary the crystal–brine boundary is flat; these lobe
structures were also observed in the sample studied dry.
The third region of interest (black circle Fig. 8R) is a
‘healed’ grain boundary (it may contain a fluid film thinner
than 100 nm, the resolution in these images). The value of
the dihedral angle is around 30�, in contradiction to the
dihedral angle given by Holness & Lewis (1997) for these
conditions (r < 1 MPa; 20�C < T < 40�C). Along this
‘healed boundary’, there are some small cracks which may
have been initiated during plunge-freezing by differences
in thermal expansion between the surrounding crystals, by
freeze-fracturing, or by volume changes during no visible
segregation.
DISCUSSION
The best cryo-coolant for wet salt samples
Numerical simulations of the cooling of our samples indi-
cate that the centre of a salt sandwich reaches the tempera-
ture of slushy ethane in about 0.01 sec, giving a cooling
rate of approximately 20 000 K sec)1. This is faster than
the values quoted by Moor (1971) as required to vitrify
water in biological specimens. Keeping in mind that the
brines in our samples have a high concentration of dis-
solved NaCl and it is not known how this affects vitrifica-
tion efficiency, it seems reasonable to expect that the brine
in our samples is well frozen.
Our experiments performed with the synthetic ‘salt sand-
wiches’ plunge-frozen in slushy nitrogen and slushy ethane
(Fig. 3) show clearly that the brine filling the grain bound-
ary has been well frozen. On the other hand, the experi-
ments performed with the standards give evidence that
crystalline phases segregated during the plunge-freezing in
slushy nitrogen. The experiment performed with a natural
polycrystalline salt (Figs 8 and 9) plunge-frozen in slushy
nitrogen is also well frozen at the grain boundary, but in
another sample there is hint that the brine filling the grain
boundary is not fully well frozen (Fig. 4). This means that
slushy nitrogen may produce variable extents of vitrification
of the brine although the thicknesses of the investigated
samples are the same in each experiment.
These discrepancies can be explained by the proximity of
nitrogen’s boiling and melting temperatures
(TBN2 = )195.8�C and TMN2 = )210�C). Because of this
proximity, it is easy to exceed the boiling point of the
nitrogen just with a slight change of heat energy. There is
only limited control of the amount of heat we import into
the slushy nitrogen when the sample and eventually the
sample holder are plunged in the bath of slushy nitrogen.
If the amount of heat energy is enough to exceed the
nitrogen boiling temperature, the Leidenfrost effect (Fau-
chadour et al. 1999) occurs: the nitrogen boils in the
neighbourhood of the sample during the early phase of the
plunge-freezing. This covers the sample with a fine vapour
layer, which thermally insulates the sample.
The use of slushy nitrogen therefore may produce vari-
able results. Thus, for an optimal control of the freezing of
the brine in the grain boundaries in rock salt (around
1 mm thick and less), it is better to use the slushy ethane
which offers almost the same range of cryo-temperature as
the liquid and slushy nitrogen but for which the Leiden-
frost effect is minimal because the difference of its boiling
and melting temperatures (TBE = )88.6�C and
TME = )182.2�C).
Fig. 9. SEM photograph (BSE mode) of grain boundaries located at a triple
junction in a natural polycrystalline rock salt sample. This image shows the
area surrounding the last photo of the series presented in Fig. 6r. Arrows
and circle show areas of interest: see text for details. The imaged surfaces
are not coated.
Cryo-FIB-SEM of grain boundary in halite 69
� 2008 The Authors
Journal compilation � 2008 Blackwell Publishing Ltd, Geofluids, 8, 60–72
Freeze-drying, cooling and phase segregation
Schenk et al. (2006) showed the potential of the use of
cryo-SEM for the study of wet halite salt samples but did
not systematically study the structure of the frozen brine in
the grain boundaries. They interpreted the structures on
the freeze-fractured brine surfaces as ‘segregation pattern’
in reference to the segregation of phases (hydroha-
lite + ice) in brine which occurs during a freezing (Roed-
der 1984; Bodnar 1993). We obtained very similar
structures in initially well-frozen brine surfaces after a stage
of freeze-drying.
These results are consistent, considering that Schenk
et al. (2006) used a different procedure of sample prepara-
tion after the plunge-freezing. The samples can be some-
times covered by a condensate deposit from water vapour
in the cryo-coolant bath. Schenk et al. (2006) have system-
atically freeze-dried their samples in order to remove the
condensate. We interpret the surface pattern observed by
Schenk et al. (2006) in frozen grain boundaries to have
formed after the segregation of phases, not during too
slow plunge-freezing but after freeze-drying of an initially
well-frozen grain boundary to clean the sample surface.
Based on our observations so far, damage to the samples,
during plunge-freezing, freeze-fracturing, or during segre-
gation are minor, although more studies are required to
fully understand these effects.
Interpretation of grain boundary microstructures in wet
polycrystalline salt samples
Investigations of dried natural polycrystalline salt samples
provide indirect observations of fluid morphology in grain
boundaries pores (Urai 1985, 1987). Using the cryo-SEM
approach, it is possible to study fluid geometry in situ,
avoiding artefacts caused by efflorescence and drainage of
the fluid. Sectioning the sample with an ion beam after
plunge-freezing has been shown to provide superior images
at high resolution, in addition to a 3D view.
In 3D, large parts of the grain boundary contain a fluid
film of variable thickness, with structures ranging from no
visible fluid film at the resolution of the ‘slice and view’
technique to a fluid-filled film several microns thick. This
fluid distribution is comparable with some of the structures
reported by Schenk & Urai (2005) in wet, artificial fine-
grained polycrystalline compacted at high total stress (but
presumably at low effective pressure), where significant
open porosity can exist (cf. Schoenherr et al. 2007; Schle-
der et al. in press). This type of fluid distribution is
expected to be much less common in domal salt samples,
although very little data are available at present.
In the sample studied in Figs 8 and 9 the halite surfaces
in contact with the brine are clearly curved, and do not
contain crystallographic facets which might be expected for
halite which has been in contact with a saturated brine for
a long time. There are several possible explanations for
this. First, the facets may have been ‘rounded’ during the
(metadynamic) recrystallization during natural deformation
in the salt glacier or after sampling and storage of the sam-
ple. Secondly, the sampled part of the salt glacier may have
been dissolving very slowly at the time of sampling, after
rainfall some time ago (Talbot & Rogers 1980). It is unli-
kely that these structures are due to damage to the sample
during storage after cutting or the cryo-SEM procedure,
because they are also observed in the samples which were
fractured and dried (Fig. 6F). Similar lobate structures on
grain surfaces have been described in naturally deformed
carnallite (Urai 1987) where they were interpreted to be
growth structures associated with diffusive mass transfer
processes, and in experiments, where the ‘grain boundary
groove’ structures observed by Den Brok et al. (2002)
were related to stress in the samples. The cusps along the
halite–brine interface could also be related to the intersec-
tion of subgrain boundaries with the film fluid.
Following Holness & Lewis (1997)and Schoenherr et al.
(2007), the dihedral angle h is the main controlling
parameter for the distribution of grain boundary fluid in
equilibrium. In the case of h >60� the grain boundaries are
fluid-free and consist of solid-solid contacts, whereas for h<60� the fluids form an interconnected network of grain
boundary triple junction tubes. In Figs 8 and 9 and in the
region surrounded by the black circle, the dihedral angle
(h) is around 30� which is much lower than the values pro-
posed by Holness & Lewis (1997) for the P–T conditions
of our samples. In addition, it is not clear if a true
triple point exists in this location, of whether the fluid
progressively thins into a nanometre-scale film, perhaps
also influenced by the anisotropy of surface energy of halite
(Moment & Gordon 1964; Jessell et al. 2007).
Implications for further studies
The main results of this paper are the validation of the tech-
nique of freezing with absence of visible segregation and
ion-beam cutting of brine-filled grain boundaries in natural
polycrystalline salt samples with a diameter of 1 mm, and
some interesting observations on the 3D structure of the
wet grain boundaries. The technique is promising for a
wide range of studies of single- or multiphase grain bound-
aries which are above their melting point in situ, such as
low-permeability reservoir rocks and claystones in which
the details of the morphology of the pore space phase dis-
tribution are poorly known (Hildenbrand et al. 2006).
In the case of rock salt, water-assisted grain boundary
processes have a major effect on the mechanical and trans-
port processes under a wide range of conditions (Urai &
Spiers 2007), and this study provides the basis for a wide
range of systematic studies of the morphology of grain
70 G. DESBOIS et al.
� 2008 The Authors
Journal compilation � 2008 Blackwell Publishing Ltd, Geofluids, 8, 60–72
boundary fluids. First, there are a wide range of natural
rock salt tectonites with correspondingly diverse grain
boundary structure, and secondly, it should be possible to
quench miniature deformation cells containing grain
boundary structures ‘under stress’, and use the same cryo-
SEM technique to study grain boundaries undergoing
active deformation. This could give an important contribu-
tion to the study of the structure of dynamically stable
brine films in grain boundaries in halite (Urai et al.
1986a). Moreover, because most cryo-SEM machines are
combined with a micro-chemical analysis tools (EDX), the
identification of second-phase impurities, and the measure-
ment of impurity concentrations in the fluid-filled grain
boundaries are possible.
CONCLUSIONS
Samples of rock salt containing brine-filled grain bound-
aries can be efficiently well frozen when dimensions are less
than about 1 mm. Coupled with an ion beam tool dedi-
cated for material excavation, cryo-SEM allows 3D ‘slice
and view’ observation of the vitrified grain boundaries in
large volumes and high resolution. Initial results of brine-
filled natural halite grain boundaries show a number of
unexpected structures such as non-faceted crystal–brine
interfaces and low dihedral angles at room temperature
and pressure, calling for further study.
ACKNOWLEDGEMENTS
We thank Zsolt Schleder, Johannes Schoenherr, Oliver
Schenk and Joyce Schmatz for help with obtaining and
preparing the natural samples, and Chris Spiers for discus-
sions on the nature of fluid films in halite. We are grateful
to Sebastian Schadler, Chris Schneijdenberg and Matthijs
de Winter and Georg Koschek for their technical support
with the cryo-SEM machines.
This project was funded by the Deutsche Forchungs-
gemeinschaft (UR 64 ⁄ 9-1). We thank also the Netherlands
Organisation for Scientific Research (175.010.2005.007)
with the support of Utrecht University and FEI Company
(Eindhoven) which funded the cryo-SEM machine located
at Utrecht University.
REFERENCES
Barnes PRF, Wolff EW, Mallard DC, Mader HM (2003) SEMstudies of the morphology and chemistry of polar ice. MicroscopyResearch and Technique, 62, 62–9.
Bodnar RJ (1993) Revised equation and table for determining the
freezing point depression of H2O-NaCl solutions. Geochimicaand Cosmochimica Acta, 57, 683–4.
Chenu C, Tessier D (1995) Low temperature scanning electron
microscopy of clay and organic constituents and their relevance
to soil microstructures. Scanning Microscopy, 9, 989–1010.
De Meer S, Spiers CJ, Nakashima S (2005) Structure and diffusiveproperties of fluid-filled grain boundaries: an in-situ study using
infrared (micro) spectroscopy. Earth and Planetary Science Let-ters, 232, 403–14.
Den Brok SWJ, Morel J, Zahid M (2002) In situ experimentalstudy of roughness development at a stressed solid ⁄ fluid inter-
face. In Geological Society Special Publications 200 (eds de Meer
S, Drury MR, de Bresser JHP, Pennock GM), pp. 73–83. Geo-logical Society, London.
Drury MR, Urai JL (1990) Deformation-related recrystallization
processes. Tectonophysics, 172, 235–53.
Durand C, Rosenberg E (1998) Fluid distribution in kaolinite- orillite-bearing cores: cryo-SEM observations versus bulk measure-
ments. Journal of Petroleum Science and Engineering, 19, 65–
72.
Echlin P (1978) Low temperature scanning electron microscopy: areview. Journal of Microscopy, 112 ⁄ 1, 47–61.
Erlandsen SL, Bittermann A, Krumweide M, Nelson RD, Chen Y,
White JG (1997) High-resolution backscatter electron detectionof cell surface molecules on human platelets using the double-
layer coating method and cryo field emission scanning electron
microscopy. Scanning, 19, 356–60.
Fassi-Fihri O, Robin M, Rosenberg E (1992) Wettability ofreservoir rock at the pore scale: contribution of cryo-scanning
electron microscopy. Revue de l’institut francais du petrole., 47,
685–701.
Fauchadour D, Pouget T, Lechaire J-P, Rouleau L, Normand L(1999) Evaluation of cryotechniques for TEM observation of
sols – application to boehmite sols used in catalysts forming.
Oils & Gas Science and Technology – Rev. IFP., 54, 513–24.Frederik PM, Sommerdijk NAJM (2005) Spatial and temporal res-
olution in cryo-electron microscopy – a scope for nano-chemis-
try. Current Opinion in Colloid & Interface Science, 10, 245–9.
Hildenbrand A, Kroos B, Urai JL (2006) Relationship BetweenPore Structure and Fluid Transport in Argillaceous Rocks. In:
IUTAM Symposium on Physicochemical and ElectromechanicalInteractions in Porous Media. Solid Mechanics and its Applica-tions (eds Gladwell GML, Huyghe JM, Raats PAC, Cowin SC)Volume 125, pp. 231–7. Springer, the Netherlands.
Holness MB, Lewis S (1997) The structure of the halite-brine
interface inferred fron pressure and temperature variations of
equilibrium dihedral angles in the halite-H2O-CO2 system.Geochimica and Cosmochimica Acta, 61, 795–804.
Jessell MW, Urai JL, Schenk O (2007) Microdynamic simulation of
grain growth in halite. In: Microdynamic Simulation with ELLE(eds Jessell MW, Bons PDD, Koehn D), pp. 209–14. Springer-
Verlag, Berlin.
Karlsson JO (2002) Cryopreservation: freezing and vitrification.
Science, 296, 655–6.Kovalevych VM, Carmona V, Pueyo JJ, Peryt TM (2005) Ultram-
icrochemical analyses (UMCA) and cryogenic scanning electron
microscopy (cryo-SEM-EDS) of brines in halite-hosted fluid
inclusions: a comparative study of analytical data. GeochemistryInternational, 43, 268–76.
Kovalevych VM, Marshall T, Peryt TM, Petrychenko OY, Zhukova
SA (2006) Chemical composition in Neoproterozoic: Resultsof fluid inclusion study of halite from salt range (Pakistan)
and Amadeus basin (Australia). Precambrian Research, 144, 39–
51.
Lohkamper THK, Jordan G, Costamagna R, Stockhert B, SchmahlW (2003) Phase shift interference microscope study of dissolu-
tion-precipitation processes of nonhydrostatically stressed halite
crystals in solution. Contributions to Mineralogy and Petrology,146, 263–74.
Cryo-FIB-SEM of grain boundary in halite 71
� 2008 The Authors
Journal compilation � 2008 Blackwell Publishing Ltd, Geofluids, 8, 60–72
Mann U, Neisel JD, Burchard WG, Heinen V, Welte DH (1994)Fluid-rock interfaces as revealed by cryo-scanning electron
microscopy. First Break, 12, 131–36.
Marko M, Hsieh C, Schalek R, Frank J, Mannella C (2007)
Focused-ion-thinning of frozen-hydrated biological specimensfor cryo-electron microscopy. Nature Methods, 4, 215–21.
Moment RL, Gordon RB (1964) Energy of grain boundaries in
halite. Journal of American Ceramic Society, 47, 570–3.Moor H (1971) Recent progress in freeze-etching technique. Philo-
sophical Transactions of the Royal Society London, 261, 121–31.
Negre M, Leone P, Trichet J, Defarge C, Boero V, Gennari M
(2004) Characterization of model soil colloids by cryo-scanningelectron microscopy. Geoderma, 121, 1–16.
Olgaard DL, FitzGerald JD (1993) Evolution of pore microstruc-
tures during healing of grain boundaries in synthetic calcite
rocks. Contributions to Mineralogy and Petrology, 115, 138–54.Peach C, Spiers CJ, Trimby PW (2001) Effect of confining pres-
sure on dilatation, recrystallization, and flow of rock salt at
150�C. Journal of Geophysical Research, 106, 13315–28.Pennock GM, Drury MR, Spiers CJ (2006) Grain boundary popu-
lations in wet and dry NaCl. Materials Science and Technology,22, 1307.
Robin M, Rosenberg E, Fassi-Fihri O (1995) Wettability studiesat the pore level: a new approach by use of Cryo-SEM. SPEFormation Evaluation, 10, 11–9.
Roedder E (1984) The fluids in salt. American Mineralogist, 69,
413–39.Sargent JA (1988) Low temperature scanning electron microscopy:
advantages and applications. Scanning Microscopy, 2, 835–49.
Schenk O, Urai JL (2004) Microstructural evolution and grainboundary structure during static recrystallization in synthetic
polycrystals of sodium chloride containing saturated brine. Con-tributions to Mineralogy and Petrology, 146, 671–82.
Schenk O, Urai JL (2005) The migration of fluid-filled grainboundaries in recrystallizing synthetic bischofite – first results of
in-situ HPHT deformation experiments in transmitted light.
Journal of Metamorphic Geology, 23, 695–709.
Schenk O, Urai JL, Piazolo S (2006) Structure of grain boundariesin wet, synthetic polycrystalline, statically recrystallizing halite –
evidence from cryo-SEM observations. Geofluids, 6, 93–104.
Schleder Z (2007) Deformation mechanisms of naturally deformedrocksalt. PhD Thesis at Geology and Endogene Dynamik,RWTH, Aachen University, Germany.
Schleder Z, Urai JL, Nollet S, Hilgers C (in press) Deformation and
recrystallization mechanisms of rocksalt as evidenced by micro-structures – a case study on Zechstein (Z1) rocksalt from Neuhof
salt mine (Germany). International Journal of Earth Sciences.Shepherd TJ, Ayora C, Cendon DI, Chenery SR, Moissette A
(1998) Quantitative solute analysis of single fluid inclusions in
halite by LA-ICP-MS and cryo-SEM-EDS. Complementary mi-crobeam techniques. European Journal of Mineralogy, 10,
1097–108.
Schoenherr J, Urai JL, Kukla PA, Littke R, Schleder Z, Larroque
J-M, Newall MJ, Al-Abry N, Al-Siyabi HA, Rawahi Z (2007)Limits to the sealing capacity of rock salt: a case study of the
Infra-Cambrian Ara Salt from the South Oman Salt Basin.
AAPG Bulletin, 91, 1–17.Spiers CJ, Schutjens PMTM, Brzesowky RH, Peach CJ, Liezen-
berg JL, Zwart HJ (1990) Experimental determination of con-
stitutive parameters governing creep of rocksalt by pressure
solution. In: Deformation Mechanisms, Rheology and Tectonics(eds Knipe RJ, Rutter EH), Geological Society of London SpecialPublication 54, 215–27.
Talbot CJ, Rogers EA (1980) Seasonal movements in a salt glacier
in Iran. Science, 208, 395–7.Tritlla J, Cardellach E (1997) Fluid inclusions in pre-ore minerals
from the carbonate-hosted mercury deposits in the Espadan
Ranges (eastern spain). Chemical Geology, 137, 91–106.Urai JL (1985) Water-enhanced dynamic recrystallization and
solution transfer in experimentally deformed carnallite. Tectono-physics, 120, 285–317.
Urai JL (1987) Development of microstructure during deforma-tion of carnallite and bischofite in transmitted light. Tectonophys-ics, 135, 251–63.
Urai JL, Spiers CJ (2007) The effect of grain boundary water
on deformation mechanisms and rheology of rocksalt duringlong-term deformation. In: The Mechanical Behaviour of Salt– Understanding of THMC Processes in Salt (eds Wallner M,
Lux KH, Minkley W, Hardy HR Jr), pp 149–58. Taylor &Francis Group, London.
Urai JL, Spiers CJ, Zwart HJ, Lister GS (1986a) Dynamic recrys-
tallisation of minerals. In: Mineral and Rock Deformation; Labo-ratory Studies; The Paterson Volume. AGU GeophysicalMonograph 36 (eds Hobbs BE, Heard HC), pp. 161–99. Ameri-
can Geophysical Union, Washington, DC.
Urai JL, Spiers CJ, Zwart HJ, Lister GS (1986b) Weakening
of rocksalt by water during long term creep. Nature, 324, 554–7.
Vizika O, Rosenberg E., Kalaydjian F (1998) Study of wettability
and spreading impact in three-phase gas injection by cryo-scan-
ning electron microscopy. Journal of Petroleum Science andEngineering, 20, 189–202.
Walther P, Muller M (1999) Biological ultrastructure as revealed
by high resolution cryo-SEM of block faces after cryo-section-ing. Journal of Microscopy, 19, 279–87.
Watanabe T, Peach CJ (2002) Electrical impedance measurement
of plastically deforming halite rocks at 125�C and 50 MPa. Jour-nal of Geophysical Research, 107, ECV 2-1–2-12.
72 G. DESBOIS et al.
� 2008 The Authors
Journal compilation � 2008 Blackwell Publishing Ltd, Geofluids, 8, 60–72