Pioneering the Approach to Understand a Trash-to-Gas Experiment in a Microgravity Environment

The general block diagram of the suborbital OSCAR system is displayed in Figure 2(A). Steam was generated from a syringe pump that pumped a controlled flow of water into a pressurized and heated line at 150°C to generate steam and mix with O₂ gas. The feedstock O₂ gas was supplied from a 6,205 kPa (900 psig) tank, with a downstream regulator set to 589 kPa (100 psig). Using an orifice (located after the O₂ regulator) and a downstream back pressure regulator (BPR) (located post reactor), the intended O₂ flow was designed to supply 1 SLPM to the system and steam at 0.1 g/min. In this work the intended “design point O₂ flow” is referenced, which was this original desired O₂ flow rate tested on the ground, prior to flight. The steam and O₂ lines were mixed so that both flow into one tube and then split into two separate heated lines and each line is connected to the inlet of a different port of the reactor. Three cartridge heaters (Heaters 1, 2, and 3) located in the reactor were set to full power (225 W) during the initial ignition event of the flight and kept on throughout the active experiment. The preheat temperature created ideal ignition conditions for the trash. This was a different design than the prior μg tests at the GRC Drop Tower and ZGF, which had more time and power to preheat the reactor core prior to μg events. The reactor held a constant pressure of 308 kPa (45 psig). Approximately 4.5 g of trash were loaded into the base of the manifold during payload assembly, and 1.1 g of trash were pre-loaded into each of two injection tubes to supply two different trash injections during the μg phases of flight. The total trash mass made available for the suborbital flight reaction was 6.7 g.

Figure 2

The waste material was composed of simulated crewed mission waste (waste mass composition: 17% clothing (100% cotton t-shirt), 13% fecal simulant, 13% food packaging (white with aluminum layer), 13% clear food packaging, 11% hygiene wipes, 11% food simulant, 9% cotton washcloth, 4% tech wipes, 4% toilet paper, and 2% toothpaste and shampoo). The waste composition was named OSCAR waste simulant (OWS) and was different from the composition used during any prior testing in ground or μg campaigns. The main difference between OWS and prior Trash-to-Gas waste simulant was the absence of urine in OWS. Supplemental moisture from the urine would quench the reaction, and there was not enough time in the short μg suborbital flight to overcome or compensate the thermal losses necessary to have the reactor reach any type of a “steady state” condition. Ignition temperatures (Medina et al., 2020) and moisture content (Meier et al., 2020) of these waste simulant materials are reported in previous publications. The reactor was designed to stay below the vaporization temperature of aluminum, which would be recoverable theoretically for reuse as a feedstock elsewhere.

Helium was originally used for the intended design point upstream trash injection gas so as not to interfere with the quantification of O₂ via gas chromatography (GC), but helium conflicted with GC detection of H₂, so only qualitative data was collected for H₂.

Product gases advanced through an outlet at the top of the reactor and then through a heat sink that also acted as a condenser. The condenser was filled with desiccant and molecular sieve material to trap any liquid or moisture. The condenser was also packed with glass wool at the outlet, to filter any solid particulates that were not captured by the stainless-steel mesh in the reactor. The cooled and desiccated product gases were then routed into several gas collection tanks (T1/T2/T3/S), which opened sequentially based on the phase of the flight experiment. The sequential capture in different tanks allowed for a comparison between products formed during different reaction and gravity conditions. The gases were analyzed from the tanks on the ground, after payload return.

The OSCAR payload layout on the FSE is displayed in Figure 2(B). The reactor and gas feed were located on the top shelf, whereas the cooling, filtration, and gas collection were in the mid-region of the payload, and the data acquisition, avionics, power, and telemetry were all located on the lower shelf of the payload.

A sketch of the reactor manifold with the reactant inputs and products output is represented in Figure 3(A). The reactor internal core diameter and height is 91 mm and 74 mm (3.59 in. and 2.91 in.), respectively, with an approximate volume of 0.5 L. The trash injection tubes are each an internal diameter and length of 13 mm and 279 mm (0.5 in. and 11 in.), respectively, with an approximate volume of 0.04 L. Trash was injected below the O₂/steam inlet so that the combustion, gasification, and reforming processes would take place as the gas traveled upward to the reactor outlet (center tube). The reactor walls were the coldest points of the reactor and therefore pyrolysis was expected to dominate in these regions. Experimental complexities included reaction time, based on trash in the base of the reactor as well as two separate trash injections, and O₂ flow termination after a timed duration. The experimental parameters presented the opportunity to observe ignition, combustion, and smoldering effects in μg, but also presented a challenge to model the exact reaction equation. The idealized thermal zones are displayed in Figure 3(B), and the reactor computer aided design (CAD) model with a cut-away view is displayed in Figure 3(C).

Figure 3

The NS-12 and OSCAR flight events are compared in Figure 4. The suborbital OSCAR experiment was designed to synchronize a reactor priming sequence with a series of flight events to prime the reactor, execute waste injection, and convert the waste from solid to gas during the μg phase of the flight. The OSCAR payload data was collected with National Instruments LabVIEW data acquisition software, in combination with the Blue Origin Integrated Payload Controller (IPC). The data acquisition system used flight and experimental trigger events to provide a fully automated waste conversion experiment to occur during flight. The initial flight event signals received from the IPC were very important to the automation and timed operation of the OSCAR experiment flow.

Figure 4

After landing, the NS payload capsule was returned to the West Texas Launch Site Vehicle Processing Facility, where the OSCAR payload was disassembled in sections. The prioritized OSCAR gas sample bottles and data acquisition system data drives were received approximately 4.5 h after T-0 launch to the Payload Processing Facility (PPF), and the rest of the OSCAR payload was returned to the PPF approximately 24 h post landing. After all gas samples were analyzed in the GC instrument (see Section “Analytical Methods”), the GC equipment was disassembled, and OSCAR hardware was then packed in a vehicle and driven back to Kennedy Space Center (KSC). Ground testing in laboratory (Earth gravity) conditions at KSC was simulated prior to the actual suborbital flight using nominal flight data which was broadcasted via the simulation software tool provided by Blue Origin (the Software Payload Kit). This flight simulation data was relayed to the OSCAR controller through the Blue Origin IPC. After returning the payload from flight and back to KSC, the recorded real flight data was used in the Software Payload Kit (SPK) to emulate the actual suborbital flight events, timings, and warnings; ground tests were repeated to get a better comparison of Earth gravity versus μg performance.

One of the most challenging operational parameters with thermochemical reactors is the start-up and shut-down operation (the non-steady-state regimes). Designing the reactor to operate during a ~3-min μg period was a challenge. This suborbital flight represents the longest operational μg test of a waste conversion reactor known at this time, and so any data will be helpful to inform the design of future spacecraft trash processing systems. The reactor extremities (trash feed, gas feed system, etc.) were built to function and operate as successfully as possible for a 3-min μg demonstration, but they do require redesign for a longer duration human spaceflight operation.

Gas Tank 1 (T1) was opened during the onset of μg, therefore it likely also collected some of the non-μg gases, or pre-combustion event gases. Gas Tank 2 (T2) was opened when T1 closed and continued to collect gas during the μg phase, and Gas Tank 3 (T3) was opened when T2 closed and the μg phase was nearing its completion. T3 was closed and the smoldering tank (S) was opened when smoldering conditions were expected to occur (O₂ turned off) and while the reactor was in a cool-down phase. As was planned, the OSCAR system was dismantled and removed from the payload capsule after flight. This caused gas volume to be lost from the reactor tank (R). To compensate for this, the small “smoldering tank” was installed after the reactor with a solenoid valve to isolate some of the reactor gas volume after the reaction was complete. During ground testing, the OSCAR system was not dismantled after tests, so the reactor gases were successfully collected directly through a port from the manifold itself, in addition to gases in the smoldering tank. Each ground test was performed in triplicate.

Analyses of gases were conducted using an INFICON Micro GC Fusion^TM with a thermal conductivity detector (TCD) that was brought onsite to the Blue Origin West Texas Launch Site for gas analysis after the payload recovery (~ 24 h post flight). The GC was calibrated to quantify O₂, N₂, CO, CH₄, CO₂, C₂H₄, C₂H₆, and detect H₂ and light hydrocarbons (C₂+). The O₂ effluent concentration was recorded in this work to indicate combustion consumption and O₂ availability in the reactor.

All other post launch analyses were performed at KSC. Solids and condensed material were collected and weighed for calculating the solid conversion ratio. Solids were analyzed with field emission scanning electron microscopy (SEM, JEOL JSM-7500F) with energy-dispersive X-ray spectroscopy (EDS). X-ray photoelectron spectrometry (XPS) measurements were conducted with a Thermo Scientific K-Alpha instrument using Al Ka radiation. For each XPS sample type, three regions were analyzed and their atomic percent (at%) values were averaged. Survey scans were collected from −10 eV to 1350 eV and C1s high resolution scans from 279 eV to 298 eV; all spectra were determined from ensemble averaging of 5 individual scans. The carbon types were used to track the degree of sample combustion of the clothing (cotton t-shirt) material in the trash solid remnants. Optical images were obtained with a Nikon Stereo Microscope MSZ-1500, and footage of the reaction was captured using a modified GoPro Hero 4 at 1080 p, 60 fps provided by Blue Origin.

The following flight discrepancies that occurred were different from the initial suborbital flight operation and OSCAR system concept design:

Nominal/simulated flight timings deviated as much as 49% from the real flight timings on a relative basis. The premature flight timing triggers received through the IPC caused the OSCAR ignition sequence to trigger early, thus impacting the combustion process. Discussion on impact to flight events due to the timing differences from simulated data using the SPK versus real flight data are described in detail in another publication (Medina et al., 2020).

These flight discrepancies (#1–5) that deviated from the original design intentions of the system of were replicated during ground testing and defined as the “Flight Conditions” in the results section. Despite the premature flight timing triggers, flight discrepancies and loss of total desired pressure in the system, the ignition of trash still occurred during the μg phase of the flight. Reactor pressure was maintained at 275 kPa (40 psig) during the combustion and thermal degradation events. Valuable data was still collected for observation and follow on the design work of the solid-to-gas conversion system.

Screenshots of the NS-12 OSCAR flight video are shown in Figure 5(A), and ground simulations in Figure 5(B) and (C). Figure 5(B) and (C) presents video screenshot results from ground testing with the following flight discrepancies incorporated into the test: updated actual flight timing triggers, O₂ tank pressure leak, no trash injection from “Trash Injection 1”, no steam injection, and a simulated clogged heater line, allowing only one O₂ inlet port into the manifold. Also for Test (C), the system was purged with air rather than with helium.

Figure 5

From the flight video screenshots of Figure 5(A), an initial flash was at first observed, denoted as “ignition,” followed by darkness (ignition + 2 s) and then a great flash which started a sustained combustion event (ignition + 3 s) that lasted for about 20 s post ignition. The tangential swirl effect for flow was very apparent during this time. Water droplet accumulation on the lid became more pronounced at ignition + 16 s, but flame and smoldering were still strong throughout reactor core. At ignition + 20 s, water accumulation increased on the viewport, and the reactor appeared black inside. At ignition + 1 min 11 s, balls of flames were observed, traveling in a sputter-like fashion, appearing to “jump” around the reactor in a connected pattern. At 1 min 32 s, an intense ignition event occurred again, immediately after the Trash Tube 2 was injected, and the “jumping” fireball effect continued for the rest of the operation until a large glow at 2 min 37 s appeared to be the “start of the end” of the combustion/flame visibility. Smoldering ensued at approximately 2 min 50 s and the viewport visibility diminished due to black soot and water droplet/condensation. Upon post-experiment retrieval, the reactor lid was opaque with a thin coat of black char.

During ground testing with a flight emulation that used helium as a purge gas during payload preparation, ignition began, followed by almost immediate smoldering, as seen in Figure 5(B). The only flame-like color was on the “bottom” (6 o’clock) of the video image, and not all around in a sputtering “dance” as in the μg flight. The reactor was dark and most of the viewport was covered with water droplets immediately. Smoke was not visible as in the flight observation, and it was also difficult to tell when injections occurred via video observations alone. Similar behavior of the ground test for flight emulation that used air as a purge gas during payload preparation was observed as seen in Figure 5(C). Ignition began, followed by almost immediate smoldering. Flames did not appear to “jump” as happened in the flight video and visibility was very poor in ground test emulations.

The suborbital flight gas production was impacted by the O₂ leak. Specifically, the reduced O₂ flow resulted in a higher O₂ residence time in the reactor and less gas being pushed through the BPR. As a result, a much larger portion of combustion gases remained in the reactor under flight conditions than would have occurred at design point operation. Consequently, the collection tanks (T1, T2, T3) had to be filled with nitrogen in order to attain an above atmospheric pressure for GC sampling, which diluted the gases during postflight analysis. Since reactor gases were not collected for the suborbital flight due to practical constraints (as described above), the O₂ leak impacted the quality of the data collected and skewed the mass balance picture for the flight and so, without as much O₂ flow, a larger portion of the product gases remained in the reactor.

Since the reactor gas volume itself was not accessible or recoverable for gas sampling immediately after the experiment, a full mass balance calculation on the OSCAR suborbital flight system was not possible. A complete picture of carbon conversion is a required focus for future flights. To highlight differences between Earth gravity and μg operation, the quality of the gas effluent throughout different stages of collection was investigated during this flight.

Figure 6 displays the intended design point O₂ flow gas production summary on the left and suborbital flight data on the right. During the design point reactions, the O₂ concentration decreased as it was consumed during the dominating combustion process. Once the fuel was consumed, the combustion process diminished. Since O₂ was still flowing after the fuel consumption, the O₂ concentration increased from T3 and reactor tank gas readings as the experiment continued. This behavior was also observed in the 5-s drop testing (Meier et al., 2020). At design point values, the O₂ flow rate provided enough O₂/mixing to perform a near complete combustion demonstration. During the suborbital flight, the O₂ concentration was higher since the system was preloaded and prepped with air (rather than helium as in the design point testing). The O₂ tank leak prior to flight caused a lower starting pressure of O₂ and subsequent flow decay was observed (rather than the constant flow rate). This divergence from the intended design point O₂ flow rate inhibited O₂ consumption and diminished the expected combustion reaction behavior. The flight experiment resulted in a lower mass conversion of trash (~28%) than the pre-flight design ground testing (~85%). Note that very little CO was observed in the μg experiment as would be expected with O₂ scarcity, except for the smoldering tank gas. This suggests that combustion was likely to be transport limited or water in the wet trash reduced temperature, rather than inhibited by O₂ concentration, since despite the reduced flow rate, flight O₂ concentrations were equal to or greater than concentrations observed at design point flow. With reduced flow rate and pressure in the flight system, there was reduced mixing in the reaction zone and potentially less water vapor removal. Video of the combustion demonstrated that most of the burning was located on the stainless-steel mesh in the middle of the reactor, which coincides with the location of the O₂ inlet.

Figure 6

For the μg experiment, the minute amounts of CO did show small increases throughout the tanks but remained < 0.6% in the collection tanks (T1/T2/T3), with 3.2% in the smoldering tank. The flow of O₂ decreased throughout the μg experiment, but concentrations did not fall below 66% in any of the tanks excepting 22% in the smoldering tank (suggesting 40–50% in the reactor). It is extremely likely that the reactor contents would have exhibited a greater degree of incomplete combustion if they had been collected.

CO₂ initially experienced a sharp increase in concentration, then a drop. H₂ production was negligible. As such, it was concluded that steam reforming was not an influential or dominating reaction in the μg experiment.

The three main issues that occurred during the flight and were further evaluated in the lab with independent and combined bases, are described in some of the follow-on graphs as numbers (1, 2, 3) for simplicity. Each event was evaluated in various perspectives to try to understand what impacts the flight discrepancies may have had on the system, since many variables and complexities were present for each case. The conditions are correlated to the following numbers below:

Flight timing trigger—which resulted in the reactor not attaining design pressure prior to ignition.

The gas production that was quantified from the flight experiment was largely consistent with lab testing under analogous conditions. As seen in Figure 7, at the design point, the majority of the synthetic gas (syngas) products were pushed into collection bottles (as opposed to remaining in the reactor). Syngas is described as the total gas products of CO, CO₂, and CH₄ from the system. The flight timing discrepancy (1) and O₂ leak (2) combined reduced gas throughput to the collection tanks. The flight timing discrepancy (1) reduced carrier gas throughput as this effect occurred primarily prior to reaction. The O₂ leak (2), reduced syngas throughput since flow decay occurred during later stages of the experiment, effectively increasing the relative amount of syngas products remaining in the reactor at the experiment end. Preloading the reactor with air (1–3) did not have a significant effect on total products versus preloading with He (1,2). When condition (2) is met, collected syngas correlates well with the flight.

Figure 7

Table 2 details quantified products for total collected gas for the μg and simulated lab runs (not including reactor). Lab – He (1–2) corresponds to the tests described earlier in this section, with the addition that the OSCAR system was prepped and purged with helium. Lab – Air (1,3) corresponds to the tests described earlier where the OSCAR system was prepped and purged with air. Trash injections were identical to flight conditions in all post-flight ground testing.

Both He loaded and Air loaded reactor runs have similar results. The range of measurement error is estimated at ±5%.

The overall amount of syngas production was not significantly different between the lab runs and the flight experiment shown in Table 2 and graphically in Figure 8 (top). There were increased amounts of CO and CH₄ observed in the lab experiments, regardless of He/Air loading at the start. It is unclear whether this was the result of μg effects. It was theoretically possible that μg mixing effects improved surface access for O₂ to oxidize trash particles. It was observed on video that in μg, combustion occurred largely at the stainless-steel mesh (about mid-height in the reactor) with trash particles floating in its vicinity. It is also possible that the higher O₂ concentration in the reactor contributed to a more complete combustion in the flight as compared to the design point. Alternatively, the unrecovered reactor gas from the flight may have exhibited higher CO content (based on the smoldering tank, Figure 8 (top)) due to a delayed combustion in the flight. As such, it is not possible to conclude there were differences in the process and parameters of completion of combustion.

Figure 8

It can be seen in Figure 8 (bottom) that the percentage of syngas in the effluent gas stream is different in the μg experiment. There are two contributors to this discrepancy. First, the reactor was observed to maintain a lower pressure (roughly 34 kPa (5 psig) lower) in the flight than in lab runs preand post- flight. The BPR setpoint was not changed, indicating a possible pressure drop between the reactor and BPR of unknown cause. Second, the O₂ flow rate into the reactor was slightly higher during the flight than lab simulations, in part due to this reactor pressure difference. While the suborbital flight filled the tanks to higher pressures, the increased amount of gas was composed of inlet gases, not syngas products. CO₂ production was the same between lab and μg conditions within experimental variation.

Combusted solid trash remnants were initially evaluated and separated with optical microscopy. While there were several types of solid trash used in the reactor, only t-shirt and food packaging were identifiable and intact to allow for analyses. Among these, only the t-shirt samples provided consistent and comparable results.

Optical microscopy images of four t-shirt samples appear in Figure 9. Figure 9(A) shows a pristine t-shirt sample, Figure 9(B) shows a t-shirt sample recovered from the Lab – He (1,2) experiment, and Figure 9(C) and (D) show t-shirt samples recovered from the μg flight experiment. While both the laboratory (1 g) and μg combustion tests produced a great deal of variability (which is not surprising considering the chaotic nature of solids combustion), these images are generally representative of the recoverable samples observed. In particular, the t-shirt remnants recovered from the 1 g test tended to show a burnt edge with a distinct demarcation from burnt black to scorched brown, as can be seen on the right edge of the sample shown in Figure 9(B). In contrast, the t-shirt remnants recovered from the μg flight test tended to display a more even scorch pattern, without a distinct demarcation between burnt and scorched portions of the sample, as can be seen in the sample shown in Figure 9(C). These burn/scorch patterns are consistent with the flame environments of each piece. For instance, the natural convection provided by an Earth gravity environment allows for the complete burning of the bottom edge of a flat piece of combustible material such as the t-shirt pieces shown here. As the heated gases rise, O₂ is replaced all along the bottom edge. For a μg environment, however, convection is not directional, and simply proceeds outward from the burning surface. Thus, the solid-gas interface quickly becomes O₂ deficient, limiting the degree of combustion. Note, however, that the OSCAR system is designed to induce forced convection from the geometry of the injection of O₂ into the reactor. This, combined with the generally chaotic nature of combustion, resulted in some limited edge burn for a few t-shirt remnants recovered from the μg test. However, the amount of edge burning observed was significantly less than that for the 1 g test. Since these results are only attributed to a single μg test, this observation should be considered qualitative and preliminary. Also, it is interesting to note that both 1g and μg tests resulted in completely blackened t-shirt fragments, with weave patterns intact. Figure 9(D) shows an optical microscopy image of an example of these sorts of samples, in this case a blackened t-shirt sample from the μg test. Complete blackening such as this suggests a pyrolysis process where a portion of the trash is heated in a locally O₂-deficient environment. The fact that blackened debris was found for both flight and lab tests is consistent with the presence of pyrolysis zones in both environments.

Figure 9

XPS was used to analyze several samples from the suborbital NS-12 flight, Lab – He (1,2), and pristine samples. For T-shirt material, five sample types were analyzed: Pristine, dark, and light portions of a sample from the μg test, and dark and light portions of a sample from the Lab – He (1,2) test. A representative survey scan and corresponding high-resolution scan of the C1s peak region are shown in Figure 10(A) and (B), respectively. The survey scan, used for surface elemental analysis, in Figure 10(A) shows several labeled peaks that are the result of C, O, Si, and N (those elements that were ubiquitous to all samples analyzed). The high-resolution C1s spectrum in Figure 10(B) shows the measured data in orange, with the envelope curve overlaid as a black trace (the residual function is shown above the spectrum with a 2X scale). The deconvoluted spectrum was separated into the C-C, C-O, and C=O peaks in Figure 10(B). The envelope is the sum of these constituent curves.

Figure 10

By monitoring the chemical types of carbon present in these samples, information regarding the degree and type of combustion can be derived (Zasada et al., 2015; Wang et al., 2003; Randy Vander Wal et al., n.d.; Liu et al., 2019). The cotton from which the t-shirt material is fabricated is predominantly cellulose, a polymer of glucose. Therefore, the predominant chemical type of carbon should be C-O. Upon combustion, O₂ is removed (in the form of CO₂ and H₂O) at a greater rate than carbon is removed – especially if the reaction is O₂ deficient. This results in an increase in the C-C chemical type (char, like carbon black, is made up predominantly of C-C carbon). As the combustion proceeds, an increase in C=O should be observed due to the stepwise formation of CO₂. Conversely, if pyrolysis occurs, an increased C-C content would be expected, but without the concomitant presence of C=O (since pyrolysis occurs in the absence of O₂). A summary of the quantitative XPS results is shown in Figure 11. Figure 11(A) shows the average distributions of the ubiquitous elements for each sample type, and Figure 11(B) shows the average atomic percent of carbon chemical types for each sample type.

Figure 11

In Figure 11(A), a trend is immediately apparent where the at% of carbon increases in the order of C(Pristine)<C(L)<C(D)<C(B). The reverse pattern can be observed for O₂ such that O(Pristine)>O(L)>O(D)>O(B). These trends are observed in both ground laboratory 1g reaction and μg reaction (though they are slightly different in value with respect to one another). These trends are consistent with the prima facie assumption that the darker the sample, the greater the degree of reaction has taken place, and that the blackened samples represent a more complete carbonization than the black portions of the partially reacted samples. Further insight can be found in the carbon chemical type data in Figure 11(B). Here the Pristine sample shows predominantly C-O, which is expected due to the high cellulose content as described above. However, as the degree of reaction increases, the C-C carbon type immediately eclipses the amount of C-O observed, increasing from (L) to (D) to (B) as with the total carbon in Figure 11(A). Also, the C-O shows a decrease in at%, similar to the overall O content (though not quite as monotonic).

In these phenomena, it is important to keep in mind that XPS is very surface sensitive, sampling only the top ~10 nm. A sample may have initiated the reaction on the surface, while appearing visually as only slightly scorched, or perhaps unreacted. There are two other salient features from Figure 11(B) worth mentioning. First, for each (L) and (D) sample, the at% of C=O remains relatively constant while the C–O at% decreases from (L) to (D) (C=O for flight measured slightly greater than that for the lab test, though these values are not significantly different). The constant C=O might simply be a manifestation of the fact that a C=O moiety is a precursor to CO₂. Thus, any additional double-bonded O₂ would form CO₂ during reaction, and leave the surface, suggesting a maximum observable at% of C=O for these types of reactions. Secondly, for both the flight and the lab tests, the blackened (B) samples showed a decreased amount of C=O relative to their scorched counterparts. This result is consistent with pyrolysis where the sample would be carbonized without the presence of O₂, and therefore without the expected amount of C=O on the surface (note, however, that C=O is still present, as is C-O due to the native presence of O in the cotton cellulose).

In summary, these data suggest that there are no significant differences between the reactions in the μg environment and the 1g (Earth gravity) environment. While the optical microscopy images originally suggested that the natural convection of 1g lab experiments results in different modes of reaction than for the μg flight experiment, any differences could not be elucidated by means of the analyses employed. The authors also investigated the solid remains with scanning electron microscopy. No morphological differences were observed between materials combusted under μg versus 1g, therefore these results are not presented here. It is interesting to note that the presence of both combustion and pyrolysis modes of reaction are corroborated by these results. The similarities between μg and 1g combustion behaviors for these tests is perhaps not surprising since μg combustion rates of thin fuels have been reported to be nearly identical to those for 1g combustion under high partial pressure O₂ conditions such as those described here (Olson, 1991; Olson et al., 2008; Olson, 1987; Olson et al., 1989).

Table 3 shows the carbon conversion, mass conversion, and pertinent parameters for the following experiments: Design-Point O₂ lab, Flight μg, Lab – He (1,2), and Lab – Air (1,2). Also shown are data for the ZGF experiments described earlier (Meier et al., 2020). Note that the mass conversion value for the flight experiment shown in Table 3 should be considered as a low confidence measurement due to the loss of the reactor gases as mentioned previously. Carbon conversion is a better assessment of the performance variation as it is based on product quantification from closed tanks. In terms of carbon conversion, the flight performed similarly to lab testing. Still, the carbon conversion metric is subject to variability in trash carbon content compounded with the experimental measurement range of error. The suborbital flight was not able to account for gases left in the reactor. For the design point experiment, this represented 10–20% of gas production. However, due to the decrease in O₂ flow, around 50% of the gases produced stayed in the reactor (according to analogous lab testing) resulting in a large portion of syngas to be unquantified. The main takeaway is that the suborbital flight exhibited carbon conversion within the same range as the lab simulations when only the gases collected in tanks is considered. It is important to note that the flight experiment was not meant to achieve the highest mass or solid conversion possible, but rather observe the trash-to-gas phenomenon in a μg environment and compare its design differences in Earth gravity.

The lab experiments had a mean carbon conversion of 15.2% and a standard deviation of 1.87%. Flight data showed a difference of 1.5% (at 13.8%), within one standard deviation. Indeed, a single-sample t-test indicated that the flight value of 13.8% could not be rejected from the interval for the measurement (t = 1.79, significance = 0.05, and 4 degrees of freedom), showing that the flight value is not in disagreement with the distribution of laboratory measurements.

The carbon conversion, mass conversion, reactor thermocouple temperatures, ignition time associated with combustion timing, pressure increased due to reaction, and the CO₂:CO ratio are also displayed in Table 3 from both suborbital flight μg and ground testing, as well as the prior μg testing at the GRC ZGF (an average from 3 runs). It is important to note a few differences from the data comparison of prior μg tests: At ZGF, High Fidelity Waste Simulant (HFWS) was used instead of OWS. HFWS included the urine in the waste simulant. The ZGF also used the trash injection as t = 0 s to count the time of ignition (since the reactor was preheated and already hot), whereas the suborbital flight μg and ground testing used the time from when the cartridge heaters were turned on. The ZGF only had one single thermocouple probe inside the reactor, whereas the suborbital flight reactor had the multi-point thermocouple. And finally, all of the following were different in ZGF testing: the trash starting mass, the O₂ flow rate was different, and the thermal reactor zone design (a copper core was used in the manifold). So, it is very difficult to compare ZGF with the suborbital data, but it is presented as a reminder to the reader of the only other known data for a reference-based trash-to-gas μg test.

It is interesting to note that for the suborbital testing, the μg thermocouple maximum temperature was recorded at the TC2 location, whereas the maximum temperature was read at TC3 for all ground tests. In addition, the smallest temperature gradients between the TC regions existed in the μg TC2, TC3, and TC4 regions compared to the Earth gravity lab testing, but the largest temperature gradient between TC1 and TC2 was in the μg flight test. Referring back to the video descriptions, it can be inferred that the tangential swirl effect for flow was more efficient in μg to mix the gases in an even hearth zone than the 1g experiments, which have the effects of gravity for convection as well as pushing the waste to the bottom of the manifold. Visual observations from the video also correlate that the “balls of flames” traveled in a more even sputter rotation around the reactor core whereas the flame from the Earth gravity tests did not travel around the reactor system. For the 1g tests, the top of the reactor may have been warmer at the TC1 location as the immediate condensation and vapor accumulation at the top near the viewport insulated the reactor system, creating a warmer vapor environment in the TC1 region than observed in μg. The midpoint of the reactor was where the feedline heater injected O₂ into the reactor. Experimental flight data confirmed model validation and a correlation between the predicted and experimental temperature profile. TC2 and TC3 received a constant feed of O₂, constituting the most active, highest temperature reaction zone. Top and bottom of the reactors (TC 4 and TC1) maintained a lower temperature profile during the sustained combustion. It can be concluded that the reactor recirculation design promoted the mixing of the gases in the mid-section zone of the reactor during μg. Figure 12 displays the observed results in which the hearth zone, or reaction zone, was expanded in μg operation, while the hearth zone in the Earth gravity testing was at a higher temperature but a smaller zone. For more details, the reader is directed to the authors’ previous work where there are several plots and explanations that help visualize these performance results from flight (Medina et al., 2020).

Figure 12

Thermocouple data confirms that the conditions were ideal for pyrolysis and gasification. Based on ignition behavior and CO₂, H₂O vapor, and heat generation, we can say that combustion was also successful. Steam reforming may have occurred due to the subtle moisture content in the OWS; however, since it was known that steam did not enter the system from the syringe pump mechanism, and H₂ was not detected, steam reforming is not considered influential or dominating in these experiments.

The CO₂:CO ratio was highest in the suborbital μg experiment. So it can be inferred that the combustion reaction was more efficient in μg than Earth gravity. Since the carbon conversion appears “statistically’ undifferentiable for these tests, it will be important to collect more μg data points to establish conclusion on theories about hearth behavior and solid to gas conversion effectiveness of solid C_xH_yO_z to CO₂ gas. It is also important to note that this data is not complete since the major part of CO was likely to be in the reactor for the μg experiment, and a reactor gas sample was not collected due to system design. Thus, additional work is required before a conclusive result may be drawn. It was expected that the pressure increase due to reaction would be lowest in the μg data point than the lab testing, since the starting pressure was lower during flight compared to any other runs before and after the flight. Correlations between the ignition time, pressure differential, conversion, and CO₂:CO ratio were not observed from this data. One would believe that increased pressure would drive the reaction rates toward completion of solid to gas, which may attribute to the increase in mass conversion at higher reaction pressures for the laboratory tests, but the importance seems to lie on the actual reactor design and in increasing the volume of the hearth zone for more efficient conversion to occur at elevated temperatures.