Alternative SEM-BEX Imaging of Rock Mini-Cores (Carbonate and Siliciclastic): Manual and Semi-Automated Acquisition

Abstract

An understanding of the textures (grain size, grain shape, porosity, etc.), composition (mineralogy), and distribution of constituent components of geological materials such as carbonate and siliciclastic sedimentary rocks is essential in their classification, interpretation, and significance in terms of their geomechanical strength and liquid/gas storage potential. In terms of scanning electron microscopy (SEM), this is limited to relatively flat areas of selected rough surfaces, or the analysis of polished thin sections. Here, we illustrate a new technique that can image large areas of the external surface of mini-cores (approximately 10 mm or smaller in diameter) drilled from carbonate and siliciclastic rock samples. The technique utilises a specially developed horizontal rotation stage within an SEM and allows the collection of high-resolution images that can be reconstructed into realistic surface representations of the mini-core surfaces. Elemental data (representative of mineralogy) can also be added using a combined backscattered electron and X-ray (BEX) detector. Currently, these reconstructions can be used as a useful tool for the analysis of both carbonate and siliciclastic geological materials. Further work may allow such reconstructions to aid in the improvement of resolution in micro-CT scans and the direct identification of mineral phases within such scans.

1. Introduction

Scanning electron microscopy (SEM) and X-ray tomography (XRT, i.e., micro- and nano-CT) are both imaging techniques that are commonly used in the examination and interpretation of geological materials such as mudrocks, sandstones, and carbonates. SEM analysis is typically carried out on relatively flat and often polished geological materials, while micro- and nano-CT are usually performed on cylindrical cored geological samples.

SEM imaging of relatively flat sample surfaces (Figure 1A), performed in conjunction with energy dispersive X-ray (EDX) analysis, can characterise the composition and therefore the mineralogy of materials over large millimetre- to centimetre-sized areas (Figure 1A,E). This presents data as easily interpreted colour-coded compositional maps, illustrating the distribution of mineral phases [1], which is uniquely informative. Importantly, SEM image acquisition can easily obtain a sub-micron resolution, sufficient to resolve clays (2 µm or smaller [2]) and many smaller pores and textural features which are typically very difficult to view with any degree of confidence.

Micro-CT can construct high-resolution 3D images of internal features and exterior surfaces (Figure 1B,F). Although nano-CT can resolve structures at the sub-micron (nanometric) scale, this is limited to volumes typically under 100 µm³ (Figure 1C,G), while micro-CT can scan larger volumes (millimetres to centimetres) but at best has a pixel resolution of 1–4 µm (Figure 1B,F). The latter is not suitable for the high-resolution analysis of clay fabrics or smaller pores. In some cases, it has proven possible to use artificial intelligence (AI) to utilise limited area nano-CT scans to produce enhanced on-the-fly reconstructions at a higher-resolution from micro-CT, although this has only currently been illustrated for relatively predictable materials such as batteries [3]. Although grayscale variation within micro-CT images can be used to infer mineral/phase differences, it is a blunt tool in comparison to characterising the composition of analysed material in an SEM-EDX analysis.

Therefore, neither standard SEM-EDX analysis nor micro- or nano-CT individually fully characterise often unpredictably heterogeneous geological materials such as sandstones and carbonates at the sub-micron to micron scale over large areas. The latter is of current relevance for characterisation of such materials in relationship to carbon capture and storage (CCS) and hydrogen storage. Both are of interest in respect to net zero carbon targets for the reduction in greenhouse gas emission and the control of global warming.

Here, we illustrate the acquisition of both imaging and elemental mapping around the exterior surface of carbonate and sandstone mini-cores (Figure 1D,H), within a large-chambered SEM, utilising a newly developed horizontal axis rotation stage (Figure 2). Data acquired can be reconstructed to display both textural features (i.e., grains, pores, cement, clays) and composition (i.e., elemental composition, mineralogy) in a 3D perspective around a partial circumference or the whole surface of any mini-core. We also compare these results with micro-CT reconstructions of the same material and discuss possible developments that should enhance the characterisation of such materials, particularly in respect to net zero carbon initiatives.

2. Materials and Methods

Two samples were the main focus of the current study. These were a 5 mm diameter mini-core of carbonate rock (Estaillades limestone, France), and a 4 mm diameter mini-core of a clay-rich sandstone (locality and age unknown). In addition, an unnamed 10 mm diameter feldspar-rich white sandstone mini-core was also briefly examined.

The general method follows Buckman [4] in using a specially designed horizontal rotation stage (Figure 2), within an FEI (Hillsboro, OR, USA) Quanta FEG 650 scanning electron microscope (SEM), equipped with an Oxford Instruments (Abingdon, UK) BEX detector for backscattered (BSE) imaging and qualitative elemental mapping with an X-Max^N 150 mm energy dispersive X-ray (EDX) detector.

Samples were superglued to a standard 10 mm aluminium SEM stub (Figure 3A,B) and mounted to the horizontal rotation stage using a built-in pin vice (Figure 2). In addition, one sample was also examined using a purpose-built mini-core holder (Figure 3C). The SEM was operated in low-vacuum mode (0.83 Torr), with a working distance of between 10 and 20 mm and an operating current of 20 kV. The sandstone sample illustrated in Figure 3B was sputter coated in gold and imaged in high-vacuum mode. All samples were imaged with the built-in Concentric Backscattered (CBS) detector, with a spot size of 4.5 and aperture selection of 6. The sandstone samples were also scanned using a combined backscattered (BSE) and X-ray BEX detector from Oxford Instruments. To maximise X-ray flux, the latter utilised a spot size of 5 in association with the widest available aperture setting (position 1). In both instances, a horizontal field of view of 750 µm was used, with a frame size of 1536 × 1024 pixels for backscattered (BSE) images and 1024 × 704 pixels for BEX images (combined Unity-backscattered (UBSE) and X-ray detector): representing a pixel resolution of under 1 µm (0.488 and 0.732 µm, respectively). A higher pixel resolution could be easily achieved by increasing the image pixel dimensions (e.g., 3072 × 2048, 6144 × 4096).

Individual images were either taken manually every 10° of rotation, or in the case of one BEX scan, the AZtec BEX software program (Oxford Instruments ^TM version 6.1, Abingdon, UK) was used to programme the automated acquisition of the images. In the latter case, imaging was conducted in high-vacuum mode, after gold-coating the sample, to allow the implementation of auto focus using standard secondary electron (SE) imaging.

The collected images were imported into Agisoft Metashape Standard ™ version 2.1.0 (Sankt Petersburg, Russia) multi-view 3D reconstruction software for post-processing of the collected BSE and BEX images and reconstructed following the workflow in Buckman [4]. Once images have undergone 3D reconstruction, models can be produced in surface mesh, solid, shadow, and texture (based on SEM images) modes (Figure 4 and Figure 5). Textured surface models were saved as 360° rotating .avi movies (see Supplementary Materials).

In addition, the two main mini-cores were scanned by micro-CT utilising an Rx Solutions Easy Tom 150 micro-CT system (Chavanod, France). The samples were each scanned at 100 kV, for approximately 1 h, over a single stage rotation, with 1440 X-ray projections recorded, with a pixel resolution of 5.5 µm. Data were then reconstructed using Rx propriety software (Xact64 revision 23.04 2023-06-27) and saved as a stack of 1358 .tiff images. Images were loaded into FIJI ™, where contrast and brightness for each image slice were enhanced and optimised, followed by the construction of models of the surface of the mini-cores using the FIJI plug-in 3D View. The latter were saved as 360° rotating .avi movies (see Supplementary Material), with representative still frames collected to illustrate image quality.

3. Results

3.1. SEM BSE-Based Reconstructions

Surface reconstructions based on backscattered electron (BSE) images were successfully produced for both the carbonate and clay-rich sandstone mini-cores (Figure 5). In both cases, full circumference reconstructions of selected regions were made (Figure 5B,D). All materials examined herein by BSE had a resolution of under 1 µm, with 0.488 µm per pixel being typical. This is enough to adequately illustrate all the observed grains/clays, as well as the pore size and structure, distribution, and textural features (Figure 5 and Figure 6).

3.2. BEX-Based Reconstructions

As with the reconstructions based on BSE images, BEX-based reconstructions also have a pixel resolution of under 1 µm, with 0.732 µm per pixel. BEX reconstructions of the clay-rich sandstone are illustrated in Figure 7 and Figure 8, with quartz grains, clays, iron-rich areas (pyrite) and calcium-rich areas. An enlargement of the sandstone surface (Figure 8) illustrates sub-micron resolution and elemental information from the BEX detector.

The partially scanned and reconstructed segment of a feldspar containing sandstone (Figure 9), clearly differentiates between orthoclase feldspar and altered orthoclase (blue versus green), as well as the occurrence of the formation of Fe-rich areas around the surface of some of the altered feldspar (yellow).

3.3. Micro-CT Reconstruction of Exterior Surfaces

Micro-CT reconstructions of the surface of both the carbonate and clay-rich sandstone mini-cores were successfully employed (Figure 10A,B). However, the reconstructed scans have a lower resolution (5.5 µm per pixel) than could be obtained through SEM imaging (see Figure 11). The maximum achievable resolution with the current samples is approximately 5 µm per pixel. In addition, reconstruction artefacts were also noted (Figure 10C,D).

4. Discussion

Although photogrammetry software designed for use with drones could construct realistic high-resolution 3D surfaces from SEM images taken around the circumference of mini-cores, some teething problems were observed. In some examples, the curvature of the mini-core surface was not accurately reconstructed, leading to the endpoints of each rotational scan not meeting (Figure 5D and Figure 7B). This will require further work, and possibly the use of alternative software. The latter problem stems from the lack of positioning data within the images, and the inability to reposition “cameras” (here, images) during the reconstruction phase. In addition, the reconstructions from the current software can generate spurious blank areas at the top and bottom of reconstructions (Figure 7A). These can be removed during the reconstruction phase but can also remove areas of interest from the imaged surface of the mini-core, requiring a degree of finesse. Anecdotally, it was also noted that the currently employed software program has a greater success rate utilising the greyscale BSE SEM images in comparison to the coloured BEX material. This can likely be improved through further trial and error in the settings of the reconstruction parameters. All the above may be resolved with the use of alternative reconstruction software programmes available (e.g., Adobe Substance 3D Sampler ™, Autodesk ReCap Pro™, RealityCapture™, and Zepher 3Dflow™), although none were tested herein.

On the positive side, the current examples illustrate that it is possible to capture and realistically reconstruct micron to sub-micron features through the SEM imaging of mini-core surfaces using the horizontal rotational stage, compared to the typical 1–5 microns (per pixel) of micro-CT for similar diameter mini-core. Additionally, it is possible to apply the same technique to larger core samples (at least 10 mm diameter), with the same degree of resolution, which would have correspondingly poorer resolution in micro-CT reconstructions. With the modification to the SEM mini-core holder, it may be practical to image core surfaces up to 20 mm in diameter, although this would also require an increase in the number of images taken per rotation of the stage (i.e., perhaps one image every 5° of rotation). Direct SEM imaging of the surface of the mini-cores also benefits from the lack of artefacts generated through surface reconstruction from micro-CT data (see Figure 10C,D).

This new method of obtaining SEM images and reconstructing the surface also benefits from direct elemental mapping and mineral identification through SEM-BEX. No BEX reconstructions of the carbonate mini-core were made during the current work, as initial investigation did not detect any major chemical variation within the sample. However, the use of SEM-BEX scans would be beneficial in cases of mixed dolomite–calcite carbonates, and where a substantial component of clays or other siliciclastic materials are present. In such cases, software programs such as AZtecMatch (Oxford Instruments ^TM, Abingdon, UK) could be used to directly identify mineral phases obtained through BEX maps [1]. Buckman and Krivtsov [1] also demonstrated how AZtecMatch can potentially identify other minerals such as feldspars, kaolin, and heavy minerals, which could equally be directly applied to reconstructions of sandstone mini-cores.

Using the current method, manual collection of individual images within the SEM is more time consuming than micro-CT. The usage of automation, in this case using an AZtec workflow for scanning core in the SEM, would be more time effective, although still slower than micro-CT scanning and reconstruction of the sample surface.

5. Further Work

This paper demonstrates the resolution and quality of data (images and elemental maps) that can be collected across the whole surface of geological mini-cores using SEM. This is a new technique that has several potential implications for use of such collected data in conjunction with micro-CT, that will form the basis of further developmental work.

In the first case, we plan to co-locate the SEM-based imaging with micro-CT reconstructed external surface models for geological mini-core materials. This has the potential to greatly enhance the surface resolution of micro-CT reconstruction, concentrating on using artificial intelligence (AI) and image analysis techniques to extrapolate and improve internal 3D resolution of pores, grains, clays, and cement. This would be similar to the “DeepRecon Pro” (ZEISS, Oberkochen, Germany) that uses AI in micro-CT reconstructions, achieving faster results at higher resolution [3]. Machine learning has previously been used in workflows involving geological thin sections [5], suggesting that such an approach would be beneficial in further developing the interpretation of the current material. This would enhance the use of such scans in the modelling of porosity and permeability, as used in projects such as carbon capture and storage (CCS), and hydrogen storage. Both are important in relation to net zero carbon aspirations, combating climate change through the reduction in greenhouse gas emissions.

In addition, BEX reconstructions and AZtecMatch will be used in a similar fashion with micro-CT scans, enabling the identification of actual mineral phases (i.e., plagioclase feldspar, orthoclase feldspar, quartz, dolomite, etc.) in 3D CT reconstructions. Although this could be performed manually, we would expect to harness machine learning/AI, consequently enabling a better understanding of the geomechanical and flow properties within mini-cores in relation to their constituent grain types, through the application of known physical and chemical properties (i.e., hardness, surface roughness, chemical bonding, wettability, etc.) of identified mineral phases and their distribution, and relationships with pore network characteristics within each mini-core. Other techniques are available to help interpret mineralogical composition or phase changes within micro-CT, including spectral X-ray tomography [6] and phase contrast imaging, with the latter often based on synchrotron X-ray tomography [7,8]. However, phase contrast imaging is more sensitive within softer organic/plastic materials (see [9]) and does not directly identify minerals/phases. The dual use and combination of SEM-EDX/BEX and micro-CT data is expected to have greater potential value in interpretation and subsequent modelling. Additionally, other methods of microanalysis/imaging, such as micro-X-ray fluorescence [10,11,12], may also be usefully combined with the SEM-BEX and micro-CT workflow illustrated herein.

SEM-EDX-based analysis, using more traditional methods, over limited ‘flat’ areas, could also be used to enhance micro-CT 3D reconstructions. However, the new technique outlined herein has the advantage of being able to automatically scan large areas of the external surface of mini-cores by SEM-EDX/BEX, and the potential to co-locate merge and enhance detail around the periphery of such cored material, adding important extra value to both the surface and internal reconstruction of micro-CT models.

Finally, future work will also include investigating other commercially available 3D reconstruction packages such as Adobe Substance 3D Sampler ™, Autodesk ReCap Pro™, RealityCapture™, and Zepher 3Dflow™, and the possibility of developing in-house software. This will ascertain the best workflow for the reconstruction of core-shaped (cylindrical) materials, with minimal artefacts, the highest resolution, and the best surface rendering potential.