Microgravity-induced wet chemical etching of borosilicate glass enhances the process rate

Abstract

In the paper, microgravity simulators—Random Positioning Machine (RPM) and Rotary Wall Vessel (RWV) – were used to investigate the potential influence of quasi zero-gravity force onto the wet chemical etching of borosilicate glass substrates. For the first time, straight microchannels of different widths were etched utilizing RPM and RWV in a solution of hydrofluoric acid (HF) with additives (40% HF: 65% HNO3, 10:1, v/v). Substantial differences compared to the reference etching could be observed and the highest etching rate could be obtained for RPM (3.06 μm/min), next reference with a stirrer (2.01 μm/min) and RWV simulator (1.71 μm/min). For both RPM and RWV, also the potential aging of the solution, defined as the decrease of etching rates in a function of etched samples, was found to be the smallest (circa 10–11%). Apart from the kinetics of the etching process, the quality of the etched microchannels was also a subject of this research. As a result, good quality surfaces of the samples were observed for RPM, RWV and a reference with stirrer. When it comes to the RPM, the microfabricated structures were the deepest, but simultaneously, characterized by uniform edges and devoid of any accidental imperfections. Research on RWV, in turn, showed that the structures obtained utilizing this microgravity simulator have the smallest etching edge and have a potential to be reproduced the most precisely. Results presented in this paper can be a significant base for the further development in the field of microgravity-induced chemical research and reaction kinetics with microgravity simulators.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-025-20579-5.

Introduction

There has been a growing interest in space sciences, driven by the opportunity to study how microgravity and radiation affect life, processes, and various phenomena. This research has accelerated the development of new methods and tools that can simulate space environments more closely. The field of study covers a wide range of topics, from cutting-edge medical issues like cell culturing and tissue engineering^1–5 to more practical, everyday concerns such as the potential for manufacturing in space^6,7.

Before manufacturing in space can be developed, there are many technical challenges to overcome⁷. These issues are directly related to the properties of materials and the kinetics of chemical reactions in a microgravity environment. For example, a metal 3D printer developed for the International Space Station (ISS) had to use a wire-based printing technology instead of a powder-based system. This modification was necessary to manage the microgravity conditions^6.7.

Despite the challenges, microgravity can actually be beneficial for certain experiments. A promising area of research is protein crystallization^8–10. In a microgravity environment, protein molecules can form crystalline structures more slowly and in a more orderly fashion because there are no gravity-driven forces like sedimentation and convection. This allows scientists to grow larger, higher-quality crystals, which can help in the development of new biotherapeutic drugs. Researchers have also recently studied the physicochemical properties of colloids in space, finding that the absence of downward diffusion in microgravity reduces the rate of colloidal crystallization¹¹. Experiments mentioned herein suggest that other chemical reactions may also behave differently in microgravity. One area of particular interest is material processing, such as wet chemical etching, which is a common technique for micromachining glass and silicon for microelectronics and lab-on-a-chip devices.

Both NASA and the European Space Agency (ESA) have shown a serious interest in microscale techniques, as seen in projects like “Tissue Chips”^12–15. Beyond the potential for in-space fabrication, the transfer of these new methodologies to Earth-based applications could be a major breakthrough, leading to significant process improvements here on Earth.

For creating microstructures in glass substrates for Micro-Electro-Mechanical Systems (MEMS), Micro-Opto-Electro-Mechanical Systems (MOEMS), and lab-on-a-chip (LOC) devices, the standard method is wet chemical isotropic etching using hydrogen fluoride (HF) solutions^16–18.

Typically, a highly concentrated HF solution (40–49%) is used for the etching process. Sometimes, hydrochloric acid (HCl) is added to the HF to improve the final surface quality. While various types of glass are used in microengineering, borosilicate glass (such as BOROFLOAT^® 33 or Pyrex^®) is the most common choice for fabricating microsensors and microactuators^19.20.

The etch rate of borosilicate glass varies significantly, ranging from 1 to 10 μm/min depending on the glass type and method used²¹. While heating the solution or annealing the glass can increase the etch rate, these methods are often not recommended. Heating the solution poses safety risks due to HF vapors, and annealing can result in a rough surface^22.23.

The main challenge in wet chemical etching is the masking process. Depending on the desired structure and required etching time, a mask can be made from photoresists, metals (like Cr/Au or Cr/Cu), or silicon^23–25. If the etching time is too long, photoresist masks can develop significant defects. A popular recent solution is using adhesive polymer masks created through a process called xurography^26–28. After applying the mask, the substrate is heated to about 180 °C to improve adhesion and prevent delamination.

Given that glass microfabrication is fundamental to developing microelectronic structures, investigating this process in the context of space environments is a compelling area of research. This study explores, for the first time, how the wet chemical etching of borosilicate glass is affected by simulated weightlessness.

The questions guiding this research are twofold: first, whether it is possible to fabricate microstructures in space using standard micromachining techniques; and second, whether a microgravity environment can enhance the etching process in Earth-based laboratories. Since there is currently no literature on the effects of microgravity on wet chemical etching—and existing research on etching rates shows significant discrepancies^16,18—this work provides a preliminary insight into the subject.

Materials and methods

LOC substrate—design, technology and assembly

In the studies, borosilicate glass substrates with dimensions of 76 × 26 × 1.1 mm were used (BOROFLOAT^® 33 Schott, Mainz, Germany). As the research is fundamental, the typical microfluidic patterns, i.e. microchannels, were prepared for the microgravity-induced microfabrication (Fig. 1).

Fig. 1.

Mask designed for wet chemical etching of glass substrates with straight microchannels of different widths (1.0, 1.5 and 2.0 mm).

As shown in the Fig. 1, different microchannels widths were proposed within a single substrate to assess the potential impact of the microgravity-influenced etching process. The straight microchannels were placed in three locations—on the right, in the center, and on the left of the substrate surface to enable an analysis of the potential etching uniformity.

Xurography-based masking of the substrates was employed to obtain the microfluidic patterns. Utilizing AutoCad software and CNC plotter, the designs of the microchannels were cut in a thin polymer adhesive foil (Avery Dennison Graphics Solutions, Mentor, OH, USA) and placed on the substrates. Next, wet chemical etching in a solution of hydrofluoric acid (HF) with additives (40% HF: 65% HNO₃, 10:1, v/v) was conducted to obtain spatial microfluidic structures.

As the process of wet chemical etching in HF solutions requires special safety rules, a dedicated mounting system was developed to allow appropriate experimentation with microgravity simulators. At first, a special stand ensuring etching of two glass substrates simultaneously was done utilizing a 3D printing technique. The stand consisted of two similar parts – a base and a cap which could provide quick installation and removal of the substrates utilizing standard laboratory tweezers, as well as stability during the microgravity-induced etching process. The stand (Fig. 2) was 3D printed out of HF-resistive resin material (FusionGRAY, ASIGA, Australia) with the use of Digital Light Processing printer (DLP) model ASIGA MAX X35. During the etching process, the stand was placed into internal Polytetrafluoroethylene (PTFE) container, holding the HF etching solution and the substrates to be etched (Fig. 2). Moreover, an external container was used to act as a protective barrier, safeguarding against potential leaks. Similarly, an absorbent material positioned between the external and internal containers was the additional protection with simultaneous stabilization functions.

Fig. 2.

Stand designed for the etching on RPM and RWV: (a) a cap/base model, (b) whole stand structure with glass substrates attached at a glance, (c) 3D printed stand placed into the internal PTFE container prior to the etching process.

Each of the wet chemical etching conducted in the conditions of simulated microgravity was assisted by a substantial reference. The results were compared with both etching process implemented in a standard Petri dish, as well as with additional variant utilizing electromagnetic stirrer. Based on the Authors experience with glass micromachining, it was assumed to conduct the experiment in a 2-hours period. In order to provide significant statistical analysis, there were 60 microchannels etched by the selected method.

Microgravity simulators—working principle and experimental assumptions

Space-based experimentation is very limited. For this reason, diverse methods have been developed to imitate the weightlessness state. Drop towers and parabolic flights can be considered as inapplicable due to their time limitation, thus rotating systems, e.g. Random Positioning Machine (RPM) and the Rotary Wall Vessel (RWV) are becoming the first choice to obtain quasi zero-gravity conditions.

In both RPM and RWV systems, the operating principle relies on averaging the gravitational vector to near zero^29–31. The RPM consists of two frames: an outer and an inner frame, which are connected and independently driven by two motors (Fig. 3). The sample is mounted within the inner frame. From the point of view of the sample, the gravity vector starts to change direction in all three axes when both frames are set in rotational motion. The resultant force of gravity does not change and is constantly equal to 1 g. However, average force of gravity is being reduced to zero over time.

Fig. 3.

RPM used for the simulated microgravity-induced etching: (a) scheme with indication of angular speed observed for frame 1 and frame 2, (b) our self-developed RPM instrument at a glance.

The RWV operates on the same principle as RPM, but it has only one axis of rotation. In this case, the rotation of the vessel also contributes to averaging the gravitational force, but the effect is limited to two axes instead of 3-D volume (Fig. 4). The axis of rotation has to be perfectly perpendicular to gravity vector, in order to avoid offset in averaged gravitational force (when the axis is parallel to gravity vector, the RWV does not work).

Fig. 4.

RWV used for the simulated microgravity-induced etching: (a) scheme with indication of angular speed, (b) our self-developed RWV instrument at a glance.

For the purpose of the experiments, our own laboratory versions of the simulators were used. The specifications of each instrument are presented in the Table 1. The most importantly, the rotation speed of both RPM and RVW can be adjusted in the range of 0–10 rotations per minute. Based on the literature reports concerning the applications of the simulators for the biomedical and biochemical research, it was decided to set the rotational speed to 10 RPM for both of the instruments^3–32. Lower values could be insufficient to effectively reduce gravitational influence, whereas higher values could result in excessive centrifugal forces, preventing an accurate simulation of microgravity.

Table 1.

Basic parameters of the microgravity simulators—RPM and RWV.

Parameter	RPM	RWV
External dimensions	259 × 260 × 283 mm	120 × 120 × 152 mm
Internal frame dimensions	100 × 260 × 200 mm	ɸ 120 mm
Speed adjustment range	0 ÷ 10 RPM	0 ÷ 10 RPM
Maximum rotor load	4 kg	0.2 kg
Maximum moment of imbalance	4 kG × cm	not specified
Power supply	230 V~50 Hz < 15 VA	230 V~50 Hz < 6 VA

Calibration graphs of RPM and RWV are provided as a supplementary material (S1, S2).

Evaluation of the etched structures—measurements, optical and scanning electron microscope (SEM) analysis

A micrometer screw gauge (No. 2109 F, Mitutoyo, Japan) was used to check the rate of the microgravity-induced chemical etching of the glass and indicate the depth of the etched microchannels. Each channel was examined and measurements were taken at multiple locations to confirm the repeatability of the results. Etching speed was calculated by dividing the obtained depth by the process time, assuming a linear progression of the etching process.

To assess the quality of the etched structures, both an optical microscope and a scanning electron microscope (SEM) were used. The optical microscope equipped with a 6.3 MP camera (DLT-Cam PRO, DELTA Optical, Polska) allowed for quick and efficient examination of basic visual features, such as edge uniformity, surface defects, and the precision of the designed geometries utilizing DLTCamViewer software. SEM (JSM IT-100, Jeol, Japan) enabled a detailed analysis, at the micro- and nanostructural levels, as well as the examination of edge profiles, and the presence of potential imperfections utilizing InTouchScope software.

Results

As mentioned earlier, wet chemical etching of the glass substrates was performed in the conditions imitating microgravity utilizing RPM and RWV instruments (Fig. 5). The reference etching was provided in a typical Petri dish, as well as with electromagnetic stirrer. Ten glass substrates were prepared for each of the etching process. Every substrate was designed to contain six microchannels (Fig. 1). It was assumed to etch in a 2-hours period to appropriately evaluate the etch rate and notice any significant structural changes.

Fig. 5.

Wet chemical etching of the glass substrates on RPM: (a) set-up, (b) example etched microchannels.

Based on the structures measurements with micrometer screw gauge, etching rate was indicated for all the sixty microchannels (six per glass substrate) and presented as a mean value (Fig. 6).

Fig. 6.

Graph representing etching rates of the samples depending on the etching method. Etching parameters – etching solution: 40% HF: 65% HNO3, 10:1, v/v, RT, rotation speed for RPM and RWV: 10 RPM, for stirrer—1000 RPM. (Sample number corresponds to the borosilicate glass substrate number. Graph presents the mean value of the etching rate with a view to six microchannels localized on the glass substrate).

Next, mean value, standard deviation, maximum and minimum value, as well as evaluation of the percentage difference between maximum and minimum value of the etching rate were calculated for the samples (Table 2; Fig. 7). The last of the calculation was done to assess if there is any substantial change in the etching effectiveness. As the same etching solution was used to etch all the samples, maximum etching rates were obtained for the first etching of the samples, minimum, in turn, for the last etched samples. On that basis, it was possible to observe potential decrease of the etching rate, thus aging of the solution.

Table 2.

Wet chemical etching of glass substrates in RPM and RWV compared with reference.

	RPM etching rate [µm/min]	RWV etching rate [µm/min]	Reference etching rate (no stirrer) [µm/min]	Reference etching rate (with stirrer) [µm/min]
Mean value	2.85	1.54	1.26	1.69
Standard deviation	0.13	0.14	0.09	0.21
Maximum value	3.06	1.71	1.39	2.01
Minimum value	2.74	1.52	1.18	1.52
Min/max × 100%	89.54	88.89	84.89	75.62

Experiment parameters—etching solution: 40% HF: 65% HNO3, 10:1, v/v, RT, rotation speed for RPM and RWV: 10 RPM, for stirrer—1000 RPM (based on the average etching rate calculated per microchannel and per borosilicate glass sample).

Fig. 7.

Graph representing etching rates of the samples depending on the etching method: (a) mean values, (b) maximum values. Etching parameters—etching solution: 40% HF: 65% HNO3, 10:1, v/v, RT, rotation speed for RPM and RWV: 10 RPM, for stirrer—1000 RPM. (Based on the average etching rate calculated per microchannel and per borosilicate glass sample).

With a view to the obtained results, it can be observed that the highest etching rates, for all the samples, could be achieved for the process performed utilizing RPM. The mean value of the etching rate on RPM (2.85 μm/min) was much greater than the second result represented by the reference etching with magnetic stirrer (by nearly 40%). Etching rate of RWV was slightly lower than that obtained for the reference with stirrer (by circa 9%), but higher than the reference without the stirrer (by circa 18%). In the case of the maximum etching rates observed for all the methods, the statistics looks similar, therefore the maximum value could be obtained for the RPM (3.06 μm/min), next for the reference with stirrer (2.01 μm/min), RWV (1.71 μm/min), and reference without the stirrer (1.39 μm/min).

In order to conduct a visual inspection of the isotropy of the microchannels, SEM imaging was performed. The images were acquired at the channel terminations to enable the evaluation of regions incorporating right-angled geometries, which serve as critical sites for assessing structural uniformity and potential fabrication-induced deviations (Fig. 8).

Fig. 8.

SEM images of the etched microchannels with the use of: (a) RPM, (b) RWV, (c) reference with stirrer, (d) reference without the stirrer—top view.

Assessing isotropy and wall quality further allows the detection of defects such as rounded corners, surface roughness, or asymmetries, which can adversely affect mass transport, heat transfer, or biochemical interactions occurring within the system. Therefore, it is also important to carefully examine the edge profiles, as their sharpness and definition provide additional insight into fabrication accuracy and potential performance limitations (Fig. 9). For this reason, visual inspection constitutes a key step in the characterization and validation of fabrication processes in microfluidic technologies.

Fig. 9.

SEM images of the etched microchannels with the use of: (a) RPM, (b) RWV, (c) reference with stirrer. On the left—microchannel view, on the right—magnified edge profile.

Discussion

Presented herein results may seem interesting, especially in the context of RPM applicability for the wet chemical etching or other similar chemical processes at which reaction kinetics is an important factor. Probably 3-axis rotation imitating weightlessness state could have favoured dissolving of the glass surface. Similar etching rates obtained for RWV and reference with stirrer seems also consistent with the scientific expectations, since for both of the methods, a single-axis rotation was provided, with the only difference related to the plane. Thus, the physical phenomena revealing during the process enhanced the etching rate compared with reference without the stirrer.

As shown in the Fig. 6, with each subsequent sample the averaged microchannel etching rate decreases, regardless of the method. Nevertheless, the smallest difference between the etching rates obtained for the first and the last sample can be observed for RPM and RWV (decrease by circa 10%-11%). Probably, microgravity simulation provided by constant 1-axis or 3-axis rotation could have reduced the effects of the solution aging onto the etching rate or simply, prevented from rapid deterioration of the solution quality. Perhaps, lack of sedimentation inhibited uncontrolled separation of the solution, allowing for its uniform composition in the etching area³³.

With regard to the depth of the etched structures, some differences based on the process nature could have also been observed. Certainly, the deepest microchannels were etched for the RPM (maximum value – 367 μm), next, reference with stirrer (maximum value – 241 μm), RWV (maximum value – 205 μm) and reference without stirrer (maximum value – 167 μm). Moreover, in all the samples, tiny differences in the structures depth relating to the microchannels location (Fig. 1) could be indicated. Percentage relation had shown that in all the cases, the differences, regardless of the location (on the right, in the center, on the left of the substrate) were less than a 1%. However, for all the samples, the greatest values of the etched microchannels depths were measured for the external locations (on the right and on the left of the substrate). These indications may seem random but can also be connected with the mixing of the solution. For the reference etching with a stirrer, the nature of the mixing process is based on three phenomena – distribution, dispersion, as well as diffusion within a solution³⁴. Presence of the stirrer provides turbulence and kinetic energy of the vortices depends on the RPM value. As the magnetic stirrer is placed in the center of the Petri dish, vortex which is formed performs turbulent motion in its outer area, which explains the deepest microchannels obtained in some distance from the substrate center. Moreover, based on dispersion, vortices break further into smaller ones, enhancing the etching process at the external parts. In the case of microgravity simulators (RPM, RWV), thanks to appropriate preparation of the etching stands (minimization of air bubbles, protective sponges), the microgravity simulation could be imitated and moving of entire solution was inhibited. Nevertheless, the presence of the vortices at the molecular level cannot be neglected and microscale mixing based on diffusion should be considered in this regard. According to the literature^35.36, the smallest possible vortex size can be described by the Kolmogorov scale. The size of the vortices depends strongly on the properties of the fluid, including kinematic viscosity. In liquids with a viscosity similar to that of water, i.e., as in the case of our etching solution, the Kolmogorov scale (thus vortex size), is equal to 30–100 μm. Based on diffusion mechanism, the center of the vortex is described by a laminar flow, however random turbulences of the microscale vortices (30–100 μm) occurring at the external parts of the substrates cannot be entirely excluded. Thus, the deepest microchannels observed herein at the external parts of the glass substrates can be justified, too.

When it comes to the microscopic inspections of the microchannels, isotropy of the etching can be visible for all the samples. However, based on our knowledge and experiences, the “prettiest” etched surfaces might be observed for both RPM and RWV microgravity simulators, as well as reference with stirrer (Fig. 8). The worst, in turn, was visualized for the reference without stirrer (Fig. 8). Certainly, the size of etched edge for RPM is the biggest, but it is strictly connected with the microchannel depth – 342 μm (mean value), which is rather a deep structure. Nevertheless, the surface view, as well as homogeneity of the edge seems worth of notice. Interestingly looks also the microchannels structure obtained on the RWV. As shown in the SEM image, the sharpest edge of the etched microchannels can be observed herein, which can be important in the case of providing cavities of a precise geometry, even square-based. When it comes to the reference etching, especially without the stirrer, the surfaces look as expected and some accidentally etched parts can be seen. In all the structures (with a special focus put on RPM and RWV), no considerable surface defects and unusual imperfections could be observed. Rather smooth and hydrophilic character of the surface was noted (Fig. 9).

Conclusion

In this paper, simulated microgravity environments were used to investigate the potential influence on the wet chemical etching of borosilicate glass structures. Two simulators were applied for this purpose – RPM and RWV, in assistance with reference etching in a Petri dish with and without the electromagnetic stirrer. The studies were focused on a microfabrication of typical microfluidic structures – straight microchannels, as it was fundamental research. Xurography technique, as well as a solution of hydrofluoric acid (HF) with additives (40% HF: 65% HNO₃, 10:1, v/v) were used to implement the etching process. In order to ensure the studies on microgravity simulators that provide the rotation of the desired speed, the dedicated stand and safety containers were prepared.

The major aim of this research was to verify, if the microgravity environments can have an impact onto the etching process, especially in the context of etching rate and quality of the etched structures. Sixty microchannels have been etched in the selected processes and it has been shown that the highest etching rate (3.06 μm/min) can be obtained for the RPM, next reference with a stirrer (2.01 μm/min) and RWV simulator (1.71 μm/min). Moreover, interesting observation has been done towards potential aging of the solution. As the same etching solution was used to etch all the borosilicate glass samples, maximum etching rates were obtained for the first etching of the samples, minimum, in turn, for the last etched samples, regardless of the etching method. Nevertheless, it was RPM and RWV at which the difference between the etching rates was the smallest (circa 10–11%) that may suggest the positive influence of the simulators on the solution quality, probably relating to the lack of sedimentation and uniform composition of the solution in the etching area, as a result of inhibited gravity force. The research on microchannels structures depth relating to their location on a glass substrate was also conducted. However, substantial differences were not observed in this regard. The deepest microchannels observed at the external substrate parts (not in its center) can be connected with mixing phenomena. For the reference etching with stirrer, the dispersion, distribution, as well as diffusion of the solution could be responsible for this effect, whereas, in the case of RPM and RWV, solely the potential presence of the molecular level vortices, defined by a Kolmogorov scale and ranging from 30 to 100 μm could be considered.

Microscopic observations of the samples have shown that good quality surfaces of the samples can be observed for RPM, RWV and reference with stirrer. Due to the highest etching rate of the RPM, the microchannels etched based on this method were the deepest (circa 342 μm) but simultaneously, the edge of the structures was uniform, without any accidental imperfections. Moreover, in the case of RWV, the sharpest edge of the structures could be observed, differing from the reference with stirrer significantly. It may seem that the most precise, even square-based geometry could be obtained utilizing this method which can be essential in the case of microfluidic structures.

Based on the results obtained in this paper, it seems that further research in this area, especially in the context of microgravity-induced kinetics of chemical reactions, can be valuable. Equally important could also be the studies on silicon microengineering, or other materials that exhibit high anisotropy during etching. In summary, as shown in this paper, microgravity simulators may play an important role in the microfluidic structures fabrication, especially utilizing the process of deep wet chemical etching.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

A.K. proposed the idea of the research and wrote the main manuscript text. K.B. conducted experiments and prepared SEM images. M.B. developed RWV and RPM instruments. All authors reviewed the manuscript.

Data availability

The Authors declare that the data supporting the findings of this study are available within the paper. Should any raw data files be needed in another format they are available from the corresponding author upon reasonable request.

Declarations

Competing interests

The authors declare no competing interests.