cryogenic vitriﬁcation and 3d serial sectioning using high ... · study of the 3d morphology of a...

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


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.



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.




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.




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.




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.




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.




(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.




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.




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




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.




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.




cryogenic vitriﬁcation and 3d serial sectioning using high ... · study of the 3d morphology of a...

Documents