The Development and Use of a Natural Gas/Oxygen Burner Rig for Environmental Barrier Coating and Ceramic Matrix Composite Technology Maturation

The burner rig facility located within the Materials Research Laboratory at the National Aeronautics and Space Administration's Glenn Research Center is used for high temperature materials research including oxidation, corrosion, erosion, and impact. Within the facility are six Mach 0.3 to Mach 1.0 capable, computer-controlled, jet-fueled combustors in individual test cells. These burners operate on jet-A and preheated air. A schematic showing a representation of a standard burner rig is presented in Fig. 1. These burner rigs are used as an efficient means of subjecting advanced materials for aero applications to the high-temperature, high-velocity, and thermal cycling environment that more closely approximates engine operating conditions. Moreover, modifications and/or additions have been implemented in the burner rigs to study other environmental effects such as salt corrosion, foreign object damage, and solid particle erosion [1–5]. One limitation to the current, atmospheric, jet-fueled burner rigs is that the maximum sample temperature that can be achieved during normal operation is approximately 2500 °F (1371 °C). This restricts the ability to test higher temperature capable materials. For example, next generation environmental barrier coatings (EBCs) and ceramic matrix composites (CMCs) are targeted for temperatures >2700 °F (>1482 °C). Another limitation is the lower water vapor content present in the jet-fueled burner rig combustion, approximately 10% H₂O, limits the study of material recession, which is a critical life-limiting mechanism for silicon (Si)-containing materials such as silicon carbide (SiC)-based CMCs.

To combat these limitations, and to bridge the gap between other laboratory methods (furnaces, high-heat flux lasers, etc.) and engine rig testing, a new natural gas/oxygen (NG/O₂) fueled combustion rig was developed for high temperature environmental durability studies at atmospheric pressure. The NG/O₂ rig can simulate the high-temperature, high-heat flux, and high-velocity thermal environments encountered in gas-turbines. The NG/O₂ rig has a maximum surface temperature capability >3000 °F (>1649 °C) and results in higher water vapor content compared to the jet-fuel atmospheric burner rigs.

This work provides an overview of the development for the new NG/O₂ burner rig, initial characterization, an initial study to characterize the suitability of an EBC/CMC material system for use as a combustion liner for commercial supersonic technology (CST), and a brief synopsis of how the NG/O₂ rig is supporting high temperature EBC/CMC technology maturation under NASA's Hybrid Thermally Efficient Core (HyTEC) project.

The equipment for the NG/O₂ burner is a custom unit made by Carlisle Machine Works (Millville, NJ) that can consume up to 700 SCFH (standard cubic feet per hour) of natural gas, and 1500 SCFH of oxygen. A picture of the unit is displayed in Fig. 2. An interchangeable selection of 702, 705, or 707 series lathe burner heads can be used to produce flames with different diameters. The 702 series is ¾ inch while the 705 and 707 are 1 inch and 1½ inches in diameter, respectively. Figure 3 shows the three different burner heads (nozzles) that can be utilized in the rig. The hoses, solenoid valves, gas regulators, mass flow controllers, controls, check valves, and E-stops are integrated into the power station platform.

The rig has a dual electrode pilot to ignite and sense the flame. If no flame is detected, the system will shut off the natural gas and oxygen flow and alert the operator. A programmable logic controller (PLC) (Allen-Bradley^®, Milwaukee, WI) with touchscreen was supplied with the burner for operator control and provides ignition, flame-out alarm, safety shutdown, and gas flow control as shown in Fig. 4.

The automation software platform wonderware was used to merge the supplied Allen-Bradley^® PLC with a Modicon PLC to develop one common program for a Human Machine Interface (HMI) screen as shown in Fig. 5. The HMI enables additional operator control and monitoring of functions.

The rig is integrated into the building natural gas supply for a continuous flow of fuel while the oxygen is supplied using an on-site oxygen generator system (Oxygen Generating Systems International, North Tonawanda, NY). Implementing an oxygen generator system to provide an uninterrupted supply of oxygen enables continuous (24/7) operation for long-term durability studies.

The oxygen generating system supplies ∼93% purity oxygen to the rig. The process of generating oxygen enriched air (∼93% O₂ and ∼7% N₂) begins by passing filtered, 110-psig shop air from the building supply through a refrigerated air dryer (Parker Hannifin, Lancaster, NY) to remove water vapor. The dried air is transferred into a 120-gal capacity service air holding tank prior to being fed into the oxygen generators. Three OG-500 oxygen generators are used to separate oxygen from the dried shop air using OGSI's Pressure Swing Adsorption Technology which centers around a Zeolite molecular sieve [6]. The oxygen enriched air is then stored in two 400-gal capacity holding tanks that are piped directly to the rig. Figure 6 shows the main components of the oxygen generating system: air dryer, oxygen generators, service air holding tank, and the oxygen enriched air (∼93% O₂ and ∼7% N₂) holding tanks.

Wonderware Historian is used for data acquisition, data storage, and monitoring the status of the rig. All controls, functions, alerts/alarms, and test specific data (pyrometers, etc.) are displayed in real-time and stored to the computer for postprocessing. The storage of active alerts/alarms is used to notify the operator and aid in troubleshooting in case of a rig shutdown when operating unattended. The HMI also enables the ability to set upper/lower pyrometer (temperature) limits, and a test timer to safely shut down the rig if an operator is not present.

Figure 7 shows the NG/O₂ burner during its initial operation.

As discussed previously, the NG/O₂ burner rig uses ∼93% purity oxygen. Oxygen enriched combustion provides several benefits over the use of air (21% O₂, 78% N₂, and 1% trace gases) including increased combustion efficiency, lower emissions, and higher combustion temperatures [7]. A basic combustion analysis was performed using the NASA developed Chemical Equilibrium with Applications (CEA) program [8]. The program is used to calculate chemical equilibrium product concentrations and thermodynamic properties for complex mixtures.

The composition of natural gas can vary widely depending upon production region and even supplier. Table 1 shows the values that were used in the initial analysis. These values were obtained from the John Zink Combustion Handbook [9] and were determined to be a reasonable approximation to the range of values that can be found in the Ohio (U.S.) region. Since methane is the major component of natural gas (>93 vol. %), the CEA analysis was also run using 100 vol. % methane to see if any significant difference in flame temperature and/or gas species concentration is predicted.

The adiabatic flame temperature as a function of equivalence ratio is shown in Fig. 8 for combustion in air, and oxygen enriched air. The calculation was performed with both the fuel and oxidizer at 77 °F (298.15 K) and 14.7 psi (1 atm). The benefit of using oxygen enriched air is evident from the higher adiabatic flame temperature. It is also observed that there is no significant difference in the predicted flame temperature for the natural gas composition in Table 1 and 100 vol. % methane.

It is important to recognize that the actual flame temperature will be cooler than the adiabatic flame temperature due to incomplete combustion, heat loss to the surrounding, dilutants entering the flame, etc. Figure 9 shows a first attempt to measure the flame temperature using a platinum bead thermocouple placed approximately 5.5 inches from the nozzle. Test articles are typically placed between 5 and 6 inches from the nozzle exit depending upon sample fixturing, orientation, etc. Platinum has a melting temperature of ∼3200 °F which limited the ability to obtain thermocouple measurements above an equivalence ratio of ∼0.25.

The gas species concentration (mol. %) as a function of equivalence ratio, Φ is given in Fig. 10. The primary species of interest is H₂O which is calculated to range from ∼10% to 40% depending on equivalence ratio. For comparison, the H₂O concentration in the jet-fueled burner rigs is ∼10% or less. Similarly, no significant difference in predicted gas species concentration for natural gas and 100 vol. % methane is observed. The results of the equilibrium analysis provide a first insight into the combustion environment over a range of equivalence ratios. However, in application, chemical kinetics are required to understand the rates of reactions and how the gas species concentration, temperature, and velocity changes with time. This will be briefly introduced and discussed in Sec. 2.3.

Chemical equilibrium analysis is valid for the limit of infinitely fast reaction rates, or in other words, it is a thermodynamic limit. In actuality, the combustion process is more complex, and the process is governed by finite rates. To develop a better understanding of the combustion process, simulations are being performed using chemical kinetics mechanisms. For simplicity, and based on the results from CEA, 100 vol. % methane is used to approximate natural gas, and appropriate kinetics mechanisms for methane–oxygen combustion are being investigated and utilized.

OpenNCC (a publicly releasable version of the National Combustion Code [10–12]) and FUN3D (NASA Langley's CFD Solver [13]), are used to perform the simulations. Figure 11 shows one quarter of the 702 (¾ inch) burner nozzle that is modeled with symmetry conditions. All initial experiments in the rig have utilized the ¾ inch nozzle. Future work will be required to investigate the use of the 705 (1 inch) and 707 (1½ inches) nozzles.

Figure 12 shows the overall geometry of the computational domain. A region with a high mesh density exits between the nozzle exit and approximately 0.5 m from the nozzle exit to capture mixing of the fuel, oxidizer, and surrounding air. Figure 13 displays an example of the information extracted from the simulations showing gas temperature, gas velocity, and H₂O mole fraction as a function of distance from the nozzle exit. Future computational work will focus on incorporating an EBC/CMC sample geometry for conjugate heat transfer studies which will be coupled with experimental efforts.

An initial measurement of the heat flux in the NG/O₂ burner rig was conducted with a 64-Series Gardon Gauge water-cooled heat flux sensor (MEDTHERM Corporation, Huntsville, AL) as shown in Fig. 14(a). The sensor was positioned normal (90 deg) to the flame at an initial stand-off distance of 5.5 inches from the nozzle as displayed in Fig. 14(b). Three heat flux curves were generated by holding the oxygen constant at three different flows: 470 (minimum oxygen flow), 600, and 700 SCFH. The natural gas flow was then increased from a starting value of 25 SCFH (the minimum flow used for ignition). The results are shown in Fig. 15(a) for the three oxygen flow rates as a function of natural gas flow, and in Fig. 15(b) as a function of fuel/air (F/A) ratio, and equivalence ratio, Φ.

The measured heat flux increases with increasing natural gas flow (with constant oxygen flow) until a maximum is reached, and then begins to decrease with a further increase in natural gas flow. When plotted as a function of F/A ratio, or equivalence ratio, Φ, the three curves coincide with one another showing a primary dependence of heat flux on the F/A ratio (or similarly, equivalence ratio, Φ) where the maximum occurs at slightly fuel-rich mixtures, Φ ≈ 1.1. The measured heat flux ranges from ∼30 to 210 W/cm². For comparison, the peak heat flux measured using the same sensor in the jet-fueled burner rigs is ∼100 W/cm² which highlights the higher heat flux capability of the NG/O₂ burner rig.

It should be noted that care must be taken when interpreting the heat flux results in Fig. 15. The Gardon Gauge sensor is actively cooled using water, so the measured values represent a cold-wall heat flux. In addition, the NG/O₂ burner rig environment is mixed mode, consisting of convective and radiative heat transfer. The measurement can be highly sensitive to local convection, and while water-cooled sensors such as the Gardon Gauge can be successfully implemented in mixed mode environments, corrections may be required [14,15]. Nonetheless, these measurements serve as an initial assessment, and provide a relative comparison to the jet-fueled burner rigs.

To combat this phenomenon, EBCs were developed [17].

Due to the importance of understanding material recession, particularly for gas turbine materials such as SiC/SiC CMCs, an initial experiment on monolithic SiC (Hexoloy^® SA, Saint-Gobain, Worcester, MA) was performed to assess the capability of the NG/O₂ burner rig for oxidation and recession studies. This initial set of experiments is also used as an initial evaluation to see how the NG/O₂ burner rig compares to other laboratory facilities.

The dimensions of the SiC bars used for the oxidation and recession tests were 152.4 mm (6 inches) × 25.4 mm (1 inches) × 6.35 mm (0.25 inches) (l × w × h). The bars were held in place using a vice-clamp fixture with a nozzle exit to sample distance of 139.7 mm or 5.5 inches. An illustration of the test setup is shown in Fig. 16. Five samples were tested at the leading edge (LE) temperatures and equivalence ratios listed in Table 2. Note that one sample, sample 2, was pre-oxidized prior to testing in the rig. Since the stand-off distance was kept constant for every sample, different temperatures were achieved by adjusting the equivalence ratio, Φ. The LE temperature was measured using a Williamson Corporation (Concord, MA) ratio pyrometer. A second ratio pyrometer was aimed at the intersection of the midplane of the sample width and the midplane of the flame as denoted by the crosshairs in the inset of Fig. 16. A FLIR A6798sc thermal imaging camera was used to understand the thermal gradients within the sample during testing.

All samples were tested using the 702 (¾ inch) nozzle and an example of a SiC bar during a test is shown in Fig. 17. The specific weight change for the five SiC samples are shown in Fig. 18 and the linear volatilization rates are given in Table 3. The specific weight change was calculated using a hot zone equivalent to the nozzle diameter (1.905 cm or ¾ inch). Material recession is assumed to primarily occur within this region of high velocity combustion gases; however, there is some expansion of the flame after the nozzle. The effect of this expansion will need to be further understood using the coupled experimental and computational efforts.

Figure 19 shows a comparison of the linear volatilization rates obtained for SiC in the NG/O₂ rig to those obtained in other laboratory scale facilities: thermogravimetric analysis [18,19], jet-fueled burner rig (Mach 0.3) [20], and high-pressure burner rig (HPBR) [21]. Due to the presence of thermal gradients, and the fact that linear volatilization is temperature dependent, an average temperature of the sample within the flame region (1.905 cm × 2.54 cm) was determined using the IR thermal imaging. As such, two curves are shown for the NG/O₂ rig data where the linear volatilization rates are plotted as a function of LE temperature, and the average temperature in the flame region.

In general, the average temperature was approximately 100–150 °F lower than the LE temperature. Based on Fig. 19, it is observed that the overall volatilization rates obtained in the NG/O₂ rig lie somewhere between those obtained in the atmospheric Mach 0.3 burner rig and the HPBR. This highlights that the NG/O₂ rig can be a reasonable screening tool for understanding recession in a laboratory combustion environment prior to testing in more expensive engine rig tests. The SiC oxidation–recession tests also serve as a calibration or baseline dataset toward understanding the performance of the rig.

where Q is the activation energy and R is the universal gas constant [21]. Future work will couple the experimental oxidation–recession results with the combustion analysis and simulations to better understand the volatilization rate based on the temperature, gas velocity, and partial pressure of H₂O. Additional computational methods should also be explored in the future to better understand how oxidation and recession will vary in the presence of a thermal gradient, but this initial analysis is expected to give a first approximation of the SiC oxidation–recession in the NG/O₂ rig.

The CST project at NASA is looking to develop tools, technologies, and knowledge that will help overcome the technical barriers to practical commercial supersonic flight such as sonic boom, fuel efficiency, high-altitude emissions, and airport community noise [22]. EBC/CMC technology has the potential to aid supersonic transport through increased fuel efficiency and high-altitude emission reduction due to their low density (∼⅓ the density of Ni-based superalloys), and high temperature capability that can lead to reduced cooling requirements. Recent work on the emissions at high power conditions applicable to supersonic cruise showed that changes to the fraction of combustor air required for liner cooling significantly affect NOx emissions [23]. Air used to cool the combustor liner is not available to mix with the fuel which leads to increased flame zone fuel–air ratio that leads to increased NOx emissions. As such, the ability to reduce liner cooling due to the use of higher temperature capable materials such as EBCs and CMCs will result in reduced emissions.

Under the CST project, EBC/CMC systems are being investigated to understand the upper use temperature of the system without cooling air, with back side cooling, and effusive cooling techniques. Preliminary work in the uncooled configuration has been performed on a melt-infiltrated (MI) SiC/SiC CMC. The MI SiC/SiC CMC is protected by an EBC consisting of a Si bond coat and an ytterbium disilicate (Yb₂Si₂O₇) topcoat. A 3 inch × 3 inch sample was tested in the NG/O₂ rig at an EBC surface temperature of 2700 °F (1482 °C). An illustration of the test setup is shown in Fig. 20. The overall test setup for the 3 inch × 3 inch panel was similar to the SiC oxidation–recession testing but utilizes a different sample fixture. Based on knowledge of the thermal conductivity of the system, it was estimated that the thermal gradient through the Yb₂Si₂O₇ topcoat and Si bond coat was approximately 300 °F leading to a CMC interface temperature of 2400 °F which is the upper use temperature of the MI CMC.

Two EBC/CMC samples were tested: (1) to 25 h and (2) to 250 h with periodic shutdowns for nozzle cleaning and rig inspections. The samples were then destructively evaluated to characterize the thermally grown oxide (TGO) thickness since TGO growth will be life-limiting to the EBC system [24,25]. It is worth noting that the initial purpose of this study is to look at the temperature capability and durability of the EBC/CMC under relevant combustion conditions. Future work will be required to focus on thermomechanical aspects.

Figure 21 shows polished cross section of the 25-h and 250-h samples taken at the center of the sample in the hot zone. The average TGO thickness for the 25-h sample was 1.3 μm and the average TGO thickness for the 250 h sample was 4.3 μm. This corresponds to a parabolic oxidation growth rate, kp of 0.074 μm²/hr. Qualitative assessments of this EBC show that the critical TGO thickness resulting in spallation is somewhere between 20 and 30 μm. Assuming 20 μm as the limit, the life of the EBC due to oxidation would be reached after approximately 5400 h under these conditions: 2700 °F (1482 °C) EBC surface, 2400 °F (1316 °C) CMC interface temperature, with no backside cooling (free convection). It must be noted that the synergy between other environmental/failure mechanisms can result in shorter lifetime.

In the 250-h sample, a crack initiated through the EBC as shown in Fig. 22. Based on IR thermal imaging, the crack initiated approximately 18 h into the test and propagated down to the Si bond coat–CMC substrate interface. The crack was located slightly above the center of the sample but still within the hot zone. The difference in TGO thickness under the EBC and within the crack is evident. The average TGO thickness on the exposed CMC substrate in the cracked region was 8.8 μm which was approximately twice the thickness of the TGO under the EBC in an uncracked region of the sample. This highlights the reduction of the TGO growth rate under the EBC and shows the importance of maintaining the integrity of the EBC as defects such as cracks can lead to deleterious effects.

Investigations on NASA-developed modified Yb₂Si₂O₇-based EBCs to improve oxidation life are underway [25]. Previous work showed that modified Yb₂Si₂O₇-based EBCs tested under steam cycling, 90% H₂O/10% O₂ at 2400 °F (1316 °C) showed approximately 20× improvement in oxidation life compared to the baseline Yb₂Si₂O₇ EBC [25]. Slowing the TGO will be critical for the durability and life requirements for commercial supersonic transport.

The HyTEC project aims to develop small core turbofan engine technologies to enable fuel burn reductions, additional use of electric airplane systems through power extraction, and to advance engine operability and compatibility with sustainable aviation fuels [26]. Technology advancement under NASA's HyTEC project will provide environmental and cost benefits for the next generation of single-aisle class aircraft. The technology portfolio for HyTEC includes High Pressure Compressor, Enhanced Combustor Materials, High-Temperature Turbine Materials, Advanced High Pressure Turbine Aerodynamics, and Turbofan Power Extraction [26]. Within the materials portfolio, CMC combustor liners are being developed to increase performance and durability, and to reduce cooling and emissions. In addition, CMCs and EBCs capable of withstanding higher temperatures, up to 2700 °F (1482 °C) CMC and 3000 °F (1649 °C) EBC are being developed for turbine blades and vanes. In Phase I of the HyTEC project, CMC and EBC technology will be tested in laboratory-relevant environments and advanced to a technology readiness level (TRL) of 4 or 5, and then further advanced to TRL 6 in Phase II through integration of the technologies in an engine core demonstrator.

The NG/O₂ burner rig is being utilized for the high temperature EBC/CMC development with a target EBC surface temperature of 3000 °F (1649 °C) and an EBC/CMC interface temperature of 2700 °F (1482 °C). The interface temperature is controlled by adjusting the backside cooling air flowrate and fuel/air ratio to achieve the desired thermal gradient. NASA oxide-based high temperature EBCs have been developed to enable temperatures above the melting point of the Si bond coat (1414 °C) used in the baseline and modified Yb₂Si₂O₇-based EBCs [27]. The high temperature EBCs have been demonstrated to a TRL of 3 using steam cycling and laser thermal gradient testing. Technology advancement to TRL 4 is achievable through successful testing in the NG/O₂ combustion environment. Several high-temperature capable EBC architectures are being investigated within the HyTEC project. An example of the NASA-developed high temperature EBC during a test in the NG/O₂ rig is shown in Fig. 23. These tests are being performed on 3 inch × 1 inch samples with two spacers for insulation to minimize lateral heat loss with the flame impinging on the center of the sample. The target duration of a given test is 250 h. Tests are run continuously with periodic shutdowns to clean the nozzle of carbon build-up and rig inspections. Successful architectures that survive the 250 h in the NG/O₂ burner rig will be further developed and tested in a newly designed materials test section in NASA's CE-5 Combustor Test Stand to achieve TRL 5. The CE-5 test stand will expose the developed EBC/CMC technology to relevant temperatures (3000 °F/1649 °C) and pressures (up to 400 psig/∼27 atm) that can be observed in gas-turbine environments.

Figure 24 shows the newly designed materials test section for the CE-5 test stand. The new test section is modular and has dimensions of 3.5 inches in width, 3.5 inches in height, and 5.8 inches in length. The test section can accommodate a wide range of samples from flat plates to more complex airfoil geometries. The samples can be actively cooled, and various ports have been integrated for instrumentation such as thermocouples, pyrometers, and IR thermal imaging cameras.

A new atmospheric burner rig facility using natural gas and 93% oxygen was developed to test and characterize current and next generation gas-turbine materials. The facility is cost-effective and is an efficient means of screening advanced materials in a high temperature combustion environment. The facility was designed to run short term screening tests, and long-term durability tests with the ability to run continuously (24/7). Coupled experimental and computational analyses are utilized to characterize and develop a fundamental understanding of the rig. The rig also bridges a gap between other laboratory methods such as: furnaces, high-heat flux lasers, atmospheric jet-fueled burner rigs, and more expensive engine rig testing. The NG/O₂ burner rig is currently utilized by NASA's Aeronautics Research Mission Directorate (ARMD) projects to mature EBC/CMC technology with a focus on next generation, high temperature EBC (3000 °F/1649 °C) and CMC (2700 °F/1482 °C) development.

• Hybrid Thermally Efficient Core (HyTEC) Project Glenn Research Center (Funder ID: 10.13039/100006193).

• Commercial Supersonic Technology (CST) Project Glenn Research Center (Funder ID: 10.13039/100006193).

• Transformational Tools and Technologies (TTT) Project Glenn Research Center (Funder ID: 10.13039/100006193).

The authors attest that all data for this study are included in the paper.