Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “oxygen diffusion”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 163 records · Page 9
All sourcesNASA NTRSDOE OSTI
NASA NTRS1989

MoS2 interactions with 1.5 eV atomic oxygen

Exposures of MoS2 to 1.5-eV atomic oxygen in an anhydrous environment reveal that the degree of oxidation is essentially independent of crystallite orientation, and that the surface-adsorbed reaction products are MoO3 and MoO2. A mixture of oxides and sulfide exists over a depth of about 90 A, and this layer has a low diffusion rate for oxygen. It is concluded that a protective oxide layer forms on MoS2 on exposure to the atomic-oxygen-rich environment of LEO.

Martin, J. A.↗
NASA NTRS2003

Investigation of Dynamic Oxygen Adsorption in Molten Solder Jetting Technology

Surface tension forces play a critical role in fluid dynamic phenomena that are important in materials processing. The surface tension of liquid metals has been shown to be very susceptible to small amounts of adsorbed oxygen. Consequently, the kinetics of oxygen adsorption can influence the capillary breakup of liquid-metal jets targeted for use in electronics assembly applications, where low-melting-point metals (such as tin-containing solders) are utilized as an attachment material for mounting of electronic components to substrates. By interpreting values of surface tension measured at various surface ages, adsorption and diffusion rates of oxygen on the surface of the melt can be estimated. This research program investigates the adsorption kinetics of oxygen on the surface of an atomizing molten-metal jet. A novel oscillating capillary jet method has been developed for the measurement of dynamic surface tension of liquids, and in particular, metal melts which are susceptible to rapid surface degradation caused by oxygen adsorption. The experimental technique captures the evolution of jet swells and necks continuously along the jet propagation axis and is used in conjunction with an existing linear, axisymmetric, constant-property model to determine the variation of the instability growth rate, and, in turn, surface tension of the liquid as a function of surface age measured from the exit orifice. The conditions investigated so far focus on a time window of 2-4ms from the jet orifice. The surface properties of the eutectic 63%Sn-37%Pb solder alloy have been investigated in terms of their variation due to O2 adsorption from a N2 atmosphere containing controlled amounts of oxygen (from 8 ppm to 1000 ppm). The method performed well for situations where the oxygen adsorption was low in that time window. The value of surface tension for the 63Sn-37Pb solder in pure nitrogen was found to be 0.49 N/m, in good agreement with previously published work. A characteristic time of O(1ms) or less was determined for the molten-metal surface to be saturated by oxygen at 1000 ppm concentration in N2.

Megaridis, Constantine M.↗
NASA NTRS1996

Effect of chemistry and turbulence on NO formation in oxygen-natural gas flames

The effects of chemistry and turbulence on NO formation in oxygen-natural turbulent diffusion flames gas flames have been investigated. The chemistry of nitric oxides has been studied numerically in the counterflow configuration. Systematic calculations with the GRI 2.11 mechanism for combustion of methane and NO chemistry were conducted to provide a base case. It was shown that the 'simple' Zeldovich mechanism accounts for more than 75% of N2 consumption in the flame in a range of strain-rates varying between 10 and 1000 s-l. The main shortcomings of this mechanism are: 1) overestimation (15%) of the NO production rate at low strain-rates because it does not capture the reburn due to the hydrocarbon chemistry, and 2) underestimation (25%) of the NO production rate at high strainrates because it ignores NO production through the prompt mechanism. Reburn through the Zeldovich mechanism alone proves to be significant at low strain-rates. A one-step model based on the Zeldovich mechanism and including reburn has been developed. It shows good agreement with the GRI mechanism at low strain-rates but underestimates significantly N2 consumption (about 50%) at high strain-rates. The role of turbulence has been assessed by using an existing 3-D DNS data base of a diffusion flame in decaying turbulence. Two PDF closure models used in practical industrial codes for turbulent NO formation have been tested. A simpler version of the global one-step chemical scheme for NO compared to that developed in this study was used to test the closure assumptions of the PDF models, because the data base could not provide all the necessary ingredients. Despite this simplification, it was possible to demonstrate that the current PDF models for NO overestimate significantly the NO production rate due to the fact that they neglect the correlations between the fluctuations in oxygen concentration and temperature. A single scalar PDF model for temperature that accounts for such correlations based on laminar flame considerations has been developed and showed excellent agreement with the values given by the DNS.

Samaniego, J. -M.↗
NASA NTRS1976

Aerospike thrust chamber program

An existing, but damaged, 25,000-pound thrust, flightweight, oxygen/hydrogen aerospike rocket thrust chamber was disassembled and partially repaired. A description is presented of the aerospike chamber configuration and of the damage it had suffered. Techniques for aerospike thrust chamber repair were developed, and are described, covering repair procedures for lightweight tubular nozzles, titanium thrust structures, and copper channel combustors. Effort was terminated prior to completion of the repairs and conduct of a planned hot fire test program when it was found that the copper alloy walls of many of the thrust chamber's 24 combustors had been degraded in strength and ductility during the initial fabrication of the thrust chamber. The degradation is discussed and traced to a reaction between oxygen and/or oxides diffused into the copper alloy during fabrication processes and the hydrogen utilized as a brazing furnace atmosphere during the initial assembly operation on many of the combustors. The effects of the H2/O2 reaction within the copper alloy are described.

Campbell, J., Jr.↗
NASA NTRS1993

Simulations of Multidimensional Burner-stabilized Flames

Detailed numerical simulations have been performed to study the structure and dynamics of downward propagating burner-stabilized flames in lean hydrogen-air mixtures. These simulations include the effects of fluid convection, detailed hydrogen-oxygen chemistry, multispecies diffusion, thermal conduction, viscosity, and heat losses to the burner. One-dimensional calculations have been carried out to investigate burner boundary conditions. Well known features of a burner-stabilized flame have been reproduced. Two-dimensional calculations show the presence of cellular structures at the burner surface. At low inlet velocities, these cellular structures are suppressed by the increased heat loss to the burner. These simulations have been examined in detail to gain understanding of the similarities and differences in structure and stability of freely propagating and burner-stabilized flames.

Patnaik, G.↗
NASA NTRS1998

Mechanism-based Modeling of Long-term Degradation

The use of composites in high temperature, long lifetime applications requires a basic understanding of composite degradation mechanisms, advances in analytical capabilities, and accurate accelerated and scaled tests. To advance all of these goals, models are proposed based on a variety of fundamental material mechanisms. Thermal, oxygen, and moisture diffusion, chemical reactions, composite micromechanics; modified laminated plate theory, and future mechanics based damage models are used. All models attempt to stay as simple and fundamental as possible. All are coupled, so that interactions between various effects are modeled implicitly. Ongoing efforts at MIT are reviewed here, with some reference to other work, but no attempt is made to do a comprehensive review. Mechanism based models are yielding understanding of the mechanisms behind observed degradation phenomena, are helping to design accelerated tests, and are the first steps towards a predictive capability.

McManus, H. L.↗
NASA NTRS2000

Structure of Laminar Permanently Blue, Opposed-Jet Ethylene-Fueled Diffusion Flames

The structure and state relationships of laminar soot-free (permanently blue) ethylene-fueled diffusion flames at various strain rates were studied both experimentally and computationally using an opposed-jet configuration. Measurements of gas velocities, temperatures, and compositions were carried out along the stagnation stream line. Corresponding predictions of flame structure were obtained, based on numerical simulations using several contemporary reaction mechanisms for methane oxidation. Flame conditions studied included ethylene-fueled opposed-jet diffusion flames having stoichiometric mixture fractions of 0.7 with measurements involving strain rates of 60-240/s and predictions involving strain rates of 0-1140/s at normal temperature and pressure. It was found that measured major gas species concentrations and temperature distributions were in reasonably good agreement with predictions using mechanisms due to GRI-Mech and Peters and that effects of preferential diffusion significantly influence flame structure even when reactant mass diffusivities are similar. Oxygen leakage to fuel-rich conditions and carbon monoxide leakage to fuel-lean conditions both increased as strain rates increased. Furthermore, increased strain rates caused increased fuel concentrations near the flame sheet, decreased peak gas temperatures, and decreased concentrations of carbon dioxide and water vapor throughout the flames. State relationships for major gas species and gas temperatures were found to exist over a broad range of strain rates, providing potential for significant computational simplifications for modeling purposes in some instances.

Lin, K.-C.↗
NASA NTRS2000

Structure of Laminar Permanently Blue, Opposed-Jet Ethylene-Fueled Diffusion Flames

The structure and state relationships of laminar soot-free (permanently blue) ethylene-fueled diffusion flames at various strain rates were studied both experimentally and computationally using an opposed-jet configuration. Measurements of gas velocities, temperatures, and compositions were carried out along the stagnation stream line. Corresponding predictions of flame structure were obtained, based on numerical simulations using several contemporary reaction mechanisms for methane oxidation. Flame conditions studied included ethylene-fueled opposed-jet diffusion flames having stoichiometric mixture fractions of 0.7 with measurements involving strain rates of 60-240/s and predictions involving strain rates of 0-1140/s at normal temperature and pressure. It was found that measured major gas species concentrations and temperature distributions were in reasonably good agreement with predictions using mechanisms due to GRI-Mech and Peters and that effects of preferential diffusion significantly influence flame structure even when reactant mass diffusivities are similar. Oxygen leakage to fuel-rich conditions and carbon monoxide leakage to fuel-lean conditions both increased as strain rates increased. Furthermore, increased strain rates caused increased fuel concentrations near the flame sheet, decreased peak gas temperatures, and decreased concentrations of carbon dioxide and water vapor throughout the flames. State relationships for major gas species and gas temperatures were found to exist over a broad range of strain rates, providing potential for significant computational simplifications for modeling purposes in some instances.

Lin, K.-C.↗
NASA NTRS2015

Alumina-Forming MAX Phases in Turbine Material Systems

Coatings for high temperature turbine components are based on low conductivity YSZ thermal barriers and protective NiAl, NiCoCrAlY bond coats. Good oxidation hot corrosion resistance, intermediate CTE, and strain tolerance of Ti2AlC and Cr2AlC MAX phases are thus of special interest. Their alumina scale growth follows a cubic law in accord with FeCrAlY alloys, with oxygen grain boundary diffusivity: Dgb 1.8 x 10-10 exp(-375 kJmole) m3s. Protective cubic kinetics are also found in high pressure burner rig (6 atm., 25 ms) and TGA tests of MAXthal 211Ti2AlC. The initial portion (0.1 hr) is dominated by fast TiO2 growth (with little evidence of scale volatility in high pressure water vapor, as found for SiO2 scales). Bulk Ti2AlC and Cr2AlC substrates show promise as potential bond coats for YSZ TBCs in 1000-1200 C furnace life (500 h) tests. Cr2AlC is proving to be very resistant to 700-900 C Na2SO4 hot corrosion and is of interest for disk alloys. Preliminary diffusion bonded Cr2AlC-superalloy hybrid couples have survived 1000 hr interrupted furnace tests at 800C with no indication of cracking or debonding. Diffusion zones of -NiAl+Cr7C3 were produced in these above 1000 C, but did not grow to any great extent after 1000 hr at 800 C. Processing as coatings presents challenges, however the basic properties of MAX phases provide novel opportunities for high temperature turbine components.

MAX phase↗
NASA NTRS2022

Effect of Sulfur on siderophile element partitioning between olivine and martian mantle primary melt

Ni and Co variations in primary martian magmas exhibit anomalous incompatible behavior, which has remained an unexplained conundrum. Because martian magmas are S-rich, and some trace metals are reported to have enhanced solubility in S-bearing magmas, we have carried out a series of experiments to evaluate the effect of high-S melts on the olivine/melt partitioning of Ni, Co, Mn, V, and Cr. Near-liquidus experiments on a synthetic primary martian mantle melt (Yamato-980459 [Y98]) were completed in a piston-cylinder apparatus at 0.75 GPa. Previous studies in S-free systems illustrate that the partition coefficients for these elements are dependent chiefly on D(Mg(Ol/melt)) (the partition coefficient defined as wt% Mg in olivine/wt% Mg in melt, a proxy for temperature), and were used to calibrate a predictive expression that includes the effects of temperature [i.e., D(Mg(Ol/melt))], melt composition, and oxygen fugacity. These predictive expressions are then used to isolate any effect in DM olivine/melt due to dissolved sulfur. The results show that S might have a small effect for Co, but not enough to change Co partitioning from compatible to incompatible in our experiments. The addition of a sulfur term to the D(Co) predictive expressions shows that nearly 8000 ppm of sulfur would be required in the melt (at liquidus temperature of Y98) for D(Co) to become <1. These S contents are two times higher than those of a sulfide-saturated melt at the P-T conditions of a martian mantle source region. Therefore, the anomalous incompatible behavior observed in these primary magma suites must be due to another mechanism. High temperature, oxygen fugacity, and diffusion are not viable mechanisms, but magma mixing, assimilation, or kinetic crystallization effects remain possibilities.

shergottite↗
NASA NTRSDate not supplied

Oxygen Isotopic Composition of Refractory Inclusions from the Miller Range (MIL) 090019 Carbonaceous Chondrite

Introduction: Primitive meteorites contain Calcium-Aluminum-rich Inclusions (CAIs)that preserve the records of the earliest times in the Solar System. CAIs record O isotopic variations with time and/or location from solar (16O-rich) to planetary (16O-poor) compositions [e.g., 1, 2].However, CAIs were also affected by parent body processing to varying degrees. The goal of this work was to identify the processes that resulted in the fine-scale spatial variations in oxygen isotopic compositions within the most ancient Solar System solids. We report oxygen isotopic composition of CAIs from MIL 090019 CO3.1 carbonaceous chondrite. MIL 090019 contains a high abundance of CAIs showing mineralogical as well as textural variations [3]. Methods: We analyzed 10 CAIs with different mineralogical assemblages, classified as corundum-bearing, grossite-bearing, hibonite-bearing, melilite-bearing, anorthite-bearing inclusions and amoeboid olivine aggregates(AOA’s). The mineralogical and petrological characterization of CAIs was performed using the JOEL Hyperprobe 8530 electron microprobe at NASA JSC. Oxygen isotopic imaging of the CAIs was done using the Cameca NanoSIMS 50L ion microprobe at NASA JSC. We followed the analytical protocol described in [4]. O-isotopic maps of the CAI were acquired by rastering a ~3 pA primary Cs+beam at 16 keVover an area of 20 ×20μm2over a period of ~7 hours. Negative secondary ions of 16O−, 17O−, 18O−, 28Si−, 24Mg16O−, 27Al16O−, and 40Ca16O−were simultaneously acquired using electron multiplier detectors at a mass resolving power sufficient to resolve the 16OH−interference from the 17O−peak, where the contribution of 16OH−was <0.1%.An electron flood gun was used to mitigate sample charging during the analyses. We used San Carlos olivine and Madagascar hibonite as isotopic standards to correct for the instrumental mass fractionation. The O-isotopic ratios were corrected for the quasi simultaneous arrival (QSA) effect and the detector dead time. All reported errors are 1 sigma. Results: The MIL 090019 carbonaceous chondrite hosts refractory inclusions varying in their mineralogies and textures[2, 3]. We found the oxygen isotopic compositions of mineral phases in MIL 090019 CAIs show large variations with Δ17O values varying from -27 to 0‰. A corundum-bearing inclusion records isotopic variations with corundum, melilite, perovskite, and anorthite showing Δ17Ovalues of -20 ±2.5, -10.2 ± 6.5, -10.7 ± 1.3,and -1.2 ± 3.2‰ respectively. A grossite-hibonite-bearing inclusion also records O-isotopic heterogeneity, where grossite, hibonite, Mg-rich spinel, and melilite show -3.7 ± 2.7, -22.9 ± 2.5, -21.4 ± 2.6, and -17.3 ± 2.8‰ respectively, whereas perovskite grains record a heterogenous isotopic composition ranging from -22.6 to -13.9‰. A hibonite-bearing inclusion shows average Δ17Ο values of -17.6 ± 5.3 ‰ and -15.1 ± 4.8 ‰ for hibonite and spinel respectively. The two perovskite grains also show relatively 16O-poor composition (Δ17Ο = -11.8 ± 6.4 ‰ and -8.6 ± 5.2 ‰). Anorthite has ~0.1 wt% Na2O and ~0.52 wt% FeO and a Δ17Ο value of -7.2 ± 5.0 ‰, whereas diopside records a Δ17Ο value of -3.5 ± 5.0 ‰. A hibonite-spinel-rich inclusion contains spinel and hibonite that are16O-rich (Δ17Οvalues -21.3 ± 2.8‰ and -19.4 ± 3.0‰, respectively). Melilite shows heterogenous oxygen isotopic composition, with an inner 16O-rich region and an outer 16O-poor region separated by a sharp boundary, with a bulk Δ17Ο= -16.2 ± 2.1‰, whereas two perovskite grains are relatively 16O-poor (Δ17Ο= -13.2 ± 3.7‰ and -7.1 ± 3.4‰).A spinel-rich inclusion contains perovskite grains with Δ17Οranging from -27.0 to -22.8 ‰and spinel with average Δ17Οvalue of -21.4 ± 2.8 ‰. In a melilite and spinel-rich CAI, perovskite, spinel, diopside, melilite, and Al-Ti-rich pyroxene record Δ17Ο values of-26.9 ± 4.6‰, -22.4 ± 4.3‰,-23.4 ± 4.9, -19.3 ± 3.4, and -19.0 ± 4.3 ‰ respectively. Olivine and spinel components of 2 AOA’s record a 16O-rich composition with Δ17O~-23‰. Discussion: The coexistence of 16O-rich and 16O-poor minerals within CAIs has been attributed to differing degrees of O isotopic exchange, reflecting a wide range of O diffusion rates[5]. However, the O isotopic heterogeneity observed in these CAIs cannot be explained by oxygen diffusion in nebular or parent body settings alone. The isotopic imaging of CAIs from the MIL 090019 meteorite shows variations in the O isotopic composition recorded in their mineral components, suggesting that these CAIs record the oxygen isotopic heterogeneity in the nebular gas from which they condensed. The coexistence of 16O-rich and 16O-poor mineral components in the CAI-forming region suggests that the gas in this region was not well-mixed and consisted of distinct 16O-rich and 16O-poor gaseous reservoirs.

P Mane↗
NASA NTRS1981

Energetic oxygen and sulfur in the Jovian magnetosphere

Measurements made in the Jovian magnetosphere by the cosmic ray subsystem on Voyager 1 and 2 are reported. Energy spectra of oxygen ions in the energy range 1-20 MeV/nuc between 5 and 20 Jovian radii are presented, and phase space densities are calculated. A steep positive radial gradient in the phase space density of the energetic oxygen ions is observed, indicating an inward diffusive flow. The upper limit on the rate at which oxygen ions with greater than 400 MeV/nuc-G diffuse across 10 Jovian radii is calculated to be 5 x 10 to the 21st ions per second, indicating that about 10 to the -7th of the ions from Io are accelerated to over 400 MeV/nuc-G and diffuse to 10 Jovian radii. Observations also suggest that oxygen and sulfur ions in the Io plasma torus diffuse radially outward, are nonadiabatically accelerated in a region outside 17 Jovian radii, and diffuse inward and outward from the accelerated region.

Gehrels, N.↗
NASA NTRS1981

Assessment of high temperature metallizations for I(2)L and CMOS technologies

High temperature barrier metallizations were assessed and tested for I(2)L and CMOS applications. Life tests were accelerated to 375 C in view of the -55 C to +300 C temperature range established for engine-located electronics without fuel cooling. The gold-refractory metallizations evaluated were Au-TiW-PtSi, Au-TiW/TiO2/TiW-PtSi and Au-TiW(N)-PtSi. These metallization systems were thermally annealed to at least 375 C for up to 250 hours. The critical requirement for stable diffusion barrier is the TiW grain size. Small grain (250A-500A) films were observed to be stable up to 375 C. Deposition to TiW diffusion barrier in the presence of oxygen and nitrogen also results in an effective diffusion barrier. Auger electron spectroscopy profiles of the PtSi indicates some penetration by the TiW. In the case of PtSi/TiW interface, the redistribution of oxygen further passivates the system by forming a TiO2 layer at the interface.

Christou, A.↗
NASA NTRS2017

Diffusion of Redox-Sensitive Elements in Basalt at Different Oxygen Fugacities

The terrestrial planets and moons of our solar system have differentiated over a range of oxygen fugacity conditions. Basalts formed from magmas on the Earth cover a range of more oxidized states (from approximately IW (iron wustite) plus 2 to approximately FMQ (fayalite-magnetite-quartz) plus 3) than crustal rocks from Mars (IW to approximately IW plus 3), and basalts from the Moon are more reduced than both, ranging from IW to IW minus 2. The small body Vesta differentiated around IW minus 4. Characterization of redox sensitive elements' diffusivities will offer insight into behavior of these elements as a function of f (fugacity of) O2 for these planetary bodies. Here, we report a systematic study of the diffusion of redox-sensitive elements in basaltic melts with varying oxygen fugacities (fO2) for trace elements, V, Nb, W, Mo, La, Ce, Pr, Sm, Eu, Gd, Ta, and W. Since fO2 is an intensive variable that is different for the reservoirs of various planets and moons in our solar system, it is important to characterize how changes in redox states will affect diffusion. We conducted experiments in a piston cylinder device at 1300 degrees Centigrade and 1 gigapascal, at the University of Rochester and NASA Johnson Space Center. We buffered some experiments at Ru-RuO2 (FMQ plus 6.00), and conducted other experiments within either a graphite or Mo capsule, which corresponds to fO2s of either FMQ minus1.2, or FMQ minus 3.00, respectively. Characterizing the diffusivities of redox sensitive elements at different fO2s is important because some elements, like Eu, have varying valence states, such as Eu (sup 2 plus) and Eu (sup 3 plus). Differences in charge and ion radii may lead to differences in diffusivities within silicate melts. This could, lead to formation of a Eu anomaly by diffusion, the magnitude of which may be controlled by the fO2. Characterization of trace element diffusion is also important in understanding trace element fractionation. We found, during the course of our investigation, that not only did the diffusivities of the redox sensitive elements change with fO2, but that the diffusivities of all other analyzed elements also changed. This indicates that not only do changes in valence influence trace elements diffusivities but that the structure of melt may have changed with varying oxygen fugacity, probably due to changes in the speciation of the major element Fe.

Szumila, I.↗
NASA NTRS1998

Enhanced Glow Discharge Production of Oxygen

Studies starting in late seventies have shown Mars atmosphere can be used as a feedstock for oxygen production using simple chemical processing systems during early phases of the Mars exploration program. This approach has been recognized as one of the most important in-situ resource utilization (ISRU) concepts for enabling future round trip Mars missions. It was determined a decade ago that separation of oxygen can be accomplished efficiently by permeation through a silver membrane at temperatures well below 1000 K. This process involves adsorption of atomic oxygen on the surface and its subsequent diffusion through a silver lattice via an oxygen concentration gradient. We have determined recently that glow discharge can be used to liberate atomic oxygen from Mars atmosphere and that the oxygen can be collected through a silver permeation membrane. Recently, we demonstrated a substantial increase in energy efficiency of the process by applying a radio frequency discharge in combination with a silver permeation membrane. The experiments were performed using pure carbon dioxide in the pressure range equal to Mars surface conditions. Energy efficiency was defined as the ratio of the energy required to dissociate a unit mass of oxygen from carbon dioxide to the (electrical) energy consumed by the overall system during the dissociation and collection process. The research effort, started at NASA Langley Research Center, continued with this project. Oxygen production apparatus, built and operated under the research grant NAG1-1140 was relocated to the Atomic Beams Laboratory at ODU in July 1996, being since then in fall operation.

Ash, Robert↗
NASA NTRS2003

A Role of the Reaction Kernel in Propagation and Stabilization of Edge Diffusion Flames of C1-C3 Hydrocarbons

Diffusion flame stabilization is of essential importance in both Earth-bound combustion systems and spacecraft fire safety. Local extinction, re-ignition, and propagation processes may occur as a result of interactions between the flame zone and vortices or fire-extinguishing agents. By using a computational fluid dynamics code with a detailed chemistry model for methane combustion, the authors have revealed the chemical kinetic structure of the stabilizing region of both jet and flat-plate diffusion flames, predicted the flame stability limit, and proposed diffusion flame attachment and detachment mechanisms in normal and microgravity. Because of the unique geometry of the edge of diffusion flames, radical back-diffusion against the oxygen-rich entrainment dramatically enhanced chain reactions, thus forming a peak reactivity spot, i.e., reaction kernel, responsible for flame holding. The new results have been obtained for the edge diffusion flame propagation and attached flame structure using various C1-C3 hydrocarbons.

Takahashi, Fumiaki↗