Exoland Simulator, a Laboratory Device for Reflectance Spectral Analyses of Planetary Soil Analogs: Design and Simulation

Abstract

In planetary science, visible (Vis) and near-infrared (NIR) reflectance spectra allow deciphering the chemical/mineralogical composition of celestial bodies’ surfaces by comparison between remotely acquired data and laboratory references. This paper presents the design of an automated test rig named Exoland Simulator equipped with two reflectance spectrometers covering the 0.38–2.2 µm range. It is designed to collect data of natural/synthetic rocks and minerals prepared in the laboratory that simulate the composition of planetary surfaces. The structure of the test rig is conceived as a Cartesian robot to automatize the acquisition. The test rig is also tested by simulating some project trajectories, and results are presented in terms of its ability to reproduce the programmed trajectories. Furthermore, preliminary spectral data are shown to demonstrate how the soil analogs’ spectra could allow an accurate remote identification of materials, enabling the creation of libraries to study the effect of multiple chemical–physical component variations on individual spectral bands. Despite the primary scope of Exoland, it can be advantageously used also for tribological purposes, to correlate the wear behavior of soils and materials with their composition by also analyzing the wear scars.

1. Introduction

The paper presents the design, set up, and simulation of a test rig called Exoland Simulator with two reflectance spectrometers mounted on it and devoted to the spectral analysis of surfaces, as soil analogs. In wider terms, a test rig is a customized research and testing device created for a particular purpose, often in a laboratory setting. Very often, a test rig is used for simulations that allow the validation of theoretical models and scientific observations [1,2,3]. It serves as a hardware prototype designed to test specific functionalities of a new proof-of-concept or technology, with the primary focus on gathering reliable data under controlled and ideal conditions. A test rig can have applications in different fields, which can range from automotive testing [4,5] and aerospace [6], to electronics and material science [7,8], from educational fields [9] to medical purposes [10,11,12].

The primary aim of the proposed laboratory device is the analysis of planetary soil analogs (PSA). The structure of the Exoland Simulator is based on a Cartesian robot derived from the modification of a Fused Deposition Modeling (FDM) 3D printer, generally used for producing laboratory prototypes as in [13,14].

A fundamental approach for the exploration of our Solar System (including Earth) is represented by remote sensing. Remote sensing in the field of spectroscopy is based on the collection of electromagnetic radiations that are emitted or reflected by a body and relies on the fact that, at different wavelengths of the electromagnetic spectrum, each object reflects or emits a specific intensity of light, which depends on the physical or compositional properties of the target [15].

The remote sensing techniques performed through telescopic, space, and ground-based missions have provided much of our knowledge about the Solar System, and in particular of the composition and structure of solid planetary surfaces, since this technique allows analyzing a planetary object in a geo-scientific context [16,17].

Among the plethora of different remote sensing techniques, the one exploiting the visible and near-infrared (Vis–NIR) range of the electromagnetic spectrum is widely used for investigating planetary atmospheres and to identify surface compounds [16,18,19,20].

Considering the power of Vis–NIR reflectance spectroscopy to identify and map surfaces from a distance, several spectrometers have been carried onboard spacecrafts, e.g., the Compact Reconnaissance Imaging Spectrometer for Mars (CRISM) on the Mars Reconnaissance Orbiter (MRO) spacecraft. This is a visible–infrared (Vis–IR) hyperspectral imager that aims at characterizing the geology and mineralogy of thousands of sites at high spatial and spectral resolution [21]. The Mars Multispectral Imager for Subsurface Studies (Ma_MISS) on the ESA Rosalind Franklin rover is a miniaturized Vis–NIR spectrometer aimed at acquiring spectral data on the Martian subsurface [22], and the SuperCam coupled with Vis–NIR spectroscopy on the Mars 2020 Perseverance rover payload has the scope of identifying primary and secondary minerals (e.g., quartz vs. amorphous silica) [23]. Therefore, several planetary laboratories were built and equipped with spectral test rigs to collect data and create a large spectral database of rocks and minerals to interpret and transfer data from spacecrafts [24,25,26].

Many works have been carried out using Vis–NIR spectroscopy to collect and characterize different compounds for later comparison with remote data. Sulfates and carbonate minerals were analyzed since they are important for understanding the geological and climatic evolution of several planetary bodies [27,28]. Numerous studies have been performed on sulfur- and ammonium-bearing minerals in relation to possible micro-organism habitability in these environments [29,30,31]. Laboratory spectra of clays/phyllosilicates, altered basalts, mafic minerals (e.g., pyroxene), and glass materials were largely studied as important Mars-analog materials [32,33,34]. The interest in the presence of water above or below the crust of planetary bodies, in different states of aggregation (solid or liquid), has motivated the study of hydrated minerals and their evolution at different pressures and temperatures to assess the possible storage or release of liquid water [35].

In some cases, the evolution of mineral spectra at high and low temperatures and the effect of grain size were analyzed [32,36]. The spectra of hypothetical mineralogical mixtures are then interpolated with remote sensing data to decipher the composition of planetary soils. However, the spectrum of a planetary surface (like the Earth’s) is the result of the sum of several mineralogical/amorphous phases with different physical and textural properties. Therefore, a much more complex spectrum than that of a single phase or component is obtained.

In this perspective, the conceptualization and construction of the Exoland device aims at implementing spectral libraries of rocks and minerals with a series of reflectance spectra in the 0.38–2.2 µm range with a different and progressive degree of complexity, from a single mineral composition to a series of different rock composition mixes, considering different physical parameters such as grain size and/or temperature. This approach will make it possible to assess the effect of variations of multiple components on individual spectral bands for a better and more accurate remote identification of the selected materials and, finally, to identify which bands and spectral ranges are the most diagnostic and influenced by different physical–chemical properties. A prior work in the scientific literature reports the conceptualization and study of an automated test rig for spectral analyses of soil analogs [37]. In this prior paper, a laboratory goniometer system for the assessment of the anisotropic reflectance and emittance behavior of soils was presented. This system is based on a commercially available robotic arm to collect measurements over the full hemisphere, performing multi-angular measurements under controlled illumination. The arm has a spectrometer and a thermal camera mounted on it, and these allow the acquisition of either hemispherical measurements or measurements in the horizontal plane. Another goniometer system was presented in [38]. Different from the prior work, the test rig proposed in the present paper exploits a different geometrical measuring approach by using a Cartesian robot and four moving axes: three sliding ones and a rotating one. In addition, a heating system is included to change the temperature of the samples on the worktable.

Vis–NIR spectroscopy being a rapid, user-friendly, and non-destructive technique, it is used also in many industrial fields and applied sciences such as tribology [39,40,41], product and quality control [42,43], the food industry [44,45], the study of fertilizers [46], chemistry [47], and the pharmaceutical industry [48]. Regarding the use of visible and/or infrared spectroscopy in tribology, the main applications relate to the study and tribological characterization of lubricants and oils [39,40] or thin films [41] to the optimization of the lubrication and vibration behavior in tribological joints as bearings [49]. This technique is also useful for studying the wear scars on the mating surfaces of a tribological pair to analyze the wear behavior, the wear mechanism, and the presence of particles or impurities at the interface. This can be performed both for industrial tribological pairs and for human joints or prostheses, such as knee or hip joints [50]. Another application can deal with the study of the compounds on worn surfaces that cannot be attributed directly to the wear progress, but derive from different phenomena. From this perspective, the automated simulator proposed in this paper can be advantageously used for tribological purposes both on tribological pairs and on different materials as liquid or solid lubricants and ground soil samples.

Therefore, the aim of this work is to present and discuss the design, set up, and motion simulation of a Vis–NIR spectral test rig, based on a Cartesian robot, to analyze planetary soil analogs. The test rig is designed to collect the data of natural/synthetic rocks and minerals prepared in the laboratory that simulate the composition of planetary surfaces. For this reason, the device is called the Exoland Simulator. The same device can be also advantageously used for tribology purposes to analyze worn surfaces or friction interfaces. The ability of the test rig to trace some designed project trajectories is also investigated by simulating specific working conditions and presenting some results. Furthermore, some preliminary spectral results are presented to emphasize the future research directions or possible applications.

2. Materials and Methods

In this section, the main components and methods to design and set up the new proposed automated test rig are presented and described together with the proposed acquisition method and design of the trajectories. Furthermore, a brief insight about Vis–NIR spectroscopy is introduced.

The proposed test rig for spectral acquisitions in the Vis–NIR range is conceived with the following main elements:

• An acquisition system (spectrometers and optical system);
• A positioning system (Cartesian robot);
• A handling system.

The acquisition system consists of two reflectance spectrometers and an optical system consisting of a fiber and a collimator.

The positioning system is represented by a Cartesian robot, which is obtained through the modification of a 3D printer. This system is able to automatize the movement of an end-effector, which can be located at any point of the workspace of the Cartesian robot. The end-effector must be equipped with the collimator, and this can be achieved by designing a proper handling system. This handling system is aimed at working as a case for the collimator, allowing its movement while it is connected to the spectrometers and their objectives through the fiber and allowing its rotation with respect to the worktable. The sliding movement of the collimator must cover the 3 directions of the workspace, performing acquisitions on planes parallel to the worktable at different heights. To automatize the movement of the optical system mounted on the end-effector of the Cartesian robot, an electronic control board was properly programmed.

2.1. Vis–NIR Reflectance Spectroscopy and Acquisition System

Spectroscopy studies the interaction between matter and electromagnetic radiation. Reflectance is the potency of a material to reflect electromagnetic radiation (EMR). The union of these features gives rise to four fundamental vibrational modes: symmetric stretch ($\nu 1$), symmetric bend ($\nu 2$), asymmetric stretch ($\nu 3$), and asymmetric bend ($\nu 4$). Bending vibrations include molecular scissoring, rocking, wagging, and twisting. The Vis–NIR range was chosen to realize the acquisition system of the Exoland Simulator as it provides the identification and information of many minerals and molecules [51], such as the following:

• Molecular water (adsorbed, interlayer, water ice) and hydrated salts (e.g., sulfates) due to overtone and combinations of H₂O/O–H groups in the 1.4–1.6 µm range [28];
• Ortho- and chain silicates through Fe²⁺ crystal field transitions that produce large absorption bands around 1.0 and 2.0 µm (pyroxene or olivine) [52];
• Sheet silicates (phyllosilicates) due to the hydroxyl vibrational mode OH at 1.4 µm combined with the Al–OH or Fe/Mg–OH modes at ≃2.2 and ≃2.3 µm, respectively [33];
• Carbonates through overtone and the combination of the ${\mathrm{CO}}_3$ groups $3\nu 3$ at 2.3 µm [28];
• ${\mathrm{NH}}_4^{+}$ compounds due to the presence of overtone and the combination at ≃1.08, 1.56, 2.01, and 2.2 µm [53].

As the acquisition system must work in an overall range of wavelengths equal to 0.38–2.2 µm, two different spectrometers were selected during the design phase: an SR2 model (0.38–1.08 µm), shown in Figure 1a, and an NIRQuest+ model (0.9–2.2 µm), shown in Figure 1b (both Ocean Insight Inc., Orlando, FL, USA). The collimator is a Gershun Tube, as shown in Figure 2, and it is connected to the two spectrometers by an optical fiber.

The collimator is able to collect light from an area of about 10 cm of diameter and allows changing the angle of view between 1 and 28 degrees. The optical system must be moved with respect to a worktable, where the soil samples will be placed, while the two spectrometers will be placed in a steady position on the table, where the test rig is positioned.

2.2. Design of the Positioning System: Modification of a 3D Printer

The positioning system is designed to move the optical system in a three-dimensional workspace, a task that can be performed by a Cartesian robot (Figure 3). For convenience, to set a proper Cartesian positioning system, a suitable 3D printer can be used and properly modified for the specific task. In fact, a 3D printer is a kind of Cartesian robot used in the manufacturing field. It has only prismatic joints for the main movements, and therefore, it is able to move the end-effector only along straight lines on the three axes. Therefore, a 3D printer represents an adequate solution to build a specific Cartesian robot for the automated spectral experimental acquisitions proposed in this paper. Furthermore, a 3D printer normally has a planar worktable for the arrangement of the samples and an end-effector normally used for the extrusion of filaments, which can be modified for mounting the handling system and the collimator to receive reflected data from the desired point in the workspace.

The low cost, easy availability, and adaptability allowed identifying a Fused Deposition Modeling (FDM) 3D printer as the Cartesian robot to be modified and used as the positioning system of the newly proposed test rig. In particular, the chosen 3D printer was an Artillery Sidwinder X1 (Artillery, Hong Kong, China), which is a medium-sized printer with 400 mm of stroke along the z-axis and a worktable of 300 mm × 300 mm. This means that the available workspace for the automated spectral acquisitions was 300 × 300 × 400 mm³.

The axes move thanks to rubber wheels running on V-slot tracks. The structure has a main base case where the control boards and cables are laid out. As can be observed in Figure 3, the worktable can move along the y-axis through proper guides, and the vertical beams act as supports for the movements along the z-axis of the horizontal rods, which, in turn, work as the support for the slider with the end-effector mounted, which slides along the x-axis. For motions along z, there are two motors and a belt transmission with pulleys that connect the two upper ends of the threaded rods. Therefore, the considered Cartesian robot has three native sliding axes, but in this work, it was modified by adding a fourth axis, which allows the rotation of the handling system, mounted on the end-effector, to obtain spectral acquisitions at different incidence angles, as will be disclosed in the next paragraph. The stroke-end switches are inductive. In addition, the worktable is equipped with a heating system, and this is a very useful function that could allow studying the variation of the spectral data with temperature.

2.3. Design of the Collimator Handling System

To integrate the optical system into the end-effector, a proper handling system was designed to support the collimator and to allow its movements, considering the mounting dimensions. In addition, the handling system was also conceived to give the collimator the possibility of rotating to obtain acquisitions at different angles. This rotation, managed by a stepper motor, allows evaluating the influence of the angle between the collimator axis and the normal to the worktable, where the samples are positioned. For this purpose, the handling system was designed as an assembly of a handling support for the collimator and of a support for the motor and the control board, as reported in Figure 4.

Figure 4a shows the proposed handling system for the collimator, and Figure 4b reports the designed support for the fourth motor and the control board.

The final proposed test rig is shown in Figure 5.

2.4. Design of Trajectories

To allow geo-acquisitions or tribology measurements, the test rig proposed and described here must be properly programmed to perform acquisitions following specific trajectories properly chosen for the specific application. These acquisition paths can be defined and transferred to the Cartesian robot through a G-code program. Therefore, the imposed trajectories are implemented point-by-point by programming the position of the Cartesian robot’s end-effector in the workspace.

The acquisition paths can be designed as planar trajectories in the plane of the worktable, as Figure 6 shows, or can be set as three-dimensional curves. In the case of the soil analogs’ geo-acquisitions, the most suitable trajectories to allow spectral analyses are planar ones, which are repeated at different heights along the z-axis.

The positioning system is, thus, governed by a matrix containing the coordinates of the end-effector’s positions including the stop times for each desired acquisition. The desired paths can be defined at different heights on the z-axis for comparing the planar acquisitions at different distances from the samples.

The project trajectories of the end-effector in the workspace are defined also over time: as an example, Figure 7 shows the x position of the end-effector over time. The program for the positioning of the handling system can be properly modified at each acquisition, based on the defined best acquisition paths.

The prismatic joints of the robot that allow the translation in the x and z directions are consecutive, but the joint for the translation in the y direction moves separately with respect to the others. To solve this problem, the governance of the robot was split into two parts according to the block diagram reported in Figure 8.

To simulate the motion of the Exoland Simulator and its ability to follow given paths, a specific project trajectory was considered with the main characteristics reported in

Table 1. Characteristics of the project trajectory.

Features	Data
Number of positions in x	3
Number of positions in y	2
Number of total positions	6
Acquisition height	10/40/90 mm
Collimator admittance angle	0°
Stop time in each position	0.2 s
Total simulation time	118.08 s

. With this kind of trajectory, the system was properly programmed and the end-effector moved along this path, registering the differences between the assigned and executed trajectory along the three axes, over time.

3. Simulations on the Positioning System and Results

In this section, the results of the tests and simulations conducted on the Cartesian robot are reported with the aim of investigating the ability of the positioning system to reproduce the project trajectory. This trajectory was implemented by using the Matlab code and was transformed into a G-code by means of an algorithm properly developed for this purpose. The G-code program allows controlling both the translation along the three axes and the rotation of the collimator to allow multi-angle acquisitions.

Therefore, the control board was programmed to execute the implemented G-code to trace the desired trajectory.

The rotation of the end-effector, simulating the movement of the collimator around the proper axis designed in the handling support, is controlled by the robot board and actuated by a stepper motor that allows changing and setting the inclination angle.

To simulate the phases of the spectral data acquisition during the movement of the collimator along the project trajectory, a controlled LED is positioned on the end-effector, which turns on in correspondence to the time periods when the spectrometers are programmed to acquire images.

Simulations on the positioning system were performed by executing the programmed code moving the Cartesian robot along the project trajectory and registering the differences between the input trajectory and the real path followed by the end-effector.

The results of the simulations are reported in Figure 9, where the simulated trajectories are represented and superimposed on the project ones. The results are presented in terms of the variation of the three coordinates x, y, and z over time.

From these results, one can observe that the simulated and project trajectories show minimal differences and are superimposable.

This means that the test rig is able to reproduce the defined and programmed acquisition trajectories, and therefore, it results in being optimized and conformal to perform Vis–NIR spectral analyses. With these results, the positioning system of the test rig can be considered qualified.

4. Showcasing Spectral Analyses on Planetary Soils

This section showcases some of the spectral analyses that could be performed with the proposed, developed, and tested test rig. In fact, the Exoland Simulator, given the presence of the two spectrometers (Ocean Insight SR-2VIS400-25 (Ocean Optics, Inc., Orlando, FL, USA) 0.38–1.07 µm and NIRQuest+ 2.2 µm), will be able to collect reflectance spectra similar to those shown in Figure 10, Figure 11 and Figure 12. These three figures showcase three of the possible different measurements that could be performed with the test rig in order to illustrate its possibilities and suggestions for applying this instrument.

The reflectance spectra are characterized by absorption bands whose position and related parameters (area, depth, FWHM) are diagnostic of the type of excited interatomic bond and, thus, of the type and abundance of the mineral and/or rock. In addition, the different slopes of the spectra are diagnostic for glass species, as shown in Figure 10. In detail, the spectra reported in Figure 11 and Figure 12 regard natural and synthetic ammonium-bearing minerals, respectively. Ammonium-bearing minerals’ spectra are characterized by the presence of spectral features due to overtones and a combination of ${\mathrm{NH}}_4^{+}$ modes located at ≃1.06, ≃1.3, ≃1.56, ≃2.02, and ≃2.2 µm, which are useful to discriminate these salts by remote sensing data. These spectral features are sensitive to the degrees in symmetry of the ${\mathrm{NH}}_4^{+}$ group inside the crystal structure [55], the strength of hydrogen bonds [27,29], and the distance between donor–acceptor atoms in the structure [56]. Concerning the glass samples (Figure 10), they are essentially featureless spectra, but characterized by different slopes, albedos, and spectral ratios. These parameters are sensitive to the iron and composite Silicium-Calcium-iron-Magnesium contents (SCFM parameter) [32].

In addition, from the spectral analysis, the spectral ranges most affected by different physicochemical properties will be identified.

These collected data have a wide range of applications in planetology from rocky to icy planetary bodies for interpreting their surface composition. Furthermore, other spectral data collected on thin films, lubricants, and friction interfaces can be advantageously used to study the tribological behavior of surfaces, the wear mechanisms, and the contamination of surfaces.

5. Conclusions

The paper presents a test rig able to perform automatic spectral acquisitions of soils. The presence of different geological materials on the test rig worktable allows studying specific planetary soil analogs. This is the reason why the test rig is called the Exoland Simulator, which is conceived as a Cartesian robot obtained from the modification of a FDM 3D printer, adding a properly designed handling support for the optical system on the end-effector. In addition, the end-effector is designed to rotate, allowing acquisitions with different angles, and the worktable of the test rig is equipped with a heating system to study the variation of the spectral data with temperature. The two spectrometers used, covering a range of 0.38–2.2 µm, are connected by an optical fiber to the collimator mounted on the robot’s end-effector, and the system can be programmed by specific G-codes, which implement planar project acquisition trajectories at different heights on the z-axis. The imposed trajectories are defined to perform experiments for spectral geo-acquisitions. The design and setup of the test rig were described together with the testing approach that was established through the definition of the acquisition methods and trajectories. The results of the motion simulation of the test rig are reported in terms of the ability of the robotic positioning system to reproduce the defined trajectories.

Therefore, this paper presented the study and development of a system devoted to the automated spectral characterization of rock/soil analogs. Despite this being the primary function of the proposed automated test rig, it can be advantageously used for tribology. For instance, it will be useful in the tribological characterization of lubricants, oils, coatings, and thin films or to study the wear behavior and mechanisms with a focus on the presence of particles or impurities at the interface and on the composition of the wear scars. Future research directions will be focused on performing spectral acquisitions directly with the test rig. This will be useful to implement spectral libraries of minerals with a series of reflectance spectra, allowing assessing the effect of variations of multiple components on individual spectral bands for a better and more accurate remote identification of materials.