Sputtered highly effective iridium catalysts: a new approach for green satellite propulsion

This work demonstrated the large potential of sputtered iridium metal for catalytic reactions shown by the example of decomposition of hydrogen peroxide (H2O2) for space propulsion systems. For this purpose, iridium was coated onto Al2O3 pellets by a sputter process under varied process parameters. Depending on previously selected parameters, the obtained metal-loaded pellets offer closed- and/or open-shell structures. Catalytic productivity of these first-generation iridium-sputtered catalysts was estimated in laboratory experiments and compared to platinum-loaded pellets. Under optimized sputter-process conditions, the reactivity is significantly improved compared to the platinum-impregnated pellets. The better catalytic productivity can be explained by the increased active surface area of the iridium layers on the pellets. The surface morphology and the microstructure of the iridium coating can be actively controlled by the sputter pressure. The results are in accordance with the sputtering process pressure tendency described by the Thornton Structure–Zone Model.


Introduction Application of heterogeneous catalysis of green propellants in space technology
The need for propellants with reduced toxicity and aggressiveness combined with lower required safety precautions during transportation, storage and handling is a main driver in today's propellant technology: Efforts are being made in the EU which will probably restrict the use of hydrazines (e.g. N 2 H 4 ) and dinitrogen tetroxide (NTO) in propulsion systems for space vehicles in the near future [1]. These toxic substances are still widely used in mono-or bipropellant propulsion systems and have been flight-proven over decades [2]. A green propellant substitute for hydrazine in monopropellant-based propulsion systems is rocket-grade hydrogen peroxide [3]. Although hydrogen peroxide has been used since the early days of rocketry and aviation, industrial production has made progress and opened the market to commercially available hydrogen peroxide of high purity and in sufficient quantities [4].

Noble metals as active catalyst in space applications
Hydrogen peroxide decomposes in an exothermic reaction to water steam and oxygen gas by help of catalytically active materials. Upon contact of the catalyst with aqueous solutions of hydrogen peroxide, the decomposition reaction will start and will be proportionately accelerated by rising reaction temperatures. Therefore, high-concentration solutions of hydrogen peroxide can be used as a mono-or bipropellant in propulsion systems [5]. Typical materials that are used as active catalyst phase are members of the platinum group (platinum, iridium), which are mainly chemically dispersed on ceramic catalyst beads (e.g., aluminum oxides) [6]. Besides the efficiency of combustion of those catalysts at even ambient temperatures, most of the state-of-the-art catalysts are used in a pre-heated mode. Noble metals of the platinum group have as common features suitable resistance to oxidation, sufficient thermal stability, and a high affinity toward catalytic peroxide decomposition [7]. These features make them valuable key players in catalytic-driven monopropellant propulsion systems, working in pulse trains or in continuous-firing mode [8]. Compared to bipropellant combinations and hydrazines as monopropellant, the calculated adiabatic decomposition temperature for highly concentrated hydrogen peroxide is reduced (* 950°C) and will therefore limit the risk of high-rate metal oxidation or ablation [9]. In the following experimental concepts and considerations, this value is also used as benchmark.
Although manganese oxide catalysts have been used in the early days of rocket propulsion for the decomposition of hydrogen peroxide [10], we focus within this paper on noble metals as more ''modern'' catalysts, especially on iridium and platinum.

Sputtered films as catalyst
Besides the above-mentioned methods to use platinum or iridium, many applications in the form of thin films for catalytic applications are known [11][12][13]. The main reason for this usage is the costsaving compared to bulk material. One method to deposit thin catalytic films is a physical method such as PVD (physical vapor deposition, sputtering) [11,[14][15][16]. However, the sputtering method to deposit platinum or iridium for catalytic applications is currently rarely used; most of the publications describe the usage of sputtered platinum and/or iridium as catalysts in proton exchange membrane fuel cells. As early as 1979, Beni et al. [17] demonstrated that electrocatalysts for oxygen evolution electrodes can be produced by sputtered Ir-O films. Mukerjee et al. [18] studied in 1993 the effect of sputtered iridium on the kinetics of oxygen reduction. Atanasoski et al. [19] compared different rare earth materials sputtered on Pt-coated nanostructured thin films. Alvisi et al. [20] could achieve significantly higher activity for MeOH oxidation with sputter-deposited platinum films compared to conventional catalysts. The goal of Hirano et al. [21] was to minimize the Pt-loading in catalysts cells. Ganske et al. [22] and Tapalov [23] investigated the catalytic activity of different Ir/Pt ratios. In two different publications, Slavcheva et al. described the influence of sputter parameters on the surface structure and on the catalytic effects of sputtered platinum [24] and iridium [25]. A systematic variation of process parameters such as power, pressure, and target-tosubstrate distance is described in publications by Wessling et al. [26][27][28], including a simulation of the sputter process in one publication [29].
Most of the current publications concerning sputtered iridium catalysts are dealing with the usage of these layers for Proton Exchange Membranes (PEM), water splitting, Oxygen Evolution Reaction (OER) and Hydrogen Evolution Reactions (HER) [30].
In this study, we experimentally investigated sputtered iridium films for the decomposition of aqueous hydrogen peroxide solutions for later application in propulsion systems. The porosity of the deposited iridium layers should be tuneable by adjusting the process gas pressure during the sputtering process. Transmission electron microscopy (TEM), scanning electron microscopy (SEM) and energy-dispersive X-ray spectroscopy (EDX) measurements were performed to examine morphology, microstructure, thickness and elemental composition of the obtained iridium layers. Following the sputtering process, an experimental decomposition test and thermal stability measurements of the prepared iridium pellets are of high interest. Oxygen-rich environments and increasing temperatures that are achievable in a model combustor chamber are of a challenging in nature concerning material oxidation or ablative effects.
First simulations of these environmental conditions can roughly be performed in thermogravimetric (TG) and dynamic differential calorimetry (DSC) apparatus under oxygen-rich gas flow and high temperatures. Within this experimental approach, worst-case conditions (1000°C, oxygen-rich atmosphere) were chosen. Experimental outcomes give a first impression at material stability and thermal properties of the produced pellets.

Experimental Preparation of pellets and sputtering process
The iridium thin films were prepared by radio frequency (RF) magnetron sputtering using equipment from Aurion Anlagentechnik with a high-purity (99.9%) iridium target. The distance between the target and the substrate was set to 120 mm, and the inclination angle between the target and the normal to the substrate was set to 40°and a sputtering power of 300 W was used. As process gas, high purity argon (99.999%) was used. The process pressure during coating was actively controlled by a throttle valve, in front of the pump combination of rotary pump and turbo molecular pump, independently from the argon flow. After loading the chamber with the substrates (Al 2 O 3 pellets or flat glasses), the system was pumped down to 5.0Á10 -5 hPa to remove the remaining gases. The sputter processes started after the argon gas flow was set to 50 sccm and the chosen process pressure was stabilized. Two different batches (named A and B in the following text) were coated using different process pressures (p A-= 3Á10 -2 hPa and p B = 5.0Á10 -3 hPa). To achieve better uniformity of the films, the substrates were rotated at 8 rpm during sputtering. Neither heating nor cooling was applied to the substrates. The pellets were placed on a structured substrate holder (metal lattice), as shown in Fig. 1.
The coatings were performed in two steps to ensure a coating of the complete pellet surface. After the first step, the coating chamber was flooded with air, opened, and the pellets were turned upsidedown by hand. Afterward, the second process step was started, again with pumping down and using the same process conditions. To estimate the coating thickness, some flat glass substrates were coated additionally, and on these glass substrates the film thickness was measured using a Dektak XT profilometer from Bruker, equipped with a stylus of radius 12.5 lm. The reproducibility of the layer thickness on flat substrates is (for identical sputter process conditions) better than ± 2%. However, the thickness distribution on the Al 2 O 3 pellets is strongly depending on the individual surface morphology of the fissured pellets. Additional uncertainties of the layer thickness occur due to the described turning of the pellets to ensure the complete coating of the pellets. Both batches are treated in the same way.
To achieve comparability (i.e. surface morphology, dimensions) in the decomposition experiments, iridium was coated on the same type of pellets (diameter ca. 2 mm, length 5 mm) which were used as reference platinum pellets. For this purpose, industrially manufactured Al 2 O 3 pellets loaded with 1% of platinum were used (formerly Degussa, F 26 P). Blank pellets for the sputtering process were obtained the following way: Pt-loaded pellets from Degussa were treated twice with boiling aqua regia and washed with purified water to remove all platinum before sputtering. The success of complete removal of platinum was proven with EDX analysis at different positions on the pellets. This procedure removes loose Al 2 O 3 particles in/on the pellets, too.

SEM and TEM analysis
Cross-sections of iridium films were investigated using an FEI TECNAI T20 transmission electron microscope (TEM) equipped with a LaB 6 filament. An electron-transparent lamella of the iridium coated sample (flat glass and pellet) perpendicular to the film plane was prepared by focused ion beam (FIB) using a Helios NanoLab 600i, followed by argon ion etching with a PIPS II polishing system. The surfaces of the iridium-coated pellets were analyzed using a Phenom ProX Desktop Scanning Electron Microscope including energy-dispersive X-ray spectroscopy (EDX) to determine the elemental composition.

Thermal analysis
Thermogravimetric (TG) and dynamic differential calorimetry (DSC) measurements were performed with a NETZSCH Jupiter STA 449F in oxygen-rich gas flow (50 mL min -1 ), in Al 2 O 3 crucibles. In a typical experiment, Ir-loaded pellets were heated at a rate of 10 K/min to the final temperature. The measuring curves were corrected by subtracting a previous measured buoyancy curve which eliminated lifting effects caused by the gas flow at raising temperatures. Evolved gaseous species were detected by Fourier-transform infrared spectroscopy (FTIR spectroscopy) with a coupled BRUKER ALPHA FTIR spectrometer.

Catalytic effects
Perhydrol (solution of 30 wt % H 2 O 2 in water) was used in the experiments as model liquid in a gas collecting apparatus built in our laboratory. Concentration of the aqueous H 2 O 2 solution was validated before initialization of the test sequence by density measurement. The reaction chamber was a doublewalled glass cylinder with water as thermostating liquid occupying the annular space between the double walls. The water constantly circulated through a thermostat at ambient conditions (22°C). A thermometer displayed the temperature of the peroxide solution. The apparatus was sealed by a glass flange, providing gas-tight conditions. Collection and readout of the volume of the evolved gas were done visually by a glass syringe (100 mL, Fortuna Optima) where the syringe barrel and plunger were sealed with PTFE (see Fig. 2).
The whole apparatus was washed three times with water and ethanol separately to remove any traces of impurities and flushed to dryness with nitrogen afterward. Prior to the start of each experiment, gas tightness of the apparatus was proven by applying a slight positive pressure of nitrogen gas and checked for potential gas emission using leak detection spray (OKS 2801). Experiments were performed successively with 10 pellets of each batch and Pt-reference material in 60 mL of freshly bottled perhydrol. All pellet charges were used sequentially according to Table 1 and tested within one day. Each setup was cleaned, dried and checked before a new pellet charge was investigated.
The liquid was then allowed to equilibrate to thermostat temperature in the darkened reaction vessel. Time was recorded at first contact of pellets with the liquid.

Experimental investigations of the catalytic productivity of Ir-coated pellets
Results of the experimental investigations of the catalytic productivity of the two different processed pellet batches compared to Pt-loaded samples are summarized in Table 1. In general, the amount of released oxygen gas was measured at three different points in time (excluding the starting point at 0 s), and perhydrol temperature was recorded at the end of each measuring procedure. Measurement errors are principally based on reading errors and scaling effects of the syringe volume during data acquisition. Here, measured values of higher volumes are less faulty than those obtained in lower regimes. It is therefore assumed that the obtained values show an inaccuracy of approximately less than ± 1 mL.
During the experiment, photographs of the reaction during perhydrol decomposition were taken. In Fig. 3, an example image of the ongoing reaction (left) with used pellets (right) are presented.
For better comparison, Fig. 4 shows closeup photographs of nine pellets before (Fig. 4a) and after hydrogen peroxide treatment (Fig. 4b). Compared to the freshly prepared material, no significant optical changes in the overall appearance of the pellets are detectable. Moreover, no solid remnants were found at the bottom of the glass vessel, indicating ongoing ablation effects during reaction progress under the given conditions. In Fig. 5, obtained data points are plotted in a turnover vs time curve and displayed by a logarithmic curve. Under the given experimental conditions, gas release capability of the platinum pellets was equal to that of pellets of batch B. The curve that reveals much higher gas volumes was obtained with the batch A pellets. Increasing liquid temperature and gas release indicates a spontaneous, self-accelerating exothermic reaction. This finding is in accordance with the assumption of an increased active metal surface area on the pellets, originating from different process pressures during sputtering. On glass substrates, the coating time of 1080 s leads to nearly equal layer thicknesses of approximately 190 ± 10 nm at both used sputtering pressures. The corresponding relative mass of the iridium layer has been calculated as about 0.5 wt % for a typical cylindrical pellet, considering an iridium density of 20 g/cm 3 . Therefore, an effect of the layer thickness for the observed increase of the catalytic    efficiency can be ruled out. The cause lies rather in some aspect of the iridium layer that differs as a function of the different used process pressures.

Thermal analysis of sputtered Ir-coated pellets under oxygen-rich conditions
In Fig. 6, thermogravimetric (TG, black line) and dynamic differential calorimetry (DSC) curves (dotted line) of uncoated and Ir-sputtered Al 2 O 3 pellets are presented (from top to bottom). While the TG/ DSC curves of an uncoated pellet remain unremarkable, curves of the coated samples reveal significant changes.
In general, the recorded TG curve of an Ir-sputtered sample reveals a stepwise mass decrease during heating to 1000°C final temperature. In sum, three weakly defined steps are detectable without reaching a final plateau at even higher temperatures (\ 1000°C). It is therefore assumed that, at first stage (* 100°C), mass loss is dominated by the evaporation of gases or adsorbed water which is near to the surface of the pellet.
This assumption is also validated by recorded FTIR spectra of the evolved gas species. Here, only carbon dioxide and gaseous water are detectable as gaseous species. The medium temperature range at roughly 100°C to 500°C is dominated by a weight loss of yet unknown origin.
At temperature regimes above 500°C, the observed mass loss could be explainable by slow metal oxidation into iridium(IV) oxide followed by evaporation. The observable exothermic DSC peak may originate from this oxidation process. Moreover, it is known in the literature that the iridium(III) oxide phase is maintained by chemical vapor transport of iridium(IV) oxide in oxygen-rich atmospheres at temperatures above 1000°C [31]. A similar process could take place here. This assumption is underlined by the fact that discoloration of the pellet after the heating process is observable. Moreover, the nanosized nature of the iridium crystallites could lead to changes of material properties. Similar effects on material conditions (e.g., change of boiling or melting point) that are caused by nanosized materials are extensively described in the scientific literature [32,33]. In further experiments, an Ir-coated pellet was exposed to harsh conditions in a muffle furnace: For this purpose, the sample was heated to 1000°C in air and then maintained at this temperature for 1 h. Results are presented in Fig. 7, clearly showing the complete removal of the deposited iridium layer.

Discussion
The fact that the catalytic efficiency of sputtered iridium layers depends on the sputter parameters, such as pressure, substrate temperature, composition, bias voltage, is already known. For example, Ganske et al. [22] and Topalov et al. [23] described the catalytic effect of sputtered platinum-iridium layers for hydrogen fuel cells as dependent on the Pt-Ir ratio and the sputtering power. Wessling [28,29] evaluated and simulated the electrochemical performance of sputtered iridium as dependent on the sputter pressure. The very strong dependence of the catalytic efficiency on the working gas sputter pressure seen in our study can be explained by the different layer structures for different sputter pressures: Fig. 8 shows transmission electron microscopy (TEM) cross-sections of sputtered iridium layers for the two used different sputter pressures. The iridium layers for these TEM investigations had been deposited on glass substrates with identical process parameters as Figure 6 Thermogravimetric (TG) and dynamic differential calorimetry (DSC) curves of an uncoated (above) and Ir-coated pellet (bottom) under oxygen-rich conditions. for the Al 2 O 3 pellets (except coating time), because the sample preparation for the TEM images with glass substrates is much easier and the results are more concise compared to TEM images of layers deposited on the brittle Al 2 O 3 pellets, as described later.
At low argon pressures (p B = 5Á10 -3 hPa; Fig. 8b), the iridium layers had grown in columnar fashion, and clear grain boundaries can be identified. The layer surface is very smooth. Small irregularities of the substrate are covered complete by the Iridium layer.
With increasing argon pressure (p A = 3Á10 -2 hPa; Fig. 8a), the layer structure appears fissured. Columns are separated by voids; a cauliflower-like structure appears. Also, the iridium crystals are smaller. Diffraction pattern reveals that the crystals have all different crystallographic orientations [34]. Defects of the glass surface lead to an additionally increased disoriented and rugged layer structure, with a significantly increased iridium surface area as shown in Fig. 8a, too. This influence of the process pressures on the layer morphology is mainly caused by the interaction between argon and iridium atoms. A higher sputter pressure leads to an increase of the collisions of the iridium atoms with the argon process gas atoms before reaching the substrate. This results in lower incident energies and a loss of the directionality of the iridium atoms arriving at the substrate or coating layer surface. At lower sputter pressures, the iridium atoms reaching the substrate or layer surface have a higher kinetic energy, which leads to higher surface mobility of the atoms [28,29,34]. Therefore, these atoms build more dense layers with an even surface area, following the concept described by the Thornton-Structure-Zone Model [35,36].   The described TEM analysis of iridium layers sputtered on glass is confirmed by similar analysis of iridium layers on Al 2 O 3 pellets, as shown in Fig. 9.
The sputtered layers show the same growth mechanism as on a glass substrate. In addition, the structure of the Al 2 O 3 increases the surface roughness and surface area of the iridium coating. Figure 10 shows a scanning electron microscope (SEM) picture of an uncoated Al 2 O 3 pellet at low resolution.
The Al 2 O 3 pellets have intentionally a very rough and inhomogeneous structure. Some areas are rougher than other areas. Therefore each pellet is different. However, this diversity should have no effect on the observed increased catalytic effect because the same structured pellets (loaded with platinum) had been used for the reference measurements. Additionally, an averaging takes place over the 10 pellets used in each batch.
The rugged structure of the pellet will additionally increase the disturbance of the iridium growth and may lead to an even larger iridium surface area than for coatings on glass surfaces, which may even increase the catalytic reactivity.

Conclusion and outline
Al 2 O 3 pellets were coated with iridium by a sputtering process. It was shown that the surface morphology is highly tuneable by variation of process parameters, leading to more porous and/or more closed metal structures. Surface morphology was characterized by SEM and TEM images whereas TG and DSC measurements were performed to study the thermal properties of the materials. In first laboratory experiments, Ir-coated pellets showed a high tendency toward catalytic decomposition of aqueous hydrogen peroxide solutions. Here, a clear correlation between gas release and sputtering parameters is recognizable. As a reference, platinum-loaded pellets of the same substrate type were used, although the mass of platinum was about twice the mass of iridium for these pellets. Experiments verified that sputtered Ir-coated pellets with an open porous structure have superior catalytic productivity among the examined samples. TG and DSC analysis revealed that the sputtered iridium metal tends to evaporate at dry and harsh conditions after a period of time.
Therefore, similar effects are expected in rocket motors but, depending on the experimental setup and conditions, much less marked. The results obtained so far show the highly beneficial approach of new catalyst design using sputtering technology also for other catalytic applications. Within this new approach, hot gas tests with rocket-grade hydrogen peroxide ([ 87%wt. H 2 O 2 ) and extended lab-tests (e.g. lifetime of the coated Ir-pellets) are currently in preparation.

Funding
Open Access funding enabled and organized by Projekt DEAL.

Compliance with ethical standards
Conflict of interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. The authors declare, that conflicts of interest do not exist.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licen ses/by/4.0/.