Microstructural evolution of deformation–modified primary halite from the Middle Triassic Röt Formation at Hengelo, The Netherlands

Zsolt Schléder, János L. Urai
Geologie-Endogene Dynamik, RWTH Aachen, Lochnerstrasse 4-20, 52056 Aachen, Germany

Abstract
The microstructure of halite from the subhorizontal, bedded Main Röt Evaporite Member at Hengelo, The Netherlands (AKZO well 382, depth interval of 420-460 m), was studied by transmitted and reflected light microscopy of gamma-irradiation decorated samples. Primary microstructures compare favourably with those found in recent ephemeral salt pans. Large, blocky, fluid-inclusion-poor halite grains and elongated chevrons are interpreted to have formed in the saline lake stage, while void-filling clear halite is interpreted to have formed during the desiccation stage of the salt pan. In addition, in all layers the grains are rich in deformation-related substructures such as slip bands and subgrains indicating strains of a few percent. The study of gamma-irradiation decorated thin sections shows that the main recrystallization mechanism is grain boundary migration. Grain boundary migration removes primary fluid inclusions and produces clear, strain-free new grains. Differential stresses as determined by subgrain size piezometry were 0.45 – 0.97 MPa. The deformation of the salt layers is probably related to Cretaceous inversion in the area.

Keywords: Hengelo (The Netherlands) – Triassic bedded rock salt – Synsedimentary structures – Deformation microstructures – Recrystallization



1. Introduction
The microstructure of natural rock salt is a product of a complex series of depositional, diagenetic and deformation processes. Depending on the relative importance of these, a wide range of microstructures can be formed, starting from the typical primary structures in young, shallow halite to the strongly deformed and completely recrystallized tectonites in salt diapirs. Although both synsedimentary and deformation-related microstructural processes are documented (e.g. Shearman, 1970; Lowenstein and Hardie, 1985; Urai et al., 1987; Casas and Lowenstein, 1989), details are not well understood, making the interpretation of halite microstructures difficult.

One commonly observed microstructure consists of composite halite grains locally rich in fluid inclusions, but also containing irregular patches of clear halite free of fluid inclusions. Not uncommonly, the transition from fluid-inclusion-rich material to clear halite is defined by a sharp curved surface within single halite grains. One possible explanation for this structure is a syndepositional solution-reprecipitation process in ephemeral or shallow brine system (shallow enough to become undersaturated with respect to halite when diluted by floodwater). In an ephemeral salt pan environment, a flooding stage (when floodwater arrives and partly dissolves the existing halite crust) is followed by an evaporation stage, in which the irregularly dissolved halite framework is covered by a new layer of fluid-inclusion-rich grains. Finally, when the saline lake shrinks and dries out, clear halite slowly crystallizes in the dissolution voids from the residual groundwater brine (Lowenstein and Hardie, 1985).
After burial, if the halite is deformed and recrystallized, a similar patchy core-and-mantle structure can form by grain boundary migration, where the migrating grain boundaries erase primary fluid inclusion bands. Thus the presence of patches of clear halite in fluid inclusion-rich grains can also be explained by a deformation-induced recrystallization process (Urai et al., 1987).

This problem, i.e. distinguishing between primary and secondary features in halite, has been recognized by several authors (Wardlaw and Schwerdtner, 1966; Hardie et al., 1983; Roedder, 1984) but these authors did not provide criteria for distinguishing between microstructures resulting from deformation and those resulting from recrystallization.
The aim of this paper is to present a detailed microstructural analysis of bedded salt from a core taken near Hengelo, The Netherlands (Fig. 1), in an attempt to separate primary and deformation-related microstructures, and to quantify the paleostress in the salt sequence using subgrain size piezometry.


Fig. 1. Map of study area showing the Gronau fault zone (with Late Cretaceous movements) and approximate location of the normal fault in the study area (after Geluk and Duin, 1997; NITG, 1998; Doornenbal et al., 2002). The stippled polygon indicates the Twenthe-Rijn concession area.



2    Lithostratigraphy of the Röt strata and location of the study material
Samples described in this paper are from the Röt Formation in the Twenthe-Rijn concession area, where more than 400 wells have been drilled for solution mining since 1919. The Röt Formation (Early Anisian) comprises the strata between the Solling Formation and the Muschelkalk Formation (Geluk and Röhling, 1997) (Figs. 2 and 3). The Röt Formation is divided into two members. At the base lies the Main Röt Evaporite Member, which consists of four salt layers denoted by the Akzo Nobel company as salt A to D from bottom to top (Harsveldt, 1980; van Lange, 1994; Kovalevych et al., 2002). Salt layers A and C are the thickest (25 to 30 m), while layers B and D are only a few meters thick. The D salt layer is laterally discontinuous, developed as lenses in former topographic depressions. The salt layers are separated by 1 to 2 m thick shaly anhydritic mudstones and dolomitic claystones (Harsveldt, 1980; van Lange, 1994) (Fig. 3). Lateral variations in thickness of the Röt salt are interpreted as syndepositional, reflecting the relief of the underlying formation, and not due to salt tectonics (Harsveldt, 1980; van Lange, 1994). The depositional environment of the Main Röt Evaporite Member is interpreted as an enclosed sea with alternating periods of non-clastic (pure evaporite) deposition and increased sediment influx (RGD, 1993; van Lange, 1994; NITG, 1998). Above the Main Röt Evaporite Member lies the Upper Röt Claystone Member, which is made up mainly of silty claystone with gypsum and anhydrite nodules (Fig. 3). The thickness of the Upper Röt Claystone Member is up to 200 m (van Lange, 1994; NITG, 1998).


Fig. 2. a) A large-scale profile through the Twenthe-Rijn concession area. Note that displacement along faults can be traced from the Carboniferous basement (Limburg Group) up to the base of Tertiary. The approximate location of profile “b” is indicated by the rectangle. b) Profile through the Twenthe-Rijn concession area with the location of AKZO well 382 indicated. The profile is based on maps of Geluk and Duin (1997) and Doornenbal et al. (2002). For the position of the profiles see Fig. 1, and note that the positions are not identical.

The NW-SE trending Gronau fault zone, situated approximately 15 km NE of the study area, is the main structural element in the region (NITG, 1998) (Figs. 1 and 2). Movement along the fault zone occurred since the Carboniferous and this zone has been reactivated in tectonic phases of Austurian and Saalian, Late Permian, Early-, and Late Kimmerian, Sub-Hercynian and Savian age (c.f. NITG, 1998 p. 112). At the local scale, within the Twenthe-Rijn concession area, a few NW-SE running faults with a maximum displacement of 50-100 m were reported (Harsveldt, 1980) (Fig. 2). Displacement along those faults can be traced up to the base of Tertiary, although their structure within the salt layers is not clear (Geluk and Duin, 1997; Doornenbal et al., 2002). The presence of these faults is probably related to Cretaceous inversion in the area (de Jager, 2003). The Tertiary is characterized by tectonic inactivity (de Jager, 2003).

In this paper we describe samples from AKZO well 382 in the Twenthe-Rijn concession area (Fig. 3). Samples were taken from the “A” salt layer (457.5 m), from the “C” salt layer (443.5 m and 430.9 m) and “D” salt layer (423.3 m). The diameter of the core was 10 cm; the length was between 10 and 15 cm. All photographs and illustrations presented in the paper are oriented with the top of the section towards the top of the page.


Fig. 3. Triassic lithostratigraphy of AKZO well 382 (after van Lange, 1994; Geluk and Duin, 1997). Arrows indicate the studied intervals.



3. Sample preparation and methods of study
We cut 2 x 5 x 8 cm slabs from the core, parallel to the core axis, using a diamond saw with a small amount of water (this prevents the development of microcracks in the halite, without causing significant dissolution artefacts perhaps due to the Joffé-effect, Joffé, 1928). Microstructures were decorated by gamma-irradiation in the Research Reactor of Forschungszentrum Jülich, using a technique similar to that described by Urai et al., 1985. Two sets of irradiations were carried out. One irradiation was done at a temperature of 35 ºC with a dose rate between 1 kGy/h and 3 kGy/h to a total dose of about 1.5 MGy. The other set was done at a temperature of 100 ºC with a dose rate between 4 kGy/h and 6 kGy/h to a total dose of about 4 MGy. Depending on the temperature of the gamma-irradiation the slabs became brown (35 ºC) or blue coloured (100 ºC). The colour intensity seen in the halite samples is heterogeneous, reflecting the heterogeneous distribution of solid solution impurities and other crystal defects in the halite grains (Przibram, 1954; van Opbroek and den Hartog, 1985; Urai et al., 1985; Garcia Celma and Donker, 1996). After irradiation the slabs were polished dry on grinding paper, etched with pure water for 2 seconds and quickly dried with a tissue. This etching technique removes scratches and provides a micro-relief on the surface of the slabs improving the stability of mounting. The slabs were mounted on glass plates at room temperature using epoxy (Körapox 439), and were cut into thick sections of 4 mm using the wet cutting technique. The sections were then ground down to a thickness of 1 mm with grinding paper (dry) and finally etched using the method described by Urai et al., 1987. The thin sections were studied with reflected and transmitted, plane polarized light microscopy.



4. Petrography of the halite samples
Two types of halite occur in the samples: a milky, fluid-inclusion-rich halite that comprises up to 40 volume percent of the material, and a clear, fluid-inclusion-poor halite. The two types commonly occur together, with most halite grains having one or more milky cores surrounded by clear halite. Some of the grains have no apparent milky core, and consist entirely of clear halite. The milky core invariably has a banded structure: fluid-inclusion-rich bands alternate with fluid-inclusion-poor bands (Fig. 4). The fluid inclusion bands are about 200 to 400 µm thick and contain cubic (negative crystal form) fluid inclusions. The alternating fluid inclusion bands define chevrons, cubes and hoppers.


Fig. 4. Photomicrograph of an unirradiated sample (salt layer D, 423.3 m) showing primary fluid-inclusion-rich bands (c) and zones of clear halite (d). Clear halite commonly truncates fluid-inclusion-rich domains across sharp, curved, intracrystalline boundaries. Grain boundaries show up as dark curves in thin section. Secondary phases occurring at grain boundaries are anhydrite and polyhalite. Transmitted light image, image width is 2 cm.

Chevrons were found in the A, C and D layers (457.5 m, 443.5 m and 423.3 m). The chevron halite grains are 0.5 to 2 cm long, vertically elongated grains, which are very often truncated or rimmed by clear halite (Fig. 5). Commonly, the clear rim comprises 50-70 volume percent of a chevron grain, rarely up to 90 percent. The transition from the milky, fluid-inclusion-rich core to the clear halite is defined by a sharp, curved surface (Fig. 5). In gamma-irradiated samples usually there is no difference in irradiation colour between the milky and clear halite within one grain (Fig. 6). In some cases the top of the chevrons is truncated by layers of anhydrite and polyhalite partings (Fig. 7).



Fig. 5. Photomicrographs of halite from salt layer C (443.5 m) show sharp, commonly curved boundaries between milky, fluid-inclusion-rich and clear halite regions within a single grain. The black straight lines seen in the right hand side image are cleavage planes, introduced during sample preparation. Plane polarized, transmitted light images of unirradiated samples. Image width is 2.7 mm.



Fig. 6. Microstructure in halite from salt layer D (423.3 m). Fluid inclusion bands (c), truncated by clear halite (d). Note that individual fluid inclusion bands can be traced over the clear halite regions. This microstructure compares well with those produced by dissolution-reprecipitation processes in present-day salt pan halites (c.f. Shearman, 1970; Lowenstein and Hardie, 1985). Plane polarized, transmitted light images of samples irradiated at 100 ºC. Image width is 7 mm.



Fig. 7. Overview image of a thin section of gamma-irradiation decorated sample (irradiated at 100 ºC) from salt layer C (443.5 m) photographed in transmitted light. White patches are chevron halite grains. Grain boundaries occur as dark, nearly black lines, the white polygons within grains are subgrain boundaries. Three different layers can be distinguished. Layer 1 (at the bottom) consists of chevron halite grains and is extensively truncated by dissolution pipes. The boundary between layers 1 and 2 is marked by anhydrite and polyhalite partings. The layer 2 is much alike as layer 1, and also characterized by the presence of chevrons and clear grains. The boundary between the layers 2 and 3 also characterized by the presence of anhydrite and polyhalite partings. At the top (layer 3) large, blocky halite grains occur. New, strain-free grains, or migrated grain boundaries were rarely observed, suggesting that deformation-induced recrystallization did not alter the primary structures. Image width is 4.5 cm.

Cubes and hoppers outlined by fluid inclusion bands were found in the A, C and D layers (457.5 m, 443.5 m and 423.3 m), although these occur less frequently than chevrons. The size of cube and hopper grains varies between 0.5 and 3 cm, however the larger grains (1.5 to 3 cm) very rarely contain well-developed fluid inclusion bands (Fig. 8).
Such large, blocky halite grains were found in the C layer (443.5 m and 430.9 m). In some cases, small (0.5 to 1.5 cm) grains are intercalated in these large, blocky grains. These hopper-shaped small grains are visible only by the different irradiation colouring (less intensively coloured, Fig. 8b).

Clear halite grains, rarely observed, occur as 0.3 to 1 cm crystals, locally as elongated grains and usually among chevrons, or rarely as smaller (0.1 to 0.6 cm) equiaxed grains (Fig. 9).
Anhydrite and polyhalite occur at grain boundaries of the halite grains, and commonly within grains either as inclusions or as small grains arranged into thin, curved bands. In many cases anhydrite and polyhalite have a spherulitic appearance with the laths arranged radially around a central core (Fig. 8a).



Fig. 8. Photomicrographs of large, blocky halite grains. Transmitted light images.
a) Large, blocky, clear halite grains with few fluid inclusion bands (see arrows). Dark lines are grain boundaries, the grey and red patches are anhydrite and polyhalite. Unirradiated sample from salt layer C (443.5 m). Image width is 3 cm.
b) Hopper crystals intercalated with the large, blocky grains (see arrows) as revealed by 35 ºC gamma-irradiation. Note that the subgrain boundaries were not decorated by the 35 ºC irradiation. Sample from salt layer C (430.9 m). Image width is 4.5 cm.

The majority of both clear and milky grains contains subgrains, as shown by etching and gamma-irradiation (decorated subgrain boundaries were observed only in slabs irradiated at 100 ºC and high total dose). In most grains, subgrains occur as equiaxed polygons with triple junctions of about 120° and with an average size between 250 and 400 µm (Fig. 9). In some of the large grains (e.g. C sample, 430.9 m) the subgrains are arranged into wavy, crystallographically controlled bands (Fig. 10). Much less frequently, elongated subgrains are found in grain boundary regions. Subgrain-free grains and grain boundary regions were observed in all samples, comprising up to 5 to 10 % of halite volume. Subgrain-free grains are relatively small (usually <0.5 cm) and are equiaxed and fluid-inclusion-free (Figs. 9 and 11). Subgrain-free grain boundary regions are always narrow (<0.3 cm) and less intensively coloured by 35 ºC irradiation (Figs. 12 and 13). Gamma-irradiation decorated bands, believed to be slip lamellae, were observed in a number of cases in both subgrain-rich and subgrain-free grains (Fig. 14). The direction of the decorated glide planes slightly changes direction at some subgrain boundaries (Fig. 14).
Grain boundaries are usually serrated at the contact of subgrain-rich grains, and commonly smooth at the contact of subgrain-rich and subgrain-free grains (Fig. 15). Very rarely bulged grain boundaries between subgrain-rich and subgrain-free grains were observed (Fig. 12). Arrays of fluid inclusions are common on grain boundaries (Fig. 16).



Fig. 9. Overview image of a thin section of sample gamma-irradiation decorated at 100 ºC from salt layer A (457.5 m), photographed in transmitted light. Grain boundaries occur as dark curves; white patches are fluid-inclusion-outlined chevrons and hoppers; and white polygons are subgrains. A few new, recrystallized, strain-free grains or grain regions are visible (see arrows). Image width is 4 cm.



Fig. 10. Microstructure of gamma-irradiated (irradiated at 100 ºC) halite sample from salt layer C (430.9 m), photographed in plane polarized transmitted light. Subgrains are arranged into NE-SW trending, wavy, crystallographically controlled bands, perhaps suggesting cross-slip as one of deformation mechanisms. Image width is 17 mm.



Fig. 11. Photomicrographs show the microstructures of sample from salt layer A (457.5 m). Plane polarized transmitted light images of sample gamma-irradiated at 100 ºC.
a) Migration of high-angle grain boundaries as recorded by elongated subgrains, and strain-free new material, which grows at the expense of old, heavily substructured grains. Image width is 11 mm.
b) Strain-free grains grew at the expense of deformed ones. The size of some of the new grains is comparable to that of subgrains. Image width is 7 mm.



Fig. 12. Microstructures of gamma-irradiated sample from salt layer A (457.5 m). Image width is 2.7 mm.
a) Plane polarized, reflected light image of the polished and etched surface. The grain boundary (NW-SE trending line) is interpreted to have migrated to the SW, as the area to the NW of the boundary is virtually subgrain-free. The dark spots at the grain boundary and in the highly substructured grain are fluid inclusions that decrepitated during sample preparation. Note the bulged shape of the grain boundary.
b) Same area photographed in plane polarized transmitted light. The area interpreted as swept by grain boundary migration is less intensively coloured by 35 ºC gamma-irradiation. Note that the subgrain boundaries were not decorated by the irradiation. Also note that the area, which was swept by the migrating grain boundary, is fluid-inclusion-free.


fig 13
Fig. 13. Photomicrograph of sample gamma-irradiated at 35 ºC from salt layer D (423.3 m). Minor grain boundary migration occurred in the grain in the middle of the image (see arrows), where the former {100} crystal face (grain boundary) migrated. The area swept by grain boundary migration is less colored. Grain boundaries occur as dark lines. Note that subgrain boundaries were not decorated with Na-precipitates by 35 ºC irradiation. Note also that the thickness of the thin section is approximately 2 mm, so grain boundaries not perpendicular to the plane of image appear thicker. Plane polarized transmitted light image, image width is 2.4 cm.



Fig. 14. Photomicrographs show microstructure of sample from salt layer A (457.5 m) gamma-irradiated at 100 ºC. The milky, substructured, fluid-inclusion-rich grain at the bottom part is interpreted as old grain, which is replaced by a new, clear, strain-free grain (upper left corner). Note the presence of decorated planes, interpreted to be glide planes, in both the old and new grains. Also note how the decorated glide planes change direction at some subgrain boundaries due to the slight misorientation of the subgrains. Plane polarized transmitted light image, image width 6 mm.



Fig. 15. Microstructures of sample from salt layer A (457.5 m, gamma-irradiated at 100 ºC), photographed in plane polarized transmitted light. In both images the grain boundaries occur as dark curves, while the white, milky areas are remnants of chevron grains.
a) The clear, fluid-inclusion-free regions could be explained by a recrystallization mechanism of grain boundary migration. However, the presence of subgrains both in the chevrons and in the fluid-inclusion-free region implies that the structure is syndepositional and is a product of dissolution and precipitation processes. Image width is 16 mm.
b) Similar structure as shown in image “a”. Milky, fluid-inclusion-rich (upper left corner) and clear, fluid-inclusion-free halite are separated by a sharp, curved line. The presence of subgrains in both the milky and clear parts is strong evidence that the structure was not developed by grain boundary migration. Note the serrated grain boundary at the right side. Image width is 1 mm.



Fig. 16. Photomicrographs show the microstructures of a sample from salt layer A (457.5 m, gamma-irradiated at 100 ºC), photographed in plane polarized transmitted light. In both images the grain boundaries occur as dark curves, while the white, milky patches are remnants of fluid inclusion-rich chevron grains.
a) Elongated subgrains indicative of grain boundary migration. Grain boundary migration (see arrows) erases the old, milky, fluid-inclusion-rich part of the old, deformed grains and produces strain-free regions. Since the strain-free grains are free of fluid inclusions, it seems very likely that the fluid inclusions in the deformed and consumed grain were transformed into the grain boundaries during grain boundary migration recrystallization. Image width is 8 mm.
b) Similar structure as shown in the “a” image. A grain with relatively little substructure replaces the primary, fluid-inclusion-rich, highly substructured grain. Here again, the migrating grain boundary seems collecting the fluid inclusions at the grain boundary. Fluid inclusions at the grain boundary occur as dark spots in the image. Image width is 1 mm.

An important observation is that the boundary between milky, fluid-inclusion-rich and clear, fluid-inclusion-free halite can be either inside a grain or at a grain boundary (Figs. 15 and 16). If it is inside a grain the clear, fluid-inclusion-free part is in some cases subgrain-rich, in some cases subgrain-free (Figs. 15 and 16). In many cases grain boundaries are in contact with a milky, fluid-inclusion-rich part of a grain.



5. Interpretation and discussion
5.1 Syndepositional (primary) structures
All samples contain structures interpreted as syndepositional in origin. The truncated chevron grains and the vertically elongated clear halite grains are features which compare well with those reported by Lowenstein and Hardie (1985), who described the characteristic features of salt pan evaporites. Ephemeral salt pans are normally dry, shallow depressions, filled with layered halite. The halite layers evolve by repeated cycles of desiccation, flooding, evaporative concentration and re-desiccation. The flooding stage brings unsaturated floodwater into the dry salt pan, and converts it into a brackish lake. The unsaturated water partly dissolves the old salt crust, preferentially at grain boundaries, producing a karst-like surface. As the floodwater becomes more saturated with respect to NaCl due to the dissolved salt and continuous evaporation, crystallization starts at the brine surface, where halite hopper crystals and plates form, which sink to the bottom. On the bottom of the brine pool, the sunken crystals and the old, eroded salt crystals start to grow in crystallographic continuity with seed crystals by precipitation from the concentrated floodwater. Growth competition between the overgrowing crystals produces vertically elongated chevron halite crystals (Nollet et al., 2005), with upward directed apices. In the desiccation stage, as the saline lake dries out, void-filling clear halite crystallizes from the residual groundwater brine. Lowenstein and Hardie (1985) argue that the salt pan environment is best identified by the presence of dissolution features. Clear halite truncating chevron halite grains, and vertically elongated clear halite grains in the studied cores are interpreted as dissolution structures and regarded as strong evidence for the salt pan environment.

Syndepositional structures documented in recent ephemeral salt pans have been extensively reported in ancient salt deposits (Dellwig, 1955; Wardlaw and Schwerdtner, 1966; Casas and Lowenstein, 1989; Benison and Goldstein, 2001; for overview see Warren, 1999). Our interpretation of the synsedimentary structures in the Röt salt layers is consistent with these previous works (Fig. 17).
The lowermost and uppermost samples (layer A 457.5 m (Fig. 9) and layer D 423.3 m) are rich in chevrons and primary, dissolution related-features: chevron halite layers that formed in the salt pan during concentration of the brine and clear halite that crystallized in former voids during the desiccation phase. The relative abundance of clear halite in some samples may imply that the chevron layer underwent repeated episodes of dissolution and precipitation.
In the C layer (443.5 m) the elongated, truncated chevron grains, dissolution pipes (Fig. 7), and the presence of large (1.5 to 3 cm), fluid-inclusion-poor, blocky halite grains above the truncation surface marked by anhydrite partings are also features which fit well with the salt pan model. Thus the chevron halite layers are interpreted as forming by competitive growth of bottom-nucleated crystals in the saline lake phase, with the void-filling clear halite crystallized during the desiccation phase. The horizontal truncation surface is interpreted to be a result of dissolution caused by arrival of unsaturated flood water. The presence of chevrons in the bottom layers may be explained by the shallow water environment, since water shallow enough allows fluctuations in NaCl saturation level and thus fluctuations in growth rate. The rhythmic alternation in growth rate results in the development of alternating fluid-inclusion-rich and fluid-inclusion-poor bands corresponding to high and low growth rate periods (Roedder, 1984; Handford, 1990). The blocky crystals are interpreted as forming in concentrated, relatively deep water (deep enough to prevent the development of fluid inclusion bands). The intercalated small hoppers in the blocky grains are probably sunken hopper crystals formed at the brine-air interface (Arthurton, 1973). The second sample studied from the C layer (430.9 m) that is made up entirely of large, fluid-inclusion-free, blocky crystals, has the same appearance as that found in the upper part of the C layer (443.5 m) sample (Fig. 7). Those large blocky halite crystals are also interpreted to have formed in concentrated, relatively deep water in the evaporative concentration stage of the salt pan.



Fig. 17. Diagram illustrating the evolution of the Hengelo rock salt. (a) Overgrowth of seed crystals with different orientation in the evaporative concentration stage of the salt pan cycle. Favourably-oriented crystals override those which have unfavourable orientation. Growth layering (black parallel lines) is a sequence of alternating fluid-inclusion-rich and fluid-inclusion-poor bands. The growth of the crystals continues until the saline lake dries completely out (desiccation stage) (b) Formation of dissolution surface and cavity due to the arrival of undersaturated water during the flooding stage. (c) Evaporation concentrates the undersaturated water, and thin crust of gypsum crystals precipitates from the concentrated water. With further evaporation the system arrives back to the evaporative concentration stage and the gypsum layer is overgrown by a new halite layer. Dissolution voids are filled with clear halite. (d) Layered halite rock evolved by repeating salt pan cycles. The remaining pore spaces are completely filled with clear halite at shallow burial depth. (e) As the salt deforms, the deformation-induced grain boundary migration recrystallization starts to erase the primary structures. Due to this process, the identification of synsedimentary features becomes progressively more difficult as deformation and recrystallization continue (modified after Shearman, 1970).

Extreme truncation of chevrons by clear halite, a common feature in many ancient salt pan evaporites, is best illustrated by the example of the C core (443.5 m) (Fig. 7). In this core only a few percent (up to 10 %) of chevrons are preserved, with the rest made up of clear halite filling dissolution pipes. In earlier work on ancient bedded halite (e.g. Wardlaw and Schwerdtner, 1966; Wardlaw and Watson, 1966), it was argued that it was impossible to produce such mature chevron halite layers by synsedimentary dissolution-reprecipitation alone. The main argument against the dissolution-reprecipitation process was the assumption that the entire salt layer would collapse if it were pervasively truncated by dissolution pipes. These authors thus considered recrystallization as a likely additional effect for producing those features. However, others (Shearman, 1970; Lowenstein and Hardie, 1985; Casas and Lowenstein, 1989) pointed out that the structure can entirely be explained by a repeated dissolution-reprecipitation processes, and argued that recrystallization is not necessarily required for producing that feature. In this study, the lack of evidence for extensive deformation-induced recrystallization processes (see below) in the truncated chevron layer supports the view that deformation-induced recrystallization was not involved in development of those structures.

5.2 Deformation-related structures
The majority of the grains in the samples studied contain subgrains (Figs. 7 and 9). The polygonal shape of the subgrains suggests deformation dominated by climb-controlled creep (Senseny et al., 1992), although some microstructural evidence may point to the contribution of cross-slip is also present (Fig. 10). The lack of evidence for flattened grains points to a total strain of <10 % (Jackson, 1985). Geological maps of the Twenthe-Rijn area (Geluk and Duin, 1997; Doornenbal et al., 2002) show that the salt is displaced along NW-SE trending faults, and the most plausible assumption is to associate the deformation of the salt layers with these faults and thus to the Cretaceous inversion tectonic event (de Jager, 2003).
Subgrain-free regions, elongated subgrains at grain boundaries and bulged grain boundaries provide clear evidence for strain energy driven grain boundary migration (Drury and Urai, 1990; Bestmann et al., 2005). Observations on both etched and irradiated samples show that new, less substructured halite is replacing highly substructured grains (Figs. 9, 11-15). In many cases the migrating grain boundaries consume old, milky, fluid-inclusion-rich parts of a grain (Figs. 12 and 16). Since the area that was swept by grain boundary migration is fluid-inclusion-free, it seems very likely that this process transfers brine into the grain boundaries and provides conditions for pressure solution creep (Urai, 1983; Urai et al., 1986; Spiers et al., 1990).
Although grain boundary migration is common, the recrystallized volume is only a few volume percent (Fig. 9).
Sizes of entirely substructure-free grains vary between the size of a subgrain and a few millimeters (Figs. 9 and 11), suggesting that formation of new high-angle grain boundaries by progressive subgrain rotation is also possible in salt in nature (Drury and Urai, 1990). To date, subgrain rotation has only been documented in halite experimentally deformed dry and at high temperature (Guillopé and Poirier, 1979; Franssen, 1993). Although microstructural evidence reported here points to the presence of subgrain rotation recrystallization, an EBSD analysis of selected regions (e.g. Fig. 11) is necessary and in progress to address this question in more detail.

Water content in the Röt salt is high owing to the numerous primary fluid inclusions (e.g. Fig. 9) and, as shown above, evidence for grain boundary migration is present. It would be expected that as a grain boundary sweeps through a milky, fluid-inclusion-rich area and collects the fluid inclusions, this process amplifies (since fluid increases grain boundary mobility) and continues until the majority of the grains recrystallize and the driving force is eliminated. Considering that the deformation (and creation of driving force) was at least 65 Ma years ago, it is puzzling to see that so many of the deformed grains are not recrystallized, because halite recrystallizes readily at room temperature (Schenk and Urai, 2004). One hypothesis for this phenomenon could be the change in the burial depth, and thus a change in temperature. According to the burial history of Röt strata in vicinity of Hengelo (NITG, 1998), the salt strata were deformed at a maximum burial depth of 1.2 to 1.5 km (Jurassic and Cretaceous). Keeping in mind the recent depth of the beds (<500 m), however, it is unlikely that this drop in temperature (about 30 ºC) had a major effect on the rate of recrystallization. Another explanation could be a reduction in driving force (recovery, subgrain formation) or a reduction in grain boundary mobility due to a change in fluid structure at grain boundaries (Urai et al., 1986; Peach et al., 2001; Schenk and Urai, 2004).

5.3 Differential stress and strain rate calculations
In most materials which deform by dislocation creep processes, the steady state subgrain size is inversely proportional to the stress difference (σ1 - σ3) and independent of other variables. The relation is:

(1) D = k σ -m

where D is the average subgrain size in µm, k and m are material constants and σ is the differential stress in MPa. Two recent datasets exist for experimentally deformed rock salt (Carter et al., 1993; Franssen, 1993) where subgrain size and stress were measured. Plotting those data together, the parameters of the least squares fit line become k = 215 and m = -1.15 (correlation coefficient = 0.90). In this paper we used these values for the stress calculation. For the subgrain diameter (D) calculation, subgrain boundaries from reflected light photographs of etched thin sections and transmitted light photographs of irradiated thin sections were digitized and then analyzed with the NIH Image software (http://rsb.info.nih.gov/nih-image/index.html). The software calculates the area of every subgrain, and the D value is obtained by calculating the diameter of a circle equivalent in area to that of the subgrain.
The calculated differential stress values (Table 1) are in good agreement with those reported for various bedded rock salts by Carter et al. (1993), and imply that the Hengelo salt underwent deformation at stresses of 0.5 to 1 MPa.

Table 1. Differential stress from subgrain size – σ (MPa) = 107 D -0.87 (µm).
Sample No. of analyzed subgrains Mean subgrain diameter (µm) 1 S.D.
Calculated mean values for differential stress in MPa (95% confidence)
“D” layer (423.3 m) 301 289 143 0.60– 0.97
“C” layer (430.9 m) 352 327 208 0.53 – 0.88
“C” layer (443.5 m) 343 391 223 0.45 – 0.77
“A” layer (453.5 m) 290 346 189 0.50 – 0.85


Theoretically, based on the distribution of differential stress magnitudes over the studied salt layers, it would be possible to characterize the type of deformation (simple shear vs. pure shear). This is because in case of simple shear one would expect the same differential stress values over the whole salt succession, while in case of pure shear the highest differential stress values would be expected at the top and at the bottom, the lowest values at the middle of the deforming salt succession.
Although the calculated mean differential stress values are slightly different in the four samples studied, the differences are not significant due to the overlapping range of predicted values (Figs. 18 and 19), preventing us from characterizing the type of the deformation (e.g. simple shear vs. pure shear).

During deformation of rock salt, creep processes are controlled by cross-slip (CS), climb (CL) and pressure solution (PS) (Spiers and Carter, 1996). Elaborate experiments on rock salt provided constitutive equations for dislocation creep processes (Carter et al., 1993) and for pressure solution (Spiers et al., 1990). Assuming that dislocation creep processes and pressure solution act in parallel, steady-state flow of natural rock salt can be approximated by the following equation:

(2) ε = εCS + εCL + εPS

where ε is the total strain rate in s-1, εCS and εCL are strain rates from dislocation creep processes, and εPS is strain rate from pressure solution (by definition zero for dry salt). Microstructures observed in this study are consistent with climb-controlled and the fluid assisted diffusional-creep (c.f. Fig. 1 of Schenk and Urai, 2004). Thus it seems reasonable to use the corresponding constitutive equations for climb controlled creep (equation 3) and for pressure solution creep (equation 4) to calculate total strain rate.

(3) εCL = 8.1 × 10-5 exp (-51600 / RT)(σ1 - σ3)3.4   after Carter et al. (1993)

(4) εPS = 4.7 × 10-4 exp (-24530 / RT)(σ1 - σ3) / TD3   after Spiers et al. (1990)

where strain rate (ε) is expressed in s-1, the pre-exponential constant is in MPa-ns-1, apparent activation energy is in Jmol-1, Boltzmann’s gas constant (R) is in Jmol-1K-1, temperature (T) is in ºK, differential stress (σ1 - σ3) is in MPa and grain size (D) is in mm. For the calculations we propose that the deformation occurred at a depth of ~1.5 km in the Cretaceous time. Since we are lacking of paleogeothermal data for the study area, we used that of present day (~30-35 ºC/km; Haenel, 1980), and thus T=323 ºK.
Calculated strain rate values (Table 2) show that in finer grained samples, solution precipitation creep contributes to total strain rate at least in the same order of magnitude as dislocation creep processes. In the relatively coarse-grained sample (C layer, 430.9 m) dislocation creep is the main deformation mechanism.

Table 2. Strain rate calculated from equation (3) and equation (4). Total strain rate ( ε ) is the sum of the two values (equation 2).

Sample No. of analyzed grains Mean grain diameter (mm) 1 S.D. εCL (s-1) εPS (s-1) ε (s-1)
“D” layer (423.3 m) 24
6.01
2.08
6.31x10-14 – 3.33x10-13 4.31x10-13 – 7.03x10-13 4.94x10-13 – 1.04x10-12
“C” layer (430.9 m) 3
25.91
12.59
4.20x10-14 – 2.40x10-13 4.77x10-15 – 7.96x10-15 4.68x10-14 – 2.47x10-13
“C” layer (443.5 m) 15
8.80
5.28
2.34x10-14 – 1.49x10-13 1.02x10-13 – 1.76x10-13 1.26x10-13 – 3.25x10-13
“A” layer (453.5 m) 52
4.91
2.82
3.51x10-14 – 2.07x10-13 6.25x10-13 – 1.12x10-12 6.97x10-13 – 1.32x10-12




Fig. 18. Logarithmic subgrain size vs. logarithmic differential stress of experimental data from Carter et al. (1993) and Franssen (1993). The least squares fit line of the combined data set was used to calculate the differential stress for the Hengelo samples (Table 1).



Fig. 19. Graph shows the calculated differential stress and the strain rate distribution over the inspected salt layers. For discussion see text.

Experiments on bedded and domal salts (Wawersik and Zeuch, 1986; Hunsche et al., 2003) showed that while the form of the constitutive laws governing creep rate is similar, actual creep rates in different types of salts vary over a factor of ±10. This variation in creep rate can be attributed to secondary phases, solid solution impurities, grain size and dislocation density (Pfeifle et al., 1995). In our calculations we took a factor of 10 to be a plausible range. The uncertainty in the creep rate is too large to allow detection of significant differences in strain rate across the layers (Fig. 19).


6. Conclusions
In this paper, with the aid of microstructural analysis, we differentiated between primary (synsedimentary) and secondary (deformation-related) microstructures present in the Röt salt. Most of the primary structures are still preserved and modifications by recrystallization have been subordinate. In our samples, recrystallization of salt is incipient, but it shows how associated processes erase primary structures, making identification of synsedimentary features progressively more difficult as deformation and recrystallization continues. Nevertheless, in the case of slightly deformed salts, there are already microstructures, e.g. grain boundary migration-type microstructures (c.f. Fig. 15 and 16), which, without a careful microstructural analysis, could be misinterpreted. One other important implication of this paper is the evidence for a process (grain boundary migration) that transports and concentrates primary fluid-inclusion brines into grain boundaries. This implies that the amount of brine at the grain boundaries increases as the grain boundaries sweep through a fluid-inclusion-rich area, thus giving rise to pressure solution creep, which accordingly can contribute to the total strain rate at least in the same order of magnitude as dislocation creep processes. Keeping in mind that many of the bedded halites are of salt pan origin and thus rich in fluid inclusions, it is likely that pressure solution creep is significant in halite alongside dislocation processes, at least in the early phases of halokinesis.


Acknowledgements
The authors thank Wim Paar (Akzo Nobel company) for providing the salt core samples, Manfred Thomé at the Jülich Forschungszentrum for carrying out the gamma-irradiation and H. W. den Hartog (University of Groningen) for helpful discussion on gamma-irradiation. This work was performed as a part of the SPP 1135 project (nr. UR 64/5-1-2), and was financed by the DFG (Deutsche Forschungsgemeinschaft). The manuscript benefited from thorough reviews by Timothy Diggs and Chris Spiers.


References

Arthurton, R.S., 1973. Experimentally produced halite compared with Triassic layered halite-rock from Chesire, England. Sedimentology, 20: 145-160.
Benison, K.C. and Goldstein, R.H., 2001. Evaporites and siliciclastics of the Permian Nippewalla Group of Kansas, USA: a case for non-marine deposition in saline lakes and saline pans. Sedimentology, 48: 165-188.
Bestmann, M., Piazolo, S., Spiers, C.J. and Prior, D.J., 2005. Microstructural evolution during initial stages of static recovery and recrystallization: new insights from in-situ heating experiments combined with electron backscatter diffraction analysis. Journal of Structural Geology, 27(3): 447-457.
Carter, N.L., Horseman, S.T., Russell, J.E. and Handin, J., 1993. Rheology of rocksalt. Journal of Structural Geology, 15(9-10): 1257-1271.
Casas, E. and Lowenstein, T.K., 1989. Diagenesis of saline pan halite: comparison of petrographic features of modern, Quaternary and Permian halites. Journal of Sedimentary Petrology, 59(5): 724-739.
de Jager, J., 2003. Inverted basins in The Netherlands, similarities and differences. Geologie en Mijnbouw, 82(4): 355-366.
Dellwig, L.F., 1955. Origin of the salina salt of Michigan. Journal of Sedimentary Research, 25(2): 83-110.
Doornenbal, H., Duin, E.J.T. and de Haan, H., 2002. Actualisering van de bestaande geologische kaarten van de concessie Adolf van Nassau en van de concessie Twenthe-Rijn, Buurse en Uitbreiding Twenthe-Rijn. TNO report: NITG 02-194-C, Nederlands Instituut voor Toegepaste Geowetenschappen TNO.
Drury, M.R. and Urai, J.L., 1990. Deformation-related recrystallization processes. Tectonophysics, 172: 235-253.
Franssen, R.C.M.W., 1993. Rheology of synthetic rocksalt with emphasis on the influence of deformation history and geometry on the flow behaviour. PhD Thesis, Rijksuniversiteit Utrecht.
Garcia Celma, A. and Donker, H. (Editors), 1996. The Effect of Gamma Radiation in Salt, EUR-Report 16743EN.
Geluk, M.C. and Duin, E.J.T., 1997. Kartering Twenthe-Rijn, Uitbreiding Twenthe-Rijn en Buurse Concessie. TNO report: NITG 97-69-C, Nederlands Instituut voor Toegepaste Geowetenschappen TNO.
Geluk, M.C. and Röhling, H.-G., 1997. High-resolution sequence stratigraphy of the Lower Triassic "Buntsandstein" in The Netherlands and northwestern Germany. Geologie en Mijnbouw, 76: 227-246.
Guillopé, M. and Poirier, J.P., 1979. Dynamic recrystallization during creep of single crystalline halite: an experimental study. Journal of Geophysical Research, 4: 5557-5567.
Haenel, R., 1980. Atlas of Subsurface Temperatures in the European Community. Commission European Community, Brussels.
Handford, R.C., 1990. Halite depositional facies in a solar salt pond: A key to interpreting physical energy and water depth in ancient deposits? Geology, 18: 691-694.
Hardie, L.A., Lowenstein, T.K. and Spencer, R.J., 1983. The problem of distinguishing between primary and secondary features in evaporites. In: B.C. Schreiber and H.L. Harner (Editors), Sixth Symposium on Salt. Salt Institute, Alexandria, Virginia, USA, pp. 11-39.
Harsveldt, H.M., 1980. Salt resources in The Netherlands as surveyed by AKZO. In: A.H. Coogan and L. Hauber (Editors), Fifth Symposium on Salt, Hamburg, Germany, pp. 65-81.
Hunsche, U., Schulze, O., Walter, F. and Plischke, I., 2003. Projekt Gorleben. Thermomechanisches Verhalten von Salzgestein. 9G2138110000, BGR, Hannover.
Jackson, M.P.A., 1985. Natural strain in diapiric and glacial rock salt, with emphasis on Oakwood Dome, East Texas. Report of Investigations, 143, 73 pp.
Joffé, A., 1928. The physics of crystals. McGraw-Hill, New York.
Kovalevych, V., Peryt, T.M., Beer, W., Geluk, M. and Halas, S., 2002. Geochemistry of Early Triassic seawater as indicated by study of the Rot halite in The Netherlands, Germany, and Poland. Chemical Geology, 182(2-4): 549-563.
Lowenstein, T.K. and Hardie, L.A., 1985. Criteria for the recognition of salt-pan evaporites. Sedimentology, 32: 627-644.
NITG, 1998. Geological Atlas of the Subsurface of The Netherlands, Explanation to map sheet X: Almelo-Winterswijk, Nederlands Instituut voor Toegepaste Geowetenschappen TNO, Haarlem, 143 pp.
Nollet, S., Urai, J.L., Bons, P.D. and Hilgers, C., 2005. Numerical simulations of polycrystal growth in veins. Journal of Structural Geology, 27(2): 217-230.
Peach, C., Spiers, C.J. and Trimby, P.W., 2001. Effect of confining pressure on dilatation, recrystallization, and flow of rock salt at 150 ºC. Journal of Geophysical Research, 106(13315-13328).
Pfeifle, T.W., Vogt, T.J. and Brekken, G.A., 1995. Correlation of chemical, mineralogic and physical characteristics of Gulf Coast dome salt to deformation and strength properties. Contract No. 1-92, RESPEC Inc., Rapid City.
Przibram, K., 1954. Irradiation colours in minerals. Endeavour, 13(49): 37-41.
RGD, 1993. Geologische Kaart van Nederland 1:50 000, Blad Almelo Oost/Denekamp (28O/29), Rijks Geologische Dienst, Haarlem.
Roedder, E., 1984. The fluids in salt. American Mineralogist, 69: 413-439.
Schenk, O. and Urai, J.L., 2004. Microstructural evolution and grain boundary structure during static recrystallization in synthetic polycrystals of Sodium Chloride containing saturated brine. Contributions to Mineralogy and Petrology, 146(6): 671-682.
Senseny, P.E., Hansen, F.D., Russell, J.E., Carter, N.L. and Handin, J., 1992. Mechanical behaviour of rock salt: phenomenology and micromechanisms. Int. J. Rock Mech. Min. Sci & Geomech. Abstr., 29(4): 363-378.
Shearman, D.J., 1970. Recent halite rock, Baja California, Mexico. Transaction of Mining and Metallurgy, 79B: 155-162.
Spiers, C.J. and Carter, N.L., 1996. Microphysics of rocksalt flow in nature, Fourth Conference on the Mechanical behaviour of Salt, The Pennsylvania State University.
Spiers, C.J. et al., 1990. Experimental determination of constitutive parameters governing creep of rocksalt by pressure solution. In: R.J. Knipe and E.H. Rutter (Editors), Deformation Mechanisms, Rheology and Tectonics. Geol. Soc. London Spec. Publ., pp. 215-227.
Urai, J.L., 1983. Deformation of wet salt rocks. PhD Thesis, Universiteit Utrecht, 221 pp.
Urai, J.L., Spiers, C.J., Peach, C., Franssen, R.C.M.W. and Liezenberg, J.L., 1987. Deformation mechanisms operating in naturally deformed halite rocks as deduced from microstructural investigations. Geologie en Mijnbouw, 66: 165-176.
Urai, J.L., Spiers, C.J., Peach, C.J. and Zwart, H.J., 1985. A laboratory investigation into the interaction of recrystallization and radiation damage effects in polycrystalline salt rocks, HPT Laboratorium, Insititut voor Aardwetenschappen, Utrecht.
Urai, J.L., Spiers, C.J., Zwart, H.J. and Lister, G.S., 1986. Weakening of rocksalt by water during long term creep. Nature, 324: 554-557.
van Lange, M.W.P., 1994. The development, geology and lithology of the Central-Northern part of the Hengelo rock salt solution mining area and its geotechnical characterisation. Memoirs of the Center of Engineering Geology in The Netherlands, No. 126, 77 pp.
van Opbroek, G. and den Hartog, H.W., 1985. Radiation damage of NaCl: dose rate effects. Journal of Physics. C: Solid State Phys., 18: 257-268.
Wardlaw, N.C. and Schwerdtner, W.M., 1966. Halite-Anhydrite seasonal layers in the Middle Devonian Prairie Evaporite Formation, Saskachewan, Canada. Geological Society of America Bulletin, 77: 331-342.
Wardlaw, N.C. and Watson, D.W., 1966. Middle Devonian salt formations and their bromide content, Elk Point Area, Canada. Canadian Journal of Earth Sciences, 3: 263-275.
Warren, J., 1999. Evaporites - Their evolution and economics. Blackwell Science Ltd, Oxford, 437 pp.
Wawersik, W.R. and Zeuch, D.H., 1986. Modeling and mechanistic interpretation of creep of rock salt below 200 C. Tectonophysics, 121: 125-152.