Improving co-registration of geoscientific imaging datasets with micro-sized marker structures on rock samples

de Boer, Rosa A.; Wielens, Daan H.; de Groot, Lennart V.

doi:https://doi.org/10.5194/egusphere-2025-1692

Preprints

https://doi.org/10.5194/egusphere-2025-1692

Preprints

27 Jun 2025

| 27 Jun 2025

Improving co-registration of geoscientific imaging datasets with micro-sized marker structures on rock samples

Rosa A. de Boer, Daan H. Wielens, and Lennart V. de Groot

Abstract. Polished geological samples are frequently used in geoscientific research to investigate the chemical and physical characteristics of rocks. A broad range of imaging techniques is available to analyze such samples, but when combining datasets from multiple imaging techniques, an accurate co-registration of the datasets is often challenging. In this study, we investigate this issue in the context of Micromagnetic Tomography (MMT; De Groot et al., 2018, 2021). MMT combines surface magnetometry with computed tomography (CT) to analyze the magnetic state of rock samples. By combining the spatial and dimensional information of the magnetic grains in the samples with their magnetic surface expression, the individual magnetic moments per grain can be determined. This information can be used for paleomagnetic and rock-magnetic studies. Calculating the magnetic moments of the grains strongly depends on the correct co-registration of the two datasets, which proves to be challenging. In this study, we used two test samples for the application of micro-sized marker structures, to further develop the methodology of MMT. The marker structures are applied by microlithography and Nb-sputter coating, which are standard techniques used in the semiconductor industry. We determined that the marker structure application is possible on typical MMT samples. Marker structures larger than ca. 10 μm are clearly visible under the Quantum Diamond Microscope (QDM) used for the surface magnetometry. Given a sufficient marker structure thickness, they can also be observed in the CT scans used for determining the positions and shapes of the magnetic carriers. The marker structures are useful for identifying the orientation and location of the samples during measurements and can be used for scaling and mapping of the two datasets during data processing. Nb-marker structures do not fluoresce under the QDM, which means that no magnetic interference occurs during measurements. The application procedure is time-consuming but is valuable when a sample is lacking natural marker features, it makes the data processing time in MMT significantly faster, and more precise. This method can be useful for MMT, for Quantum Diamond Microscopy in general, and for broader geological applications that require visible anchor points for sample placement or marker structures for the co-registration of multiple datasets.

Received: 09 Apr 2025 – Discussion started: 27 Jun 2025

Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this paper. While Copernicus Publications makes every effort to include appropriate place names, the final responsibility lies with the authors. Views expressed in the text are those of the authors and do not necessarily reflect the views of the publisher.

Download & links

Rosa A. de Boer, Daan H. Wielens, and Lennart V. de Groot

Status: final response (author comments only)

RC1:
'Comment on egusphere-2025-1692', Anonymous Referee #1, 01 Jul 2025
Review: Improving co-registration of geoscientific imaging datasets with micro-sized marker structures on rock samples
General Comments:

The manuscript describes the application of UV-microlithography on rock samples for co-registration of image data. Small drill cores of volcanic rock samples with about 1-5 diameter were used to apply the tested approach. The results are analyzed by surface magnetometry and 3D NanoCT and showed that the process provides good results, even though a more critical assessment seems necessary (see comments below). The problem of co-registration is a critical issue when working with multiple image techniques, so this approach seems interesting for the scientific community.

In general, the results of the manuscript are convincing and the arguments in the discussion make sense. The text is quite understandable and well structured. This work is useful research and could be published after the comments have been addressed.
Specific comments:
Line 6 and 32: What is the difference between spatial and dimensional information? Does space not include all three dimensions?

Figure 1 and lines 27ff: During the first reading it was not clear where co-registration would be necessary at this stage, since MMT and imaging of magnetic sources might be the same procedure. I would suggest combining this sentence from the abstract (“MMT combines surface magnetometry with computed tomography (CT) to analyze the magnetic state of rock samples.”) with the main text of the introduction (lines 27f: “MMT combines magnetic scans with three-dimensional imaging of magnetic sources in a sample …”

Figure 2: It is a bit hard to understand the caption. Put the two sentences regarding the scale at the end of b) or at the end of the figure caption. What sample is it (RE5, RE25 or a different)?

Lines 95f: Why is petrography not important, if mineralogy can have an effect on the surface roughness? Can you at least give the main minerals and which worked well, since this plays a role later? You could also give a reference to some mineral hardness data.

Fig 4: Add the UV-arrows to the legend. F: You might want to put a semicolon after lift-off.

L. 110: For a person not familiar to the process the term “developed” might need an explanation, which – I believe – follows, but it is not clearly connected to the development.

L. 116: Does the Ar-plasma not heat the sample?

Ll. 168: If LED and QDM share their coordinate system and nanoCT are co-registered to QDM, why is co-registration still necessary? Please explain.

L. 175: Does “thickness” refer to the depth of the feature or to the width on the 2D surface of the sample?

L. 170: “we are now experimenting” change to “this work provides results on the experiments…”

Line 174f and Figure 5a-b: Could you shed some light on how you obtained the SEM images? Normally BSE or SE images are in grey scales, so what do the colors mean? To me 5b) looks like a reflected light image.

Ll. 199ff: You say that a thicker layer of Nb was not attempted because of the expected problems, but in line 203f you claim that you expect to successfully apply thicker structures. Please explain this contradiction. Why was no attempt made with thicker markers afterwards?

L. 214: “High” resolution normally refers to small pixel/voxel size, but I believe what you mean here is low resolution with high=large pixel size.

Ll. 214ff: How successful was the co-registration? Is the 50-µm marker structure enough? What improved compared to the conventional method? How can you quantify the improvement for such a tedious process? I would have expected a more critical assessment of the results showing e.g. a comparison of co-registration with and without using the markers. Than you could quantify the shift in pixels and show the results to the reader. With the current images this is not possible for the reader.

Technical corrections:

L. 55, 57: Omit commas

L. 173: Instead of comma use semicolon or colon.

L. 240: strongly magnetic samples instead of strong magnetic …

In general, there might be some redundant commas. Please check with a native speaker.
Citation: https://doi.org/10.5194/egusphere-2025-1692-RC1
- AC1: 'Reply on RC1', Rosa de Boer, 04 Sep 2025
  
  We thank the reviewer for their comments and recommendations on our manuscript. The reviewer raises a number of valid points that we addressed in the main text as follows:
  L6 and L32: This phrasing was used to discern between the location (spatial information) of the grains and the size (dimensional information) of the grains. This has been clarified in the text.
  Fig. 1 and L27f: This has been changed to ‘surface magnetometry data with CT data’ to clarify that two datasets are being used.
  Fig. 2: The sentence about scaling has been adjusted. The sample shown is an example image from a different study, used to illustrate the type of data used in MMT.
  L95f: It is true that the surface roughness can be affected by the mineralogy. However, the marker structures adhere well to the entire sample, which contains a mixture of minerals typical for a basalt. We therefore do not distinguish between which minerals are suitable for mask structures and which are not. The total assemblage works well, as long as it is properly polished according to geological thin section polishing standards.
  Fig. 4: This has been adjusted.
  L110: An explanation has been added.
  L116: The sample holder is cooled with water to keep the temperature of the substrate low enough. While there is no temperature sensor available in the particular setup used in this work, we know from other setups where a sensor is included on the sample holder that the sample holder’s temperature does not heat up to temperatures over 50°C. To give a more specific upper bound for the sputter setup used in this work, we note that another common resist used is PMMA (polymethacrylate). PMMA has a glass temperature of 105°C, which means that if the temperature of the sample would exceed that temperature, the resist would start to flow and features printed into the resist by lithography would be compromised. We do never observe this effect (even for much longer exposure to the Ar-plasma in longer etching/deposition processes). The magnetic interference of the heating is therefore considered to be negligible.
  L168: The LED of the QDM and the magnetic surface map of the QDM share a coordinate system, because they are obtained with the same camera. We co-register the CT data to the QDM data, which can be complicated. Therefore we need the marker structures. This has been clarified in the text.
  L175: This should indeed be ‘width’ and has been adjusted in the text.
  L170: This has been adjusted in the text.
  L174f and Fig. 5a-b: This is indeed incorrect. These are optical microscopy images; this has been adjusted in the text.
  L199f: There are practical problems in applying a thicker Nb-layer. Since it is quite a complicated and time-consuming procedure to apply the marker structures, we stayed on the safe side for this case study and only tried a thinner layer. However, because we experienced no problems whatsoever during the procedure, and because of the expertise and skills of the lab operator, we expect thicker layers to be possible with a bit of procedure optimization. This lies outside the scope of the current study. This has been clarified in the text.
  L214: This is indeed incorrect and has been adjusted in the text.
  L214f: The co-registration can be successfully achieved with the applied marker structures with a 50 µm width. When multiple structures are applied at known distances from one another, they provide a highly reliable reference system for aligning the datasets. The main improvement is that the co-registration process no longer depends on the presence of natural features on the samples, that are recognizable in both types of imaging. This ensures that co-registration is possible for every sample, including those where natural, recognizable features are insufficient. It is difficult to quantify the improvement in terms of pixel shift, because the outcome of co-registration is basically binary: it is possible, or it is not. This has been added to the text.
  L55, L57, L173, L240: These suggestions have been implemented where needed.
  
  Citation: https://doi.org/10.5194/egusphere-2025-1692-AC1
RC2:
'Comment on egusphere-2025-1692', Anonymous Referee #2, 18 Jul 2025

The manuscript: “Improving co-registration of geoscientific imaging datasets with micro-sized marker structures on rock samples” is well written and brings a nice concept of the chip industry to the field of earth sciences. Below you can find my comments:
1) My main concern is related to the results and discussion section. The results section is very short. Can you elaborate more on how efficient this technique was? Furthermore, was registration improved by the marker?
2) Can you elaborate more on the need for this complex method? For which type of samples can it be important? You explain it in the introduction, but it would be good if you could add some examples.
3) How realistic do you think that this technique will become accessible for people in the field? Is there easy access?
4) Wouldn’t a proper calibration of the instruments already solve a lot?
5) Line 93: Can you elaborate on the heating step? Is it a useful technique if it is harmful to the magnetic signal? Did you check if this affected your measurements?
6) Minor - Table 2: Do you know the source power? That is more commonly used than Source Current.
7) Line 215 – 220: You need to be careful with the resolution and voxel size, as they are not the same. If the resolution is 1.38 µm, you will be able to see an object 3 µm wide.

Citation: https://doi.org/10.5194/egusphere-2025-1692-RC2
- AC2: 'Reply on RC2', Rosa de Boer, 04 Sep 2025
  
  We thank the reviewer for their constructive comments and recommendations. The points raised are relevant, and we have addressed them in the main text as follows:
  1) The co-registration can be successfully achieved with the applied marker structures with a 50 µm width. When multiple structures are applied at known distances from one another, they provide a highly reliable reference system for aligning the datasets. The main improvement is that the co-registration process no longer depends on the presence of natural features on the samples, that are recognizable in both types of imaging. This ensures that co-registration is possible for every sample, including those where natural, recognizable features are insufficient. This has been added to the text.
  2) This has been added to the discussion.
  3) Microlithography is a well-established technique in fields such as the semiconductor industry. As demonstrated in our study, laboratories equipped for this type of research can readily apply the method to geological samples as well. In terms of accessibility, the costs are not exceedingly high when compared to the costs of high-end microscopy such as SEM or nanoCT, with which this approach is intended to be combined for co-registration. If no microlithography facilities are available to a user in the field, alternative options are available as outlined in the discussion, although they come with their own limitations.
  4) The instruments used for the different imaging techniques in this study are fundamentally different. The QDM is camera-based and the CT scanner relies on an X-ray detector. Although both instruments are properly calibrated, they acquire data in very different ways. In practice, this inevitably leads to mismatches in field of view and resolution, complicating accurate co-registration.
  5) Generally, heating below 100°C does not harm the magnetic signal, as this temperature is typically below the temperature range of interest in paleomagnetic research. For example, samples in our lab are routinely heated to this temperature to remove any surface contamination. However, if the low temperature magnetic behavior of a sample is the point of interest, it is important to consider the potential effects of heating. This has been clarified in the text.
  6) The table lists the parameters that can be adjusted as part of the measurement configuration. These include source current and source voltage.
  7) It is indeed true that resolution and voxel size are not the same, this has been clarified in the text.
  
  Citation: https://doi.org/10.5194/egusphere-2025-1692-AC2

Rosa A. de Boer, Daan H. Wielens, and Lennart V. de Groot

Viewed

Total article views: 487 (including HTML, PDF, and XML)

HTML	PDF	XML	Total	BibTeX	EndNote
440	31	16	487	30	42

HTML: 440
PDF: 31
XML: 16
Total: 487
BibTeX: 30
EndNote: 42

Views and downloads (calculated since 27 Jun 2025)

Month	HTML	PDF	XML	Total
Jun 2025	45	5	3	53
Jul 2025	66	17	10	93
Aug 2025	121	4	1	126
Sep 2025	208	5	2	215

Cumulative views and downloads (calculated since 27 Jun 2025)

Month	HTML	PDF	XML	Total
Jun 2025	45	5	3	53
Jul 2025	66	17	10	93
Aug 2025	121	4	1	126
Sep 2025	208	5	2	215

Viewed (geographical distribution)

Total article views: 513 (including HTML, PDF, and XML) Thereof 513 with geography defined and 0 with unknown origin.

Country	#	Views	%

Latest update: 18 Sep 2025

Short summary

We explored a new method to improve Micromagnetic Tomography (MMT), a technique to study the magnetic signal stored in rocks. Precise co-registration of MMT datasets is crucial but challenging, so we tested the use of micro-sized marker structures. These markers help with sample orientation during measurements, speed up data processing, and improve precision. This offers benefits for MMT and other geoscience research requiring precise sample positioning and dataset co-registration.


Total:	0
HTML:	0
PDF:	0
XML:	0