PUBLISHED QUARTERLY BY THE AMERICAN SOCIETY OF MECHANICAL ENGINEERS • OCTOBER 2000
`
`1
`
`UTC 2020
`General Electric v. United Technologies
`IPR2016-01289
`
`

`

`Journal of
`Engineering for Gas
`Turbines and Pourer
`
`Published Quarterly by The American Society of Mechanical Engineers
`
`VOLUME 122 • NUMBER 4 • OCTOBER 2000
`
`556
`
`505 Preface: Special ICE Section
`TECHNICAL PAPERS-SPECIAL ICE SECTION
`Engine Design
`506 Review of Power Cylinder Friction for Diesel Engines
`D. E. Richardson
`520 Predicted Effects of Cylinder Kit Wear on Blowby and Oil Consumption
`for Two Diesel Engines
`D. E. Richardson and S. A. Krause
`526 A Probabilistic Approach to Engine Balance
`Dinu Taraza
`Intake and Exhaust System Dynamics
`533 Validation of a New TVD Scheme Against Measured Pressure Waves in
`the Inlet and Exhaust System of a Single Cylinder Engine
`M. Vandevoorde, R. Sierens, and E. Dick
`541 Comparison of Algorithms for Unsteady Flow Calculations in Inlet and
`Exhaust Systems of IC Engines
`M. Vandevoorde, J. Vierendeels, R. Sierens, E. Dick, and R. Baert
`549 Two-Dimensional Simulation of Wave Propagation in a Three-Pipe
`Junction
`R. J. Pearson, M. D. Bassett, P. Batten, and D. E. Winterbone
`Influence of the Exhaust System Design on Scavenging Characteristic
`and Emissions of a Four-Cylinder Supercharged Engine
`Ferdinand Trenc, Francisek Bizjan, Brane Sirok, and Ales Hribernik
`562 Performance Simulation of Sequentially Turbocharged Marine Diesel
`Engines With Applications to Compressor Surge
`Pascal Chesse, Jean-Franfois Hetet, Xavier Tauzia, Philippe Roy,
`and Bahadir lnozu
`In-Cylinder Processes
`570
`Intake Flow Structure and Swirl Generation in a Four-Valve Heavy-Duty
`Diesel Engine
`Kern Y. Kang and Rolf D. Reitz
`579 The Separation Between Turbulence and Mean Flow in ICE LDV Data: The
`Complementary Point-of-View of Different Investigation Tools
`Mario Amelio, Sergio Bova, and Carmine De Bartolo
`588 Comparisons of Diesel Spray Liquid Penetration and Vapor Fuel
`Distributions With In-Cylinder Optical Measurements
`Laura M. Ricart, Rolf D. Reitz, and John E. Dec
`596 Relationship Between Visible Spray Observations and DI Diesel Engine
`Performance
`Takashi Watanabe, Susumu Daidoji, and Keshav S. Varde
`Alternative Fuels Combustion and Emissions
`603 Formaldehyde Formation in Large Bore Natural Gas Engines Part 1:
`Formation Mechanisms
`Charles E. Mitchell and Daniel 8. Olsen
`
`(Contents continued on inside back cover)
`
`This journal is printed on acid-free paper, which exceeds the ANSI 239.48-
`1992 specification for permanence of paper and library materials. @?"'
`@ 85% recycled content, including 10% post-consumer fibers.
`
`2
`
`

`

`In recent years, the ever-increasing socioeconomic pressures for developing cleaner and more efficient means for
`converting chemical fuel energy into useful mechanical work have resulted in a dramatic growth in internal
`combustion engine research and development activities. As a natural outcome of this vigorous activity, a large
`number of technical papers dealing with various mechanical aspects of engine design, as well as with the basic
`thermo-fluid engine processes have been presented at the Fall and Spring Technical Conferences organized by the
`Internal Combustion Engine (ICE) Division of ASME. In 1999, the ICE Division has renewed its commitment to
`identify quality papers of long-term reference value for the journal. Therefore, a major section of this issue of the
`ASME JOURNAL FOR GAS TURBINES AND POWER is devoted to a selection of the highest quality ICE
`Papers presented at those meetings, and other papers of long-term reference value which were submitted directly to
`the journal. The contributions of the ICE Division Associates, the support of the ICE Executive Committee, and the
`ASME Editorial Staff, and above all the contributions of the authors and referees are gratefully acknowledged.
`The papers that have been included in the special ICE section of this volume have been arranged in four topical
`areas: engine design, intake and exhaust system dynamics, in-cylinder processes, and alternative fuels combustion
`and emissions. As a special note, the lead paper for the ICE section, authored by Dan E. Richardson and entitled
`''Review of Power Cylinder Friction for Diesel Engines,'' was presented at the 1999 Spring Technical Conference,
`Columbus, Indiana. This paper was selected as the most valuable technical paper presented at an ASME-ICE
`Division Meeting during 1999.
`
`Dennis N. Assanis
`Associate Editor
`ICE Division
`
`Journal of Engineering for Gas Turbines and Power
`Copyright © 2000 by ASME
`
`OCTOBER 2000, Vol. 122 / 505
`
`3
`
`

`

`Kang N. Lee1
`Chemical Engineering Department,
`Cleveland State University,
`Cleveland, OH 44115
`
`Key Durability Issues With
`Mullite-Based Environmental
`Barrier Coatings for Si-Based
`Ceramics
`
`Plasma-sprayed mullite (3 A 120 3 • 2 S iO 2) and mu:l~te!yttria-stabilize1-zirconia (:SZ)
`dual layer coatings have been developed to protect szlzcon-based ~e_ran:,zcs.Jrom envzro~(cid:173)
`mental attack. Mullite-based coating systems show excellent durability zn azr. However, m
`combustion environments, corrosive species such as molten salt or water vapor penetrate
`through cracks in the coating and attack the Si-based ceramics ~long th~ interface. Thus
`the modification of the coating system for enhanced crack-reszstan~e. zs _necess~ry for
`long-term durability in combustion environments. Q_ther key d~rabzlzty zs~ue~ include
`interfacial contamination and coating/substrate bonding. lnterfaczal contammatzon leads
`to enhanced oxidation and interfacial pore formation, while a weak coating/substrate
`bonding leads to rapid attack of the interface by corrosive species, both of which can
`cause a premature failure of the coating. Interfacial contamination can be minimized by
`limiting impurities in coating and substrate materials. The interface may be modified to
`improve the coating/substrate bond. [S0742-4795(00)03203-8]
`
`Introduction
`Silicon-based ceramics are promising candidates for hot section
`structural components of heat engines and heat exchanger tubes
`for industrial furnaces. One potential barrier to such applications
`is their environmental durability. The excellent oxidation resis(cid:173)
`tance of silicon-based ceramics at high temperatures in clean, dry
`oxygen is due to the formation of a solid, protective external silica
`scale. However, the normally protective silica scale can be de(cid:173)
`graded by reacting with impurities, such as alkali salts [1,2] or
`water vapor [3-5].
`Molten Na2S04 can deposit in gas turbine engines operating
`near marine environments or from contaminants in the fuel [ 6].
`The Na2S04 then reacts with the silica to form liquid sodium
`silicate, leading to accelerated degradation of Si-based ceramics
`[1]. In coal-fired combustion environments, combustion gases
`contain low levels of alkali salts because of naturally occurring
`minor alkali components in the coal. These alkali salts can dis(cid:173)
`solve in the silica scale and enhance the transport of oxygen,
`leading to drastically enhanced oxidation [2]. In heat engines, sub(cid:173)
`stantial amounts of water vapor, typically about 10 percent, is
`produced from burning hydrocarbon fuels in air [7]. The water
`vapor reacts with silica, forming gaseous hydroxide species, such
`as Si(OH) 4 [3-5]. In high-pressure combustion environments, the
`higher water vapor pressure generates even higher levels of gas(cid:173)
`eous hydroxide species, resulting in linear volatilization of silica.
`The linear volatilization of silica coupled with the parabolic oxi(cid:173)
`dation of Si-based ceramics results in overall paralinear kinetics
`[ 4], causing rapid degradation of Si-based ceramics. Therefore,
`the realization of the full potential of silicon-based ceramics in
`high temperature structural applications depends on the develop(cid:173)
`ment of environmental protection schemes.
`An external environmental barrier coating is a promising ap-
`
`1Resident Research Associate at NASA Glenn Research Center.
`Contributed by the International Gas Turbine Institute (IGTI) of THE AMERICAN
`SOCIETY OF MECHANICAL ENGINEERS for publication in the ASME JOURNAL OF
`ENGINEERING FOR GAS TURBINES AND POWER. Paper presented at the Interna(cid:173)
`tional Gas Turbine and Aeroengine Congress and Exhibition, Indianapolis, IN, June
`7-10, 1999; ASME Paper 99-GT-443. Manuscript received by !GT! March 9, 1999;
`final revision received by the ASME Headquarters May 15, 2000. Associate Tech(cid:173)
`nical Editor: D. Wisler.
`
`proach to protect Si-based ceramics from environmental attack.
`Mullite is a promising candidate coating material because of its
`close coefficient of thermal expansion (CTE) match and good
`chemical compatibility with Si-based ceramics. Researchers at So(cid:173)
`lar Turbines, Inc., San Diego, CA, and Oak Ridge National Labo(cid:173)
`ratory, Oak Ridge, TN, have done pioneering work on applying
`refractory oxide coatings such as alumina, zirconia, yttria, mullite,
`cordierite, etc., on SiC [8,9]. In those studies, mullite was found to
`be adherent and offer the best protection of all the refractory coat(cid:173)
`ings tested. However, those plasma-sprayed mullite coatings
`tended to crack on thermal cycling. Researchers at NASA Glenn
`Research Center, Cleveland, OH, identified the crystallization of
`amorphous phase mullite, which accompanies a volumetric con(cid:173)
`traction, as the main source for the cracking of plasma-sprayed
`mullite coatings [10]. Based on this finding, researchers at NASA
`GRC successfully eliminated most of the amorphous phase mul(cid:173)
`lite from the coating by spraying the mullite while heating the SiC
`substrate above the crystallization temperature of amorphous mul(cid:173)
`lite (-1000 °C) [10].
`These second-generation mullite coatings provide excellent
`protection in air and molten salt environment [10-14]. Mullite
`coatings, however, suffer selective vaporization of silica in the
`presence of water vapor because of its high silica activity
`(0.3-0.4) [11,15,16]. Thus, an environmental overlay coating is
`required when protection from water vapor is needed. Yttria(cid:173)
`stabilized zirconia (YSZ) was selected as a baseline overlay coat(cid:173)
`ing because of its proven performance as a thermal barrier coating
`(TBC) in combustion environments. The mullite coating in the
`mullite/YSZ coating system is somewhat analogous to the bond
`coat in conventional TBC's, in the sense that it provides bonding
`as well as oxidation protection. This paper will discuss the current
`durability issues of second generation mullite-based environmen(cid:173)
`tal coatings on Si-based ceramics and future research directions in
`this area.
`
`Experimental
`Mullite and YSZ coatings were applied by atmospheric pres(cid:173)
`sure plasma spraying onto 2.5X0.6X0.15 cm sintered a-SiC cou(cid:173)
`pons (Hexoloy™, Carborundum, Niagara Falls, NY) and reaction
`bonded silicon nitride (RBSN R. Bhatt, NASA GRC). The SiC
`
`632 / Vol. 122, OCTOBER 2000
`
`Copyright © 2000 by ASME
`
`Transactions of the ASME
`
`4
`
`

`

`substrates were roughened (Ra 2 =5-6 µm) by etching in Na2C03
`to achieve a good mechanical bond [10], whereas RBSN was used
`as processed. Fused mullite powder with the particle size of
`44-74 µm was used (Cerac, Inc., Milwaukee, WI). Typical coat(cid:173)
`ing thickness was 100-200 µm for the mullite coating and 50 µm
`for the YSZ coating. Details of the coating parameters are de(cid:173)
`scribed elsewhere [IO].
`Coated coupons were annealed in air at 1300 °C for 100 h, prior
`to the environmental exposure. Environmental exposures were
`thermal cycling in air, thermal cycling in 90 percent H20/02 at 1
`atm (simulated lean combustion environments) or isothermal ex(cid:173)
`posure in high-pressure burner rigs with or without molten salt.
`Thermal cycling tests in water vapor were to evaluate the long(cid:173)
`term behavior of coatings in lean combustion environments be(cid:173)
`cause high-pressure burner rigs are not suitable for long-term tests
`due to their high operating cost. Thermal cycling was performed
`using an automated thermal cycling furnace. Each thermal cycle
`consisted of 2 h at temperature, rapid cooling to room tempera(cid:173)
`ture, and 20 min at room temperature. Typically, samples reached
`the high temperature within 2 min and the low temperature within
`5 min in each cycle. Molten salt environments were generated
`using a high-pressure burner rig with Jet A fuel containing 2 ppm
`Na at 4 atm.
`Tested samples were mounted in epoxy, polished to 1 µmusing
`diamond suspension, and examined using Scanning Electron Mi(cid:173)
`croscopy (SEM) and Energy Dispersive Spectroscopy (EDS).
`
`Environmental Durability
`
`Air. Mullite/YSZ-coated SiC was exposed to a 2 h thermal
`cycling exposure in air at 1300 °C. Figure 1 shows the cross(cid:173)
`section after 500h exposure. Mullite coatings typically developed
`through-thickness cracks, however, they maintained excellent ad(cid:173)
`hesion and provided excellent oxidation protection. Mullite(cid:173)
`coated SiC exhibited a similar behavior under the same exposure,
`indicating that the presence of YSZ overlay coating did not affect
`the coating durability despite the large CTE mismatch between
`the two layers.
`Mullite/YSZ-coated RBSN was exposed to an isothermal oxi(cid:173)
`dation at 1300 °C in air for 50 h. A thick oxide scale and large
`pores developed at the mullite/RBSN interface (Fig. 2). EDS
`showed a significant amount of Mg in the scale. This contamina(cid:173)
`tion by MgO, which is from the RBSN, is responsible for the
`enhanced oxidation and pore formation. Similar enhanced oxida(cid:173)
`tion and pore formation of mullite-coated SiC was observed when
`the system was contaminated by Na20 or K20 from the coating
`processed with a low purity mullite powder [17]. Alkali and alka(cid:173)
`line earth metal oxides are known to degrade the oxidation resis-
`
`2 Average distance from the roughness profile to the mean line
`
`Fig. 2 Mullite/YSZ-coated RBSN after 50 hat 1300 °c in air
`
`tance of Si-based ceramics by enhancing the oxygen transport
`through silica by altering the silica network [18].
`
`Combustion Environments
`
`High Pressure Burner Rig. Uncoated, mullite-coated, and
`mullite/YSZ-coated SiC was exposed to high-pressure burner rig
`(HPBR) under a rich burn condition (equivalence ratio= 1.9) at 6
`atm and 1230 °C. Figure 3 shows the plot of weight change vs.
`time. Uncoated and mullite-coated SiC showed weight loss due to
`the volatilization of silica. The lack of weight change in the
`mullite/YSZ-coated SiC indicated that the YSZ overlay coating
`provided the protection from water vapor. Figure 4(a) shows the
`cross-section of mullite-coated SiC after the high-pressure burner
`rig exposure. Pores are observed at the interface where cracks
`intersected the SiC interface. Enhanced oxidation was observed
`around pores, indicating that water vapor penetrated through the
`cracks and reacted with SiC. The pore formation is attributed to
`the generation of gaseous silicon hydroxide species. The selective
`volatilization of silica from mullite left a porous layer of alumina
`on the surface of mullite (Fig; 4(b)).
`
`Water Vapor Cyclic Furnace. Mullite-coated SiC/SiC (Du(cid:173)
`pont Lanxide, Newark, DE) exposed to 2 h cycle exposure in 50
`percent H20/02 at 1300 °C showed weight loss, whereas mullite/
`YSZ-coated SiC/SiC exhibited parabolic oxidation [19]. The
`weight loss of mullite-coated SiC/SiC was attributed to the selec(cid:173)
`tive volatilization of silica from the mullite. This result is consis(cid:173)
`tent with the weight change behavior of mullite and mullite/YSZ
`coatings in high pressure burner rig (Fig. 3). A similar weight
`
`UicoatedSiC
`
`I
`
`lime (h)
`
`Fig. 1 Mullite/YSZ-coated SiC after 500 h with 2 h cycles at
`1300 °C in air
`
`Fig. 3 Weight change versus time for coated and uncoated
`SiC in HPBR (6 atm, 1230 °C)
`
`Journal of Engineering for Gas Turbines and Power
`
`OCTOBER 2000, Vol. 122 / 633
`
`5
`
`

`

`Fig. 4 Mullite-coated SiC after 50 h in HPBR (6 atm, 1230 °C)
`
`change behavior is expected for these coatings on sintered SiC in
`water vapor cyclic furnace because the reaction at the surface
`should not be affected by the type of substrates.
`Mullite/YSZ-coated SiC was exposed to 2 h cycle exposure in
`90 percent H20/02 at 1300 °C to evaluate the long-term durability
`in lean combustion environments. Most interfacial areas showed
`excellent adherence with limited oxidation after 100 h. However,
`at some interfacial areas, where through-thickness-cracks inter(cid:173)
`sected the SiC interface, accelerated oxidation initiated (Fig.
`5(a)). After 200 h, accelerated oxidation propagated along the
`entire mullite/SiC interface, forming a thick porous silica scale
`(Fig. 5(b)). Water vapor, the predominant oxidant in a H20/02
`environment, is known to enhance the oxidation of SiC. The silica
`scale formed in high water vapor is porous, allowing the oxidation
`to propagate readily along the interface. The porous scale is at-
`
`tributed to the generation of gaseous silicon hydroxide species.
`This is in contrast to the scale formed in dry air where the scale is
`dense and thus prevents the rapid propagation of oxidation.
`The effect of preoxidation on the coating adherence was evalu(cid:173)
`ated by preoxidizing a SiC coupon at 1300 °C for 100 h in air,
`prior to the application of mullite/YSZ coating. The coated cou(cid:173)
`pon was exposed to 90 percent H20/02 at 1250 °C with 2 h
`cycles. The cross-section after 100 h showed that the entire inter(cid:173)
`face was attacked by water vapor, forming a thick porous silica
`scale (Fig. 6). It is believed that the silica scale from preoxidation
`weakened the mullite/SiC bonding, leading to more rapid penetra(cid:173)
`tion by water vapor than in the coupon without preoxidation. Pre(cid:173)
`oxidation did not affect the durability of the system when exposed
`in air.
`
`Fig. 5 Mullite/YSZ-coated SiC exposed to 2 h cycle exposure
`in 90 percent H20/02 at 1300 °C; (a) 100 h; (b) 200 h.
`
`Fig. 6 Mullite/YSZ-coated SiC after 100 h with 2 h cycles in 90
`percent H20/02 at 1250 °C. The SiC coupon was oxidized for
`100 h in air at 1300 °C prior to the application of coating.
`
`634 I Vol. 122, OCTOBER 2000
`
`Transactions of the ASME
`
`20µm
`
`6
`
`

`

`Fig. 7 Mullite-coated SiC coupon after 50 h in hot corrosion rig at 1000 °c
`
`Hot Corrosion Rig. Uncoated and mullite-coated SiC coupons
`were exposed in a hot corrosion burner rig at 1000 °C for 50 h.
`Prior to the hot corrosion exposure, the coated coupon underwent
`600 one hour thermal cycles at 1200 °C in air to let cracks form.
`Uncoated SiC was severely deformed due to the attack by the
`molten salt, whereas mullite-coated SiC was well preserved (Fig.
`7(a)). Cross-section of the mullite-coated SiC showed that
`mullite/SiC interface was fairly intact (Fig. 7(b)). Only a limited
`attack with glassy reaction products, presumably sodium silicates,
`was observed at the interface where cracks intersected the SiC
`interface (Fig. 7(d)). This is an indication of the penetration of salt
`through some cracks.
`
`Key Issues
`Several key durability issues are identified from the environ(cid:173)
`mental durability test results. They include through-thickness(cid:173)
`cracking, bonding of mullite to the Si-based ceramic, and inter(cid:173)
`face contamination. These key issues will be discussed in this
`section to help elucidate the future research directions to improve
`the coating durability.
`
`Through-Thickness-Cracking. As-sprayed mullite coatings
`are free of macrocracks (crack width > 1 µm). However, on ther(cid:173)
`mal exposure, they develop macrocracks, the size of which can be
`as wide as 5-10 µm. It has been shown in the foregoing section
`that corrosive species, such as molten salt or water vapor, can
`penetrate through these cracks and attack the SiC, leading to ac(cid:173)
`celerated degradation of the system.
`It is believed that the development of through-thickness-cracks
`is due to stresses in the coating. The most likely source for
`stresses is the precipitation of various second phases in the mullite
`coating. Amorphous phase mullite precipitates in the coating due
`to the rapid cooling of molten mullite [10]. Even in the second
`generation mullite coating, it is likely that some residual amor(cid:173)
`phous phase mullite still remains in the coating. Volumetric
`shrinkage results during the crystallization of the residual amor(cid:173)
`phous phase mullite in subsequent thermal exposures. A signifi(cid:173)
`cant amount of alumina also precipitates in the plasma-sprayed
`mullite coating [10]. Plasma-sprayed alumina typically contains
`substantial amount of metastable alumina phases such as
`y-alumina [20]. Volumetric shrinkage results when the metastable
`alumina phases transform to stable a-alumina in subsequent ther(cid:173)
`mal exposures. The precipitation of alumina phase is also accom(cid:173)
`panied by the precipitation of silica-rich phases to maintain the
`chemical balance [10]. Both the alumina and silica-rich phases
`cause CTE mismatch stresses. Thus the precipitation of second
`phases in the mullite coating and the resulting volumetric shrink(cid:173)
`age and CTE mismatch are suggested to be the major sources for
`the stresses in the coating.
`
`Mullite/Substrate Bond. Mullite does not form a strong
`chemical bond with SiC according to our diffusion couple study
`[21]. Thus the mullite/SiC bond of as-sprayed coatings is mainly
`due to mechanical interlocking. The lack of chemical bond may
`be why the oxidation propagates readily along the interface in
`water vapor. Silica scale from preoxidation presumably further
`weakens the interfacial bond, leading to more rapid attack by the
`water vapor.
`
`Contamination.
`Interfacial contamination can degrade the
`coating durability by altering the physical or chemical properties
`of silica scale. Contaminants, such as alkali or alkaline earth metal
`oxides, are known to be most detrimental to the oxidation resis(cid:173)
`tance of Si-based ceramics [7]. They enhance the oxygen transport
`through silica by altering the silica network [18]. They also reduce
`the scale viscosity by forming silicates [22]. Pores develop at the
`interface when gases generated as a result of oxidation bubble
`through the low viscosity silica scale [17]. High interfacial poros(cid:173)
`ity can eventually lead to coating delamination. Contamination
`from coating materials can be minimized by using high purity
`mullite powder or by limiting the addition of alkali or alkaline
`earth metal oxides in the processing of Si-based ceramics [17].
`
`Future Research Directions
`
`Modification of Mullite Coating for Improved Crack Resis(cid:173)
`tance. Second phases that cause cracking may be reduced
`through process optimization. However, it may be impossible to
`completely eliminate all second phases. For example, melt grown
`mullite is always alumina-rich and thus some silica-rich phases
`will always be present to maintain the chemical balance [23,24].
`The free alumina phase is likely due to the incongruent melting of
`mullite and thus may not be completely eliminated by process
`optimization. Other approaches to improve the crack resistance
`include modifying the physical properties of mullite coating or
`sealing the cracks by applying an overlay coating with good crack
`resistance. Figure 8 shows mullite/cordierite-coated SiC after 600
`h with 20 h cycles at 1200 °C in air. Note that the crack stopped
`at
`the cordierite/mullite interface. Plasma-sprayed cordierite
`seems to be more resistant to cracking than plasma-sprayed
`mullite:
`
`Interface Modification for Improved Bonding. The mullite/
`Si-based ceramic interface may be modified to enhance the
`coating/substrate bonding. One example is Mo flash layer. A Mo
`flash layer (1-5 µm) was applied on SiC by sputtering and an(cid:173)
`nealed in Ar-5 percent H2 at 1200-1300 °C for 20- lOOh, prior to
`the application of mullite/YSZ coating. Silicon diffused into the
`molybdenum during the annealing, forming molybdenum silicide
`with varying composition through the thickness [16]. Similar re(cid:173)
`sults were reported in a Mo/SiC diffusion couple study [25,26].
`
`Journal of Engineering for Gas Turbines and Power
`
`OCTOBER 2000, Vol. 122 / 635
`
`7
`
`

`

`grateful to G.W. Leissler (Dynacs, NASA Glenn Group) for the
`preparation of plasma-sprayed coatings, M. D. Cuy (Dynacs,
`NASA Glenn Group) for the hot corrosion test, R. C. Robinson
`(Dynacs, NASA Glenn Group) for the HPBR test, and R. Bhatt
`(Army) for the supply of RBSN coupons
`
`References
`[l] Jacobson, N. S., Smialek, I. L., and Fox, D.S., 1990, "Molten Salt Corrosion
`of SiC and Si3N4 ," in Handbook of Ceramics and Composites, Vol. 1, Cher(cid:173)
`emisinoff, N. S., ed., Marcel Dekker, New York, pp. 99-135.
`[2] Pareek, V., and Shores, D. A., 1993, "Oxidation of Silicon Carbide in Envi(cid:173)
`ronments Containing Potassium Salt Vapor," J. Am. Ceram. Soc., 74, No. 3,
`pp. 556-563.
`[3] Hashimoto, A., 1992, "The Effect of H,O Gas on Volatilities of Planet(cid:173)
`Forming Major Elements: I-Experimental Determination of Thermodynamic
`Properties of Ca-, Al-, and Si-hydroxide Gas Molecules and Its Application to
`the Solar Nebula," Geochim. Cosmochim. Acta, 56, pp. 511-532.
`[4] Opila, E. J., and Hann, R., 1997, "Paralinear Oxidation of CVD SiC in Water
`Vapor," J. Am. Ceram. Soc., 80, No. 1, pp. 197-205.
`[5] Opila, E. J., Fox, D. S., and Jacobson, N. S., 1997, "Mass Spectrometric
`Identification of Si(OH) 4 from the Reaction of Silica and Water Vapor," J.
`Am. Ceram. Soc., 80, No. 4, pp 1009-1012.
`[6] Pettit, F. S., and Giggins, C. S., 1987, Superalloys II, C. T. Sims, N. S. Stoloff,
`and W. C. Hage, eds., Wiley, New York, p. 327.
`[7] Jacobson, N. S., 1993, "Corrosion of Silicon-Based Ceramics in Combustion
`Environments," J. Am. Ceram. Soc., 76, No. l, pp. 3-28.
`[8] Price, J. R., van Roode, M., and Stala, C., 1992, "Ceramic Oxide-Coated
`Silicon Carbide for High-Temperature Corrosive Environments," Key Eng.
`Mater., 72-74, pp. 71-84.
`[9] Federer, J. I., 1990, "Alumina Base Coatings for Protection of SiC Ceram(cid:173)
`ics," J. Mater. Eng.,.12, No. 2, pp. 141-149.
`[JO] Lee, K. N., Miller, R. A., and Jacobson, N. S., 1995, "New Generation of
`Plasma-Sprayed Mullite Coatings on Silicon-Carbide," J. Am. Ceram. Soc.,
`78, No. 3, pp:705-710.
`[11] Lee, K. N., Jacobson, N. S., and Miller, R. A., 1994, "Refractory Oxide
`Coatings on SiC Ceramics," MRS Bull., XIX, No. 10, pp. 35-38.
`[12] Lee, K. N., and Miller, R. A., 1996, "Oxidation Behavior of Mullite-Coated
`SiC and SiC/SiC Composites Under Thermal Cycling between Room Tem(cid:173)
`perature and 1200-1400 °C," J. Am. Ceram. Soc., 79, No. 3, pp. 620-626.
`[13] Lee, K. N., and Miller, R. A., 1996, "Development and Environmental Dura(cid:173)
`bility of Mullite and Mullite/YSZ Dual Layer Coatings for SiC and Si3N4
`Ceramics," Surf. Coat. Technol., 86-87, pp. 142-148.
`[14] Jacobson, N. S., Lee, K. N., and Yoshio, T. Y., 1996, "Corrosion of Mullite
`By Molten Salts," J. Am. Ceram. Soc., 79, No. 8, pp. 2161-2167.
`[15] Lee, K. N., Miller, R. A., Jacobson, N. S., and Opila, E. J., 1995, "Environ(cid:173)
`mental Durability of Mul!ite Coating/Sic and Mullite-YSZ Coating/SiC Sys(cid:173)
`tems," Ceram. Eng. Sci. Proc., September-October, pp. 1037-1044.
`[16] Lee, K. N., and Miller, R. A., 2000, "Durability of Mullite/YSZ-Coated SiC in
`90 percent H20/02 ," Adv. Ceram. Matrix Compos., IV, pp. 17-25.
`[17] Lee, K. N., 1998, "Contamination Effects on Interfacial Porosity During Cy(cid:173)
`clic Oxidation of Mullite-Coated SiC," J. Am. Ceram. Soc., 81, No. 12, pp.
`3329-3339.
`[18] Lamkin, M.A., Riley, F. L., and Fordham, R. L., 1992, "Oxygen Mobility in
`Silicon Dioxide and Silicate Glasses," J. Eur. Ceram. Soc., IO, pp. 347-367.
`[19] Lee, K. N., and Miller, R. A., 1997, "Modification of Mullite-Based Coatings
`on Si-Based Ceramics for Enhanced Durability," H/TEMP Review, III, NASA
`Lewis Research Center, Cleveland, OH, p. 45.
`[20] Heintz, G. N., and Uematsu, U., 1992, "Preparation and Structures of Plasma(cid:173)
`Sprayed y and a-alumina Coatings," Surf. Coat. Technol., SO, pp. 213-222.
`[21] Lee, K. N., 1998, NASA Glenn Research Center, unpublished research.
`[22] Holloway, D. G., 1973, The Physical Properties of Glass, Wykeham Publica(cid:173)
`tions, London, U.K.
`[23] Aksay, I. A., and Pask, J. A., 1975, "Stable and Metastable Equilibria in the
`System Si0rA1203 ," J. Am. Ceram. Soc., 58, Nos. 11-12, pp. 507-512.
`[24] Kriven, W. M., and Pask, J. A., 1983, "Solid Solution Range and Microstruc(cid:173)
`ture of Melt-Grown Mullite," J. Am. Ceram. Soc., 66, No. 9, pp. 649-654.
`[25] Martinelli, A. E., and Drew, R. A. L., 1995, "Microstructure Development
`During Diffusion Bonding of a-Silicon Carbide to Molybdenum," Mater. Sci.
`Eng., A, 191, pp. 239-247.
`[26] Martinelli, A. E., Drew, R. A. L., and Berriche, R., 1996, "Correlation Be(cid:173)
`tween the Strength of SiC-Mo Diffusion Couples and the Mechanical Proper(cid:173)
`ties of the Interfacial Reaction Products," J. Mater. Sci. Lett., IS, pp. 307-
`310.
`
`Fig. 8 Mullite/Cordierite-coated SiC after 600 h with 20 h
`cycles at 1200 °C in air
`
`Fig. 9 Mullite/Mo Silicide/YSZ-coated SiC after 500 h with 2 h
`cycles at 1300 °C in 90 percent H20/02
`
`The annealed flash layer was thicker and partially delaminated,
`presumably due to the volume expansion as the silicon diffused
`into the molybdenum. Figure 9 shows the mullite/molybdenum
`silicide/Sic system after 500 h at 1300 °C with 2 h cycles in 90
`percent H20/02 . Excellent durability was observed at some areas
`of the coating where the molybdenum silicide layer remained in(cid:173)
`tact. Even the silica scale at the interface attacked by water vapor
`was thinner than the scale at the unmodified interface, indicating
`that the interfacial modification delayed the water vapor attack.
`This result suggests the potential of interface modification for im(cid:173)
`proving the coating durabi)ity.
`
`Conclusions
`A mullite-based coating system is promising as an environmen(cid:173)
`tal barrier for Si-based ceramics in combustion environments. Key
`durability issues are through-thickness cracking in the mullite
`coating, a _weak mullite/Si-based ceramic bond, and interfacial
`contamination. Improvement of the crack resistance through the
`,F modification of mullite or the application of an overlay coating is
`suggested. Improved bonding may be achieved through the modi(cid:173)
`fication of interface. Interfacial contamination can be minimized
`by limiting impurities in coating and substrate materials.
`
`Acknowledgments
`The author would like to thank R.A. Miller (NASA-Glenn Re(cid:173)
`search Center) for many helpful discussions. The author is also
`
`636 I Vol. 122, OCTOBER 2000
`
`Transactions of the ASME
`
`8
`
`