Plasma in-liquid by means of anodic contact glow discharge electrolysis (aCGDE) is a growing research field allowing the selective modification of the electrode and the electrolyte. The aim of this proof of concept study is to demonstrate that auxiliary electrochemical electrodes placed in the vicinity of the plasma electrode can be modified by aCGDE (ignited at the anode by applying a DC voltage between the driving electrodes). Furthermore, we illustrate in how far such auxiliary electrodes can be used as a probe to detect products (in particular, , , and ) formed in the solution by aCGDE via electrochemical techniques. In this work, aCGDE is achieved by applying a voltage of 580 V to a small Pt wire (plasma electrode) versus a large stainless steel counter electrode. An auxiliary Pt electrochemical working electrode, operated in a three electrode configuration, is placed at different distances from the plasma working electrode. Depending on the distance, we find small changes in the working electrode structure. More importantly, we will show that, in principle, the local concentration in the electrolyte can be monitored operando. After aCGDE, the concentration changes with time and depends on the distance from the plasma electrode.
Studying plasma in-liquids has received increasing scientific interest over the last two decades for its possible application in, e.g., wastewater treatment,1–5 nanoparticle formation,2,4,6–13 catalyst material preparation,14,15 plasma electrolytic oxidation,16–18 or plasma electrolytic polishing.19,20 These investigations provide information on the modification of the plasma electrode or changes induced within the electrolyte. Gas-phase plasma can be used in a similar way to modify the plasma electrode or the properties of a substrate placed close to the plasma source. Related to catalytic applications, the gas-phase plasma can also be used, e.g., to change the mechanism of a catalytic reaction on the target (gas-phase plasma catalysis)21–23 or to tailor the structural properties of the target, for example, to form catalyst materials with distinct (electro)catalytic properties.24–27 In contrast, much less is known about the effect of in-liquid plasma on target materials.
A better understanding of the impact of in-liquid plasma on an auxiliary target electrode placed in the same electrolyte in proximity to an in-liquid plasma electrode is of interest for the following applications. First, in-liquid plasma could be used to prepare an electrocatalyst in situ. This approach would have the advantage that the electrodes do not have to be transferred after the plasma treatment from a gas-phase atmosphere into an electrochemical cell, avoiding contamination or uncontrolled restructuring of the electrode. Furthermore, the electrode potential of the auxiliary electrode can be changed during the preparation process, allowing for more specific tailoring of its structural properties. Second, an auxiliary electrode could be used as an electrochemical probe to determine the species formed by the plasma in the electrolyte, to measure the solution pH, or the voltage drop in the electrolyte.
Such fundamental aspects have, to the best of our knowledge, so far not been considered in the ongoing research, but are vital in the search for possible applications. The purpose of this work is to illustrate the influence of plasma in-liquid (ignited at the anode by applying a DC voltage between the driving electrodes) on the current–voltage behavior of an auxiliary electrode in solution placed in proximity of the plasma electrode and whether this auxiliary electrode can be used to determine products formed during the plasma. In general, plasma in-liquid can be generated by different approaches.7,28 Here, we focus on anodic contact glow discharge electrolysis (aCGDE), which is achieved by applying a high voltage between a gas evolving electrode and a significantly larger counter electrode.29–33 Pt is used as the plasma working electrode (PWE), since it was suggested to be stable under these conditions,15,30,31,34–37 which means that neither the surface structure changes significantly nor nanoparticles are extruded from the electrodes as it is observed in other systems.7–9 The auxiliary electrochemical working electrode (AE) is also made of Pt, which is an intensively studied electrochemical electrode material38–43 and therefore an optimal benchmark system. Possible structural changes (if significant) can be inferred from changes in the j–E characteristics of the electrode.38 The key products that form concomitantly during aCGDE are , , and .30,34,44 The electrocatalytic reaction of these species with Pt electrodes in alkaline electrolyte are also described in detail in the literature.45–50 It is also known that the electrode is well suited to detect products, when the electrode is controlled with a potentiostat. A key aspect is the detection of , which is usually determined after the electrolysis by means of titration.30,51–53 This approach does, however, not provide any information on the spatial distribution or temporal evolution of formed during aCGDE. Other methods are able to circumvent this issue to some extent.22,23 In total, knowing about local concentrations is, for example, important to understand possible structural changes of the working electrode induced by after aCGDE.15 In electrocatalysis, Pt is used as collector electrode, which oxidizes selectively, i.e., during the oxygen reduction reaction (ORR).54–56 As a proof of concept, in this work, a Pt wire is placed as probe at different distances from the PWE.
II. RESULTS AND DISCUSSION
The experiments were performed in a large cylindrical glass cell (diameter: 13.5 cm, height: 7.5 cm) containing 0.01 M KOH. The two electrodes for aCGDE (PWE and plasma counter electrode—PCE) and the three electrochemical electrodes (AE, reversible hydrogen electrode—RHE, and electrochemical counter electrode—ECE) are immersed in the cell as shown in the schematic illustration in Fig. 1. A glass tube (outer diameter: 2.6 cm; inner diameter: 2.2 cm; wall thickness: 2 mm) is placed around the plasma counter electrode to remove volatile products created at the PCE, an approach suggested by Hickling and Ingram.30 Even though migration of species into the electrolyte cannot be completely ruled out, most of the products contained in bubbles formed at the counter electrode during the aCGDE are believed to be transported to the electrolyte surface within the perimeter of the glass tube. The plasma was ignited at the anode by applying a DC voltage between the driving electrodes and was apparent by a distinct purple glow, confined in a gas sheath surrounding the Pt wire plasma electrode, as well as a sharp hissing noise. In our studies, the distance between the plasma working and counter electrode was between 1.2 and 1.5 cm, and the AE was placed at a distance of either 1.5 or 4.5 cm from the PWE. The applied voltage between the PWE and the PCE is given in V, while the potential at the AE is given on the RHE scale (). A detailed description of the experimental setup, materials used, and experimental procedures is provided in the Experiment section.
The effect of aCGDE on the AE was studied by comparing the variation of electrode potential () vs time () (Fig. 2, top row) with the current density () vs (Fig. 2, middle and bottom rows). Each experiment consists of a combination of the following experimental steps:
- Step A:
Cyclic voltammetry57 (measuring the current response while applying a triangular wave potential) recorded at 50 mV between 0.0 and 1.05 in the absence of aCGDE, followed by both or either one of the following steps B.
- Step B1:
Keeping the AE at 1.0 (unless otherwise mentioned) during 30 s of aCGDE (PWE anode at 580 V versus a stainless steel cathode, with a power of approximately 57 W).
- Step B2:
Keeping the AE at 1.0 for 15 min in the absence of aCGDE.
- Step C:
Recording cyclic voltammograms (CVs) between 0.00 and 1.05 in the absence of aCGDE.
Each column in Fig. 2 represents one of the experimental steps above. The time axis always counts from the beginning of each of these steps. The variation of the current density for an AE located 1.5 cm from the PWE (close to the PWE) is shown as a red curve and that for a distance of 4.5 cm as a blue curve (far from the PWE). The green curve represents the case where the AE was not exposed to aCGDE (no step B1) and was kept for 15 min at 1.0 (in step B2). The same color scheme is used throughout the article. At 1.0 , can be selectively oxidized at a Pt electrode. Note that in the literature, is usually reported to be detected at 1.2 ,43 however, at that potential, Pt surface oxidation cannot be ruled out. After aCGDE, a waiting period of 15 min was applied to ensure that the species formed by the plasma have sufficient time to distribute evenly in the electrolyte. The waiting time was chosen based on continuous cyclic voltammetry measurements performed in the same voltage range as in step A. After 15 min, the cyclic voltammogram looked identical to that recorded in step A.
Figures 2(e) and 2(i) show the potential-dependent variation of the current density during cyclic voltammetry measurements. By applying a triangular wave potential in this voltage range, usually the current (density) shows electrode specific features. These are related to the capacitive charging of the surface and charge transfer reactions (adsorption/desorption of ions) occurring at the electrode surface. The interpretation of these features is provided in the discussion to Fig. 3. Here, it is important to note that for all experiments, the potential-dependent variation of current density versus time in step A is almost identical [Figs. 2(e) and 2(i)], indicating that the experimental conditions at the start of each experiment are the same. This is even more apparent by comparing the current density versus potential curves in the cyclic voltammograms described in Fig. S1.58
Performing aCGDE while applying a constant potential at an AE placed close to the PWE [step B1: Figs. 2(f) and 2(j), red curves], a large negative current density is recorded at the AE. The current density is several orders of magnitude larger compared to the redox currents obtained without aCGDE in step A. In contrast, much less negative current densities are recorded when the AE is placed far from the PWE (blue curves). Preliminary results indicate that the voltage drop between the PCE and PWE changes the electrolyte potential not only between the plasma electrodes but also further away from these electrodes in the electrochemical cell. Hence, depending on the position of the electrochemical electrodes, the potential at the AE cannot be controlled properly. Since we record a negative current at the AE during aCGDE it is likely that the AE potential shifts in the hydrogen evolution region. In addition, the currents recorded between the PWE and PCE are fluctuating strongly, which induces changes in the electric field and, in turn, induces a strong noise in the current densities recorded at the AE in step B1. Further measurements are required to clarify these assumptions. Note that the effect of electric fields on auxiliary electrodes is not unknown and is, for example, explored in ohmic microscopy59–62 and bipolar electrochemistry.63,64 Overall, to derive a more meaningful conclusion, the effect of voltage drop on auxiliary electrodes will be discussed elsewhere using an electrochemical cell with a more defined geometry, which also allows a more precise positioning of the electrodes.
In step B2, the plasma was turned off and the AE potential was kept for 15 min at 1.0 [Fig. 2(g)]. When the AE is located far away from the PWE, the current density (blue curve) recorded at the AE is almost zero and remains constant for the rest of the waiting period. In the case where the AE is located close to the PWE, the current density (red curve) is high when the plasma is turned off. With time, the current density decreases and remains constant at slightly positive current densities during the entire waiting time. For comparison, keeping an AE at 1.0 for 15 min without exposing it to aCGDE [no step B1—green curve in Fig. 2(g)], the current density is almost zero. This difference in current densities recorded at the AE placed at different distances from the PWE demonstrates that aCGDE certainly has an impact on the current-time profiles recorded at auxiliary electrodes.
After the chronoamperometry steps (measuring the current density at a constant AE potential as a function of time—B1 and B2 or only B2), cyclic voltammetry was performed at the AE in step C [Figs. 2(h) and 2(l)]. If the plasma, waiting time, or species formed by the plasma had no effect on the current density profiles, we would expect the same behavior as in step A. The results, however, clearly show that especially in the first few seconds of the first cycle in step C, the evolution of current densities strongly depends on the waiting time as well as on the position of the AE with respect to the PWE during aCGDE. Possible changes of the pH should not play a significant role in the magnitude of the current density or the features in the CV. According to the Nernst equation this only shifts the features along the potential scale. Changes induced by variations in conductivity within the electrolyte might change the features. In order to draw detailed conclusions, more detailed studies are required, elucidating, e.g., the actual local concentrations.
As described above, the key products that form concomitantly during aCGDE are , , and ,30,34,44 whose electrocatalytic properties with Pt electrodes are well described in the literature.45–50 In order to assess whether or not the features observed in Figs. 2(h) and 2(l) are related to the presence of these species in the solution after aCGDE, we performed a separate set of cyclic voltammetry measurements in electrolytes saturated with either one of the reactants possibly formed during aCGDE or , shown in Fig. 3(a). Even though is not a product during aCGDE, it serves to remove residual from the electrolyte, which, in turn, allows studying surface specific redox processes. The gray curve shows the cyclic voltammogram (CV) of Pt recorded in saturated 0.01 M KOH, the light blue curve in saturated 0.01 M KOH, the pink curve in saturated 0.01 M KOH and the orange curve in 0.01 M KOH containing . Additional measurements with a larger upper potential limit are shown in Fig. S2.58 All CVs were recorded at a scan rate of 50 mV .
The features observed in the CVs recorded in different types of electrolytes are extensively discussed in the literature and can be interpreted as follows. The CV recorded in 0.01 M KOH (gray curve) deaerated with shows a negative current at , which is related to the hydrogen evolution reaction (HER), where water is split to form and ( ). The features between 0.05 and 0.5 are related to the adsorption of hydrogen in the negative going scan and the desorption of adsorbed hydrogen in the positive going scan ( with * being a free surface site). These features are sensitive to changes in surface crystallographic orientation. The features for are related to the adsorption of hydroxyl () or the formation of surface oxygen (O*) from adsorbed OH (OH* + ) in the positive going scan.42,65 The slightly negative current densities for in this gray curve are caused by the reduction of residual in the electrolyte in the so-called oxygen reduction reaction (ORR), where is reduced to water (). Similar CVs to the gray curve are also obtained for all electrodes investigated before aCGDE in step A shown in Fig. S1.58
In -containing 0.01 M KOH (pink curve), the positive current density between 0.05 and 1.0 is attributed to the hydrogen oxidation reaction (HOR) whereby is oxidized to form water (). The oxidation process is inhibited in the region of surface Pt oxide formation at around . In -containing 0.01 M KOH (light blue curve), the negative current density at is related to the ORR. In -containing 0.01 M KOH (orange curve) the negative current density at around is related to the hydrogen peroxide reduction reaction (HPRR), where is reduced to form hydroxid ions (O ) and the positive current density at around is related to the hydrogen peroxide oxidation reaction (HPOR) ( ). In all cases, the adsorption/desorption features observed in the CV recorded in 0.01 M KOH deaerated with (gray curve), overlap with the current density related to the electrocatalytic reactions (HOR, ORR, HPRR, and HPOR). In contrast to literature findings, the here reported current densities of the electrocatalytic reactions are low due to limited diffusion of the reactants to the electrode under stagnant electrolyte conditions.45–50
Figures 3(b) and 3(c) show the CVs from Figs. 2(h) and 2(l) (step C) recorded after aCGDE with and without waiting time, respectively. Similar results are also obtained when a lower voltage is applied between the PWE and the plasma counter electrode (550 V in Fig. S358), as well as with and without shielding the plasma counter electrode by a glass tube (Fig. S4).58 In addition, for comparison, a CV recorded in step A (before the aCGDE treatment) is included in black and a CV recorded after 15 min waiting time without applying aCGDE in green.
Comparing the black and the green curves (before and after holding the potential at 1.0 for 15 min) in Figs. 3(b) and 3(c), a larger reduction current is apparent in the green curve in the first negative going scan at around 0.8 . This feature is related to the degree of electrode oxidation at high potentials. In general, the size of the reduction peak depends on the upper potential limit of the CV as well as on the time and potential at which the electrode is kept (see also Fig. S5).66 Hence, the larger reduction peak in the green curve is caused by holding the potential at 1.05 for 15 min.
Cyclic voltammograms recorded on the AEs placed at different distances from the PWE, after aCGDE and subsequent waiting time at 1.0 (steps B1 and B2) are shown in Fig. 3(b). The CV recorded on an AE close to the PWE (red curve) is almost identical to the green curve (obtained without applying aCGDE). On the other hand, if the AE is placed further away from the PWE (blue curve), the size of the reduction peak in the first negative going scan is larger and shifted to more negative potentials (0.65 ) compared to the red and green curves. This implies that the electrode is significantly more oxidized. As mentioned above, such an effect would be expected for longer waiting times or when the upper potential limit was higher (see Fig. S558 and more details below).
Skipping the waiting time after aCGDE and recording CVs directly thereafter (only step B1) for AEs placed at different distances from the PWE are shown in Fig. 3(c). For AEs close to the PWE (red curve), in the first negative going scan a large positive current density is observed at high potentials (0.8–0.95 ). The current density decreases at more negative potentials until at ca. 0.5 , the curve follows the CV recorded before aCGDE (black curve). Compared to the CVs presented in Fig. 3(a), this large positive current density can at least at high potentials () be attributed to the oxidation of . Note that the HOR and the ORR are almost suppressed at these potentials, as shown in Figs. 3(a) and S2.58 The oxidation current is also measured at higher potentials (> 0.95 ) in the following potential cycles, as shown in Fig. S6.58 The temporal decrease of the positive current density is more clearly seen in the potentiostatic measurements in Fig. 2(g) (red curve). The CVs and the potentiostatic experiments indicate that remains in the region around the AE for a certain time until it dissipates into the electrolyte or reacts with the Pt PWE to decompose into and O.67,68
For AEs located farther from the PWE (blue curve), such a positive current density is not observed in the first negative going or subsequent scans, indicating that is not measurably present in the electrolyte. Similar to the blue curve in Fig. 3(c), obtained directly after aCGDE, a large reduction peak is observed at around 0.65 in the first negative going scan in Fig. 3(b) (steps B1 and B2). This large peak, which is observed in both experiments, suggests that the electrode surface oxidizes significantly during aCGDE. Interestingly this suggests that the plasma or the species formed by the plasma only affect the structural properties of the Pt AE, when the AE is placed far away (4.5 cm) from the PWE. For the time being, it is not clear if the increase in Pt oxide formation is induced (i) by the plasma (or plasma species) or whether it is related (ii) to changes in electrode potential induced by the electric fields during aCGDE, which could cause an uncontrolled oxidation of the electrode during aCGDE.
Overall, the CV measurements performed after aCGDE indicate that the plasma or the species formed by the plasma do not have a significant impact on the structural properties of the Pt AE. The additional current densities in the CVs, with respect to a CV recorded in saturated 0.01 M KOH (see black curves), are, therefore, attributed to the reaction of the Pt AE with the species formed by the plasma in the solution, where the reaction rates depend on the product distribution and its temporal evolution in the electrolyte. When the AE is located close (1.5 cm) to the PWE, significantly different current densities arise at the AE directly after aCGDE, which are tentatively attributed to the reaction of plasma created species (, , ) with the Pt AE. In the CVs recorded right after aCGDE (step B1) and at potentials , a more significant negative current density is observed. Assigning these current densities to the reaction of different species with the Pt AEs is not straightforward, since the total current density is a sum of HOR current density (positive), ORR current density (negative), HPRR current density (negative), and surface redox processes. At high electrode potentials, especially , the situation is more clear. The observed positive current density can basically only be attributed to the oxidation of , with minor contributions from the oxidation of . These positive current densities persist after multiple cycles (see Fig. S6).58 Finally, the different behavior of the versus curve in the first negative going scan of the CV recorded after aCGDE for AEs placed at different distances from the PWE suggests that the product distribution in the solution changes with the distance from the PWE. The change in current density in Fig. 2(g) during the waiting time after aCGDE also implies that the product distribution changes with time.
In this work, we elucidated to what extent auxiliary electrochemical working electrodes can be used to determine by electrochemical methods the impact of aCGDE on (i) the structural properties of an auxiliary electrode and (ii) how far such electrodes can be used as a probe to detect products formed during and after aCGDE. In both cases, it was not yet possible to determine these properties during aCGDE. We suggested that changes in the electrode potential of the electrochemical electrodes might be responsible for this observation. This issue has to be addressed more specifically in a more detailed study where the electrode and electrochemical cell geometries are more defined.
After the aCGDE, we did not observe any significant changes in the Pt auxiliary electrodes. Hence Pt is a suitable probe to detect . Even though the absolute local concentration in the electrolyte is not accessible yet, we were able to measure relative changes of the concentration in the electrolyte which depends on the distance from the plasma electrode as well as on the waiting time after electrolysis. This proof of concept study demonstrates that an electrochemical probe can indeed be used to detect operando and also locally in the electrolyte, which is so far not accessible with other techniques. In order to gain more quantitative information, further studies elucidating local temperature changes or convection in the electrolyte are, however, required. Overall, such information is crucial for further studies aiming at a more detailed understanding of product formation during aCGDE, the composition of the plasma, the structure formation processes at the plasma electrode, or the reaction of these intermediates with species in the electrolyte.
All experiments were performed in a five electrode setup. This consisted of a classical three electrode setup usually used for electrochemical measurements with an auxiliary electrochemical working electrode (AE), a reference electrode (here a reversible hydrogen electrode—RHE), and an electrochemical counter electrode, as well as two electrodes to generate a plasma in-liquid, namely, a plasma working electrode (PWE), and a plasma counter electrode (PCE).
Materials: Pt wires with a diameter of 0.5 mm and a purity of 99.99 % from MaTecK were utilized as AE and PWE. As electrochemical counter electrode, a Pt-sheet (10 7.5 mm) and as plasma counter electrode, a stainless steel plate (20 20 3.5 mm size) were used.
The 0.01 M KOH solution was prepared from KOH pellets (99.99 %, Sigma-Aldrich) and Milli-Q water (18.2 M cm, TOC 3 ppb).
Sample preparation: For the first measurement of each day, the Pt wires were annealed for 3 min in a propane (MTI) flame atmosphere. Since previous studies suggested that the Pt wire electrodes do not change measurably under our experimental conditions,15 the wires were not freshly prepared for the subsequent measurements on the same day.
Electrochemical cell: In order to study the influence of the PWE on the AE, all electrodes were placed in the same electrochemical cell. The latter consists of a large glass beaker (diameter: 13.5 cm, height: 7.5 cm) containing 650 ml of 0.01 M KOH, where a glass tube was placed around the plasma counter electrode (illustrated in Fig. 1). In the case where no glass tube was employed (as in the measurements shown in Fig. S458), the beaker was only filled with 500 ml electrolyte.
The Pt wire used as PWE was protected at both ends with polymer caps (fabricated from Eppendorf pipet tips). The cap around the wire apex is necessary to prevent the wire from melting during aCGDE.69 The caps were adjusted such that 2 mm of the wire are exposed to the electrolyte (surface area of 0.031 ). The PWE and plasma counter electrode were placed between 1.2 and 1.5 cm apart from one another.
The voltage between these electrodes was applied with a TDK-Lambda Power Supply (630 V/1.365 A). The power supply was controlled with a labview program. Unless otherwise mentioned, 580 V were applied for 30 s (plasma power approximately 57 W) between the PWE and plasma counter electrode. At this voltage, the plasma ignites immediately at the PWE, as compared to lower voltages.15 This can be recognized by a fluctuating purple glow within the gas sheath surrounding the PWE, as well as a sharp hissing noise.
The Pt wire used as AE was immersed 10 mm in the electrolyte solution, where the distance was controlled by putting a protective polymer cap (see above) at the upper end of the wire (surface area of 0.159 ). In this case, the apex of the wire does not need special protection, since the current densities are much lower compared to those recorded at the PWE. A homemade and freshly prepared reversible hydrogen electrode (RHE) was used as a reference electrode. The potential at the AE was controlled with a FHI ELAB potentiostat. All potentials are given on the RHE scale ().
Experimental procedure. Since the aCGDE experiments lead to an accumulation of products in the solution, namely, , , and , the electrolyte was changed after each experiment to have the same initial conditions for all experiments. To reduce the possible impact of formed at the plasma counter electrode during aCGDE on the other electrodes (especially, the AE), a solid glass tube (outer diameter: 2.6 cm; inner diameter: 2.2 cm; wall thickness: 2 mm) was placed around the electrode, allowing the majority of the formed , contained in bubbles, to leave the electrolyte within the perimeter of the glass tube.30 The glass tube does not touch the bottom of the glass beaker but is immersed in the electrolyte slightly deeper than the plasma counter electrode.
All measurements were carried out at room temperature. The temperature of the electrolyte before and after the experiment was very similar in each case, with ca. 24 C measured with a thermometer immersed in the solution. Note that after aCGDE, the temperature in the electrolyte is rather inhomogeneous, with higher electrolyte temperatures around the PWE. Since the formation of during aCGDE is inevitable, the electrolyte was not deaerated before the experiments. All experiments were performed without stirring the electrolyte.
For the electrochemical characterization, CVs were recorded before and after aCGDE. The initial potential for all measurements (unless otherwise mentioned) was set around the open circuit potential (ca. ), the lower potential limit to 0.00 and the upper potential limit to 1.05 . For some measurements, the upper potential limit was increased to 1.25 . When the upper potential limit was increased to 1.25 , the initial potential was set to 1.20 . The scan rate for all measurements was 50 mV , and the initial scan rate was always negative (negative potential scan). During the aCGDE in step B1 and in step B2, the potential of the AE was held at the initial potential (1.00 ).
The authors gratefully acknowledge support by the DFG (German Science Foundation) through the collaborative research center SFB-1316 (Project ID: 327886311) as well as the state of Baden-Württemberg and the DFG through Grant No. INST 40/574-1 FUGG. E.A. would like to thank the “Stiftung Stipendien-Fonds des Verbandes der Chemischen Industrie (VCI)” for the financial support in form of a scholarship.
Conflict of Interest
The authors have no conflicts to disclose.
Evelyn Artmann: Conceptualization (equal); Formal analysis (equal); Funding acquisition (equal); Investigation (equal); Validation (equal); Visualization (equal); Writing – original draft (equal); Writing – review & editing (equal). Lukas Forschner: Writing – review & editing (equal). Timo Jacob: Funding acquisition (equal); Resources (equal); Supervision (equal); Writing – review & editing (equal). Albert K. Engstfeld: Conceptualization (equal); Funding acquisition (equal); Supervision (equal); Writing – review & editing (equal).