Thin Solid Films 517 (2009) 5120–5129

Contents lists available at ScienceDirect

Thin Solid Films

j ourna l homepage: www.e lsev ie r.com/ locate / ts f

High emissivity coatings for high temperature application: Progress and prospect

Xiaodong He a,⁎, Yibin Li a, Lidong Wang a,b, Yue Sun a, Sam Zhang c

a Center for Composite Materials, School of Astronautics, Harbin Institute of Technology, P.O. Box 3010, Harbin 150080, PR Chinab School Materials Science and Engineering, Harbin Institute of Technology, P.O. Box 405, Harbin 150001, PR Chinac School of Mechanical and Aerospace Engineering, Nanyang Technological University, 50 Nanyang Avenue, Singapore 639798, Singapore

⁎ Correspondence author.E-mail address: hexd@hit.edu.cn (X. He).

0040-6090/$ – see front matter © 2009 Elsevier B.V. Adoi:10.1016/j.tsf.2009.03.175

a b s t r a c t

a r t i c l e i n f o

Available online 22 April 2009

Keywords:EmissivityCoatingsReview

High emissivity coatings are widely used in many cases where heat transfers through electromagneticradiation that arises due to the temperature of a body. Extensive theoretical and experimental efforts havebeen made to synthesize and investigate high emissivity coatings. The emissivity can be improved throughvarious or combined mechanisms. The characterization of the emissivity is still a fully open problem. In thispaper, we review the various mechanisms associated with the emissivity enhancement and emissivitycharacterization techniques. Based on these literature reviews, the prospect will be presented in theconcluding remarks.

© 2009 Elsevier B.V. All rights reserved.

1. Introduction

High emissivity coatings are widely used in many high tempera-ture applications to effectively transfer the heat by radiation [1–4]. Forexample, the surface of Ni-, Fe-, Co-based alloys applied in metalthermal protection systems (MTPS) [5,6] in third generation of reusedspace vehicles involves high friction heat from acute friction betweenspace vehicle surface and atmosphere, which causes obvious increaseof surface temperature up to 1000 °C during hypervelocity flight [7–9].As a result, the lifetime and performance of MTPS will be seriouslydegraded. Therefore, the high emissivity coating is intentionallydeposited on MTPS to decrease the surface temperature by radiation.

When a radiation falls on a body it may be partially reflected,transmitted, or absorbed, which is associated with reflection (R),absorption (A), transmission (T), and emission (ε). The principle ofconservation of energy ties together these radiation characteristics asA+R+T=1. According to Kirchhoff's law, at equilibrium for a givenwavelength λ and temperature T, the emissivity of any body is equal toits absorption [10], i.e. ε(ν, T, θ)=A(ν,T,θ), where ε is the emissivity, νthe frequency, and the angle θ specifies the observer angle. Thisindicates that the emissivity of an object can be indirectly obtained bymeasuring its absorptivity. For a uniform and isotropic opaque surface inthermal equilibrium, the transmission is zero, the relationship betweenthe emissivity and the reflectivity is: ε(ν,T, θ)=A(ν,T,θ)=1−R(ν,T,θ).Manymaterials canbe considered asopaquebody. The emissivity here ischaracterized to describe the surface radiative property which involvesthe transfer of heat by electromagnetic radiation arising due to thetemperature of a body. The emissivity is defined as the ratio of energy

ll rights reserved.

radiated by thematerial to energy radiated by a black body (A body thatemits the maximum amount of heat for its absolute temperature iscalled a black body, meaning that a blackbody completely absorbs allradiation incident upon it, and at the same time it emits all the energythat it absorbs with the same absorbing spectrum. That is to say, ε=1)at the same temperature, which is described by Eq. (1):

e Tð Þ =R∞0 e λð ÞEλdλ

σT4 ð1Þ

Where E is radiant heat of gray body to its surroundings, εhemispherical emissivity of the gray body (dimensionless), σ Stefan–Boltzman constant (5.67×10−8 W/(m2×K4)), T temperature (K). Theemittance is dependent on direction and wavelength, thus theradiance ability can also be evaluated by the spectral emissivity anddirectional emissivity, described as Eqs. (2) and (3) respectively:

eλ λ; Tð Þ = Eλ;actual emitted λ; Tð ÞEλ;black body λ; Tð Þ =

Eλ λ; Tð ÞEbλ λ; Tð Þ ð2Þ

eθ θ;/; Tð Þ =R∞0 Lλ;actual emitted λ; θ;/; Tð ÞdλR∞

0 Lλ;black body λ; Tð Þdλ =L θ; Tð ÞLb Tð Þ ð3Þ

A real object does not radiate asmuch as a perfect black body, whichradiates less heat than a black body and is called gray body (εb1).

Thermally emitted radiance from any surface mainly depends ontwo factors. (1) the surface temperature, which is an indication of theequilibrium thermodynamic state resulting from the energy balanceof the fluxes between the gray body surface and its surroundings; and(2) the surface emissivity, which is the efficiency of the surface fortransmitting the radiant energy generated in the surface into itssurroundings. The latter depends on the temperature (but therelationship between emissivity and temperature is not definite,

Fig. 2. The emissivity spectra at room temperature for 0.1%, 0.25% and 0.5% Cr2O3 dopedcompounds [17].

5121X. He et al. / Thin Solid Films 517 (2009) 5120–5129

depending on material nature, surface parameters and wavelength),composition, surface roughness, coating thickness, wavelength, andphysical parameters of the surface. Here we review recent progresswith emphasis on the discussion of the design methodology, followedby briefing the emissivity characterization. Lastly, the prospectsfor research and technology in high emissivity coatings will besummarized.

2. Design methodology

2.1. Doping

According to Wien's displacement law and Planck's law [11], mostenergy of the black body at high temperature is radiated in thewavelength range of 1–5 μm, however it happens that many materialshave the weak intrinsic absorption within this range. As a result, theemissivity in this region is quite low since the object's absorptivity andemissivity spectra are identical in thermal equilibrium state. Extensivework proves that total emissivity or spectral emissivity could beeffectively enhanced via doping (doping refers to the process ofintentionally introducing impurities into an extremely pure material).Several mechanisms for doping enhanced emissivity in the wave-length range of 1–5 μmwere proposed for different materials, namely,free carrier absorption mechanism (the phenomenon whereby anelectron within a band absorbs radiation by transferring from a low-energy level to an empty high-energy level) [12], d–d transitions inoctahedral coordination [13,14], distortion of the crystal lattice [15].For semiconductor material, free carrier absorption mechanismgenerally dominates the emissivity enhancement in the wavelengthregion of 1–5 μm. Fig. 1 shows the absorption coefficient as a functionof wavelength for three different SiC specimens with differentimpurity (dopant) concentrations [16]. It is obvious that increase inimpurity concentration strengthens the absorption in the wavelengthrange of 1–5 μm. As higher impurity concentration causes more freeelectrons, more electrons within a band absorb more radiation bytransferring from a low-energy level to an empty high-energy level,leading to more absorption. As such, the emissivity is accordinglyenhanced within this wavelength range as the impurity concentrationincreases since the object's absorptivity and emissivity spectra areidentical in thermal equilibrium state.

In glassy oxide coatings, different excitations are responsible forthe absorption of electromagnetic radiation [17]. It is suggested thatsuitable dopants with transition metal oxides diluted in silicateglasses may yield emissivity as high as 0.9 in the whole temperaturerange. For example, the d–d transitions of Cr3+ in octahedralcoordination play an important role on the absorption [13,14]. The

Fig. 1. Absorption coefficient as a function of wavelength of SiC. The three curves are forSiC specimens with different impurity concentrations [16].

emissivity spectra at room temperature for 0.1%, 0.25% and 0.5% Cr2O3

doped compounds as shown in Fig. 2 indicates that the concentrationof Cr2O3 dramatically enhanced the normal spectral emissivitybetween 1 and 5 μm (i.e. 2000–10,000 cm−1). This enhancement isrelated to formation or precipitation of new phases on the grainboundary due to substitution, which will change the grain boundaryarea. But the mechanism still needs to be further digested.

Another consequence caused bydopant is the crystal distortion dueto different ion radii, which also has the contribution to radiation. Prattet al. incorporated NiO into MgO–Al2O3–SiO2 system glasses heated at1020 °C for 2 h and the whole-band normal direction emissivitychanges from 0.82 to 0.90 by doping NiO [18]. When the Ni2+ isincorporated into cordierite structure and substituted with Mg2+, theorder degree of Al/Si in crystal structure of cordierite is decreased dueto different radii of Ni2+ and Mg2+, leading to the crystal structuredistortion. This in turn decreases the symmetry of lattice vibration,thus enhances the effects of anharmonic vibration of polar lattice, thecoupled action of phonon, and phonon combination radiation. There-fore, the infrared radiance of this material is enhanced by NiO doping.

Moreover, the dopant (rare-earth oxides and transition elements)could tune the spectral emissivity thus change the emissive power[19,20]. Incorporation of 2–4 wt.% Co3O4 or NiO into MgO host (a lowinfrared emissivity) produced “matched emitters”, which means itsemissive power spectrum is matched very efficiently to the responseof GaSb photovoltaic cells that convert the infrared radiation intoelectricity. Fig. 3 shows the continuous, strong radiant emissionswhich is up to 0.9 in the optimal energy range between 1 and 2 μmandminimal radiation at higher wavelengths [21,22]. The total emissivepower between 1 and 9 μm is calculated by integrating the area undereach curve. Interestingly, both the blackbody emitter and the NiO-doped MgO emitter have the same total emissive power of 144,000 Wcm−2. However, it is noticeable that the NiO-doped MgO emitteremits most of power at wavelengths less than 2 μm, instead of 1–5 μmin theoretical blackbody. The phenomenon results from intra-atomicelectronic transitions determined by the electronic configuration ofthe dopant ions and interactions with the coordinating crystal field ofthe host oxide [23].

2.2. Surface roughness

The correlation between emissivity and surface roughness hasbeen explored both theoretically [24,25] and experimentally[26,27].Wen et al. [24] modeled the effect of surface roughness on emissivityof aluminum alloys. Normally the surface involves two broadcategories: optically smooth (ideal) and rough (real) surfaces. For

Fig. 3. Theoretical blackbody emissive power spectrum at 1010 °C compared with the emissive power spectrum for an experimental 2 wt.% NiO-doped MgO emitter at 1404 °C. [22]

5122 X. He et al. / Thin Solid Films 517 (2009) 5120–5129

optically smooth surfaces, spectral emissivity is determined bycombining Fresnel's equation and Kirchhoff's law yielding

eλ =4n2

n + 1ð Þ2 + k2ð4Þ

Where n is the index of refraction and k the extinction coefficient.Fig. 4(a) compares spectral emissivity values using three differentmethods. It indicates that the experimental result shows the sametrend as calculated one. Namely, the emissivity decreases aswavelength increases in the infrared range for most metallic surfaces.

For real surfaces, the surfaces can be divided into three regionsbased on optical roughness (σ/λ, the ratio between surface roughnessto wavelength): the specular region (0bσ/λb0.2), intermediateregion (0.2bσ/λb1) and the geometric region (σ/λN1). In thespecular region, it assumes that the reflection of incident radiationis specular, and the diffraction theory is used to predict the effects ofsurface roughness on emissivity [28,29]. For a Gaussian distribution ofsurface heights, the relationship between reflectance and opticalroughness follows an exponential decay function of σ/λ, which isidentified both theoretically [30,31] and experimentally [32,33],

ρr = ρp exp − 4πσλ

� �2� �ð5Þ

where ρr and ρp are the reflectances for a rough surface and a polishedsurface, respectively. Eq. (5) reveals that the larger α could result inlower reflectance, i.e. higher emittance, which is verified in alumi-num-coated ground glass as shown in Fig. 4(b) [29]. In theintermediate region, the directional reflectance ρλ′ (θi) can bedetermined by integrating the bidirectional reflectance function, ρλ″(θi, θs) over the respective hemisphere for a given angle of incidence[34] as indicated in Fig. 4(c). According to Kirchhoff's law and energyconservation, the directional emissivity can be calculated by Eq. (6)

e Vλ θið Þ = 1− ρ Vλ θið Þ ð6Þ

Buckius et al. [35–39] theoretically confirmed that increasedroughness increases directional emissivity in the intermediate region.In the geometric region, geometric optics is used to predict emissivity,

ignoring the diffraction effects. Surfaces with repeatable groovedfinish, like V-shaped grooves [40], circular grooves [41], and pyramidalgrooves [42], are commonly used to model the emissivity enhance-ment. The slope of the grooved surface plays an important role inemissivity enhancement. For example, the inter-reflection caused by a60° V-groove in Fig. 4(c) and a small grove angle (also the slope) inFig. 4(d) could significantly enhance the normal emissivity. Overall,modeling results lead to a conclusion that the emissivity is dependenton surface roughness for σ/λb1 and on surface slope for σ/λN1.

In experiment, the emissivity enhancement due to roughness wasverified in opaque materials, such as steel [26] and Pr2NiO4 coatings[27]. Fig. 5 compares normal spectral emissivity of rough Pr2NiO4

coating and single crystal Pr2NiO4 (considered as smooth surface) inthe (a–b) plane at 1000 K, in the working temperature range of theinfrared radiator, from far to midinfrared. The spectral response of thecoating is 30% to 40% higher than that of the single crystal, which islargely dominated by the infrared absorptions produced by localizedpolarons. Clearly, the enhancement of the intrinsic absorption isobtained by increasing the optical path of the light between eachsmall cavity of the surface. Furthermore, the effect produced by therough Pr2NiO4+d coating is also clearly demonstrated in Fig. 5. Thelow normal spectral emissivity of the original semitransparentceramic in the midinfrared range (from 2500 to 5000 cm−1) iscompletely occulted after the deposit.

Although the previous achievements provide valuable insight intothe emissivity principles concerning the relationship betweenemissivity and surface roughness, the influences of surface roughnesson emissivity remain illusive, and sometimes current emissivitymodels are consistent with experimental emissivity values. Auniversal emissivity model concerning roughness effect needs to beproposed in order to completely agree with the practically measuredemissivity.

2.3. Surface texture

The spectral radiative properties were completely dependent onthe competition between the absorption coefficient (following thewavelength) and the radiative scattering. The surface texture (inmaterials science, texture is the distribution of crystallographic

Fig. 4. Classification of emissivity models based on surface roughness [24].

5123X. He et al. / Thin Solid Films 517 (2009) 5120–5129

orientations of a sample) often forming during coating deposition, hasan important effect on radiative scattering. For example, a surfacetexture with high aspect ratio feature (height to width ratioN5: 1) canserve as a collection of a lot of individual black-bodies, which exhibitan emissivity of around 0.85 [43]. The theoretical and experimentalresults of IR spectral emittance of nonsphere-shaped particles indicatethat coating composing of oriented rhenium whiskers with highaspect ratio (Height/Dimension=5) provides the high emittancevalue, compared to that of randomly oriented whisker coating [44,45].However, the texture effect on emissivity is still in argument. Recently,Rozenbaum et al. [46] found that unlike the other previously studied

Fig. 5. Normal spectral emissivity at T=1000 K and from 400 to 5000 cm−1 ofcordierite–mullite substrate (dashed line), single crystal of Pr2NiO4+d (dotted line),and rough Pr2NiO4+d coating on the cordierite–mullite substrate (solid line) [27].

ceramics, the highly textured alumina ceramic has a much loweremissivity than the single crystal between 1300 cm−1 and 1800 cm−1

(shown in Fig. 6) and the spectral emissivity of this sample wasradically different from the emissivity spectra of lowly texturedalumina. In this range, due to the weak absorption coefficient, theradiation mean path inside the sample was higher than the thicknessof the single crystal. For longer wavelengths, the emissivity increasesto become higher than those of the single crystal and tends to those ofthe lowly textured sample in the transparent zone. This is in goodagreement with the previous reports. So it is concluded that weshould consider wavelength range whenwe discuss the texture effect.

Fig. 6. Normal spectral emissivity of alumina ceramics (thickness=1 mm, T=1, 350 K)with two different textures [46].

Fig. 8. Comparison of emission power density between the photonic-crystal sample anda blackbody cavity radiator at T=750 K [52].

Fig. 7. (a) A SEM view of a 3D tungsten photonic crystal. Within each layer, the 1D rodwidth is 0.5 μmand the rod-to-rod spacing is 1.5 μm. (b) A computed absorption spectrafor an eight-layer 3D tungsten photonic-crystal sample. [51].

5124 X. He et al. / Thin Solid Films 517 (2009) 5120–5129

2.4. Surface periodicity

A three-dimensional (3D) complete photonic band gap can be usedto suppress radiation below the electronic band gap [47–50]. Mean-while, emission can be enhanced at a narrow band near a photonicband edge or a narrow allowed band. If both effects are combined, anearly ideal radiation spectrum can be obtained. Photonic crystalsrefer to the periodic optical (nano)structures that are designed toaffect the motion of photons in a similar way that periodicity of asemiconductor crystal affects the motion of electrons. Photoniccrystals are composed of periodic dielectric or metallo-dielectric(nano)structures that affect the propagation of electromagnetic waves(EM) in the same way as the periodic potential in a semiconductorcrystal affects the electron motion by defining allowed and forbiddenelectronic energy bands. Since this basic physical phenomenon isbased on diffraction, the periodicity of the photonic crystal structurehas to be of the same length-scale as half the wavelength of the EMwaves i.e. ∼200 nm (blue) to 350 nm (red) for photonic crystalsoperating in the visible part of the spectrum— the repeating regions ofhigh and low dielectric constants have to be of this dimension. Thismakes the fabrication of optical photonic crystals cumbersome andcomplex.

Interestingly, Lin et al. realized a 3D tungsten photonic-crystalsample (shown in Fig. 7(a)) using a modified silicon process [51],which can be thermally excited and exhibits emission at a narrowband [52,53]. In theory, the computed absorption spectrum for aneight-layer photonic-crystal sample in Fig. 7(b) indicates that theabsorptance is low for λN3 μm (the photonic band gap) and increasesslightly at λ∼3–4 μm due to a higher tungsten material absorption atthese wavelengths [54]. This band gap is a complete band gap, capableof trapping light fully in all three dimensions and for both

polarizations [47]. Beyond the photonic band gap regime, there aretwo strong absorptions: one is located at λ=2.5 μm with anabsorptance of ∼40% and the other located at λ=1.5–1.9 μm withan even stronger absorptance of ∼80%. As discussed above, Kirchhoff'slaw tells us that at equilibrium for a given wavelength λ andtemperature T, the emittance of any body is equal to its absorptance[10]. Thus, a strong narrow absorption is suggestive of a narrow bandemission from a photonic-crystal sample.

In experiment, the measured power-density spectrum of thesample in Fig. 8 is compared with that computed for a blackbodycavity radiator at T=750 K. The blackbody radiation follows thePlanck radiation. In the bandgap, the emission is suppressed andbelow the blackbody emission. However, photonic-crystal emission inthe peak regime is found to exceed that of the blackbody by ∼300%.Note the photonic band gap emission exhibits a narrow FWHM ofΔλ∼1.8 μmwhile black body shows a wide FWHM of Δλ∼4.5 μm. Theabsorption rate of 3D metallic photonic crystal is found to besuppressed in the photonic bandgap regime (λ∼8–16 μm) andstrongly enhanced at the photonic band edge (λ∼6 μm) The peakabsorption in tungsten photonic crystal is an effective absorption thatoriginates from the intrinsic tungsten absorption and is subsequentlyenhanced by the photonic band edge effect. The enhancement isattributed to a longer photon-matter interaction length as comparedto the metallic skin depth, the slower group velocity of light at theband edge, and the finite metallic absorption constant. It should bestressed that the photonic-crystal by no means exceeds the black-body emission in the entire wavelength spectrum, rather theexceeding happens at the neighborhood of the band-edge only [55].Although the photonic crystal has been rapidly developed, the coatingor thin film of photonic crystal has not been realized yet.

2.5. Coating thickness

As mentioned above, when a radiation falls on a body, it may bepartially reflected, transmitted, or absorbed. As such, light penetrationdepth (defined as the inverse of the absorption coefficient (α)) isclosely related to the emissivity of coatings. For the light with thewavelength λ, the depth of light penetration, dp, is associated with theextinction coefficient (k), given by

dp =1α

=λ

4πkð7Þ

Eq. (7) indicates that dp is reversely proportional to k. Thus, k isexpected to be as low as possible in order to get high emissivity ofcoatings. As a result, the coating is expected to be as thick as possible,at least thicker than dp (otherwise, the transmittance will occur). Butin practice it is impossible to fabricate unlimitedly thick coating due toproduction cost or fabrication limitation. In the past, models and

Fig. 10. Emissivity (λ=1.53 μm) versus oxide thickness at 925 °C [60].

5125X. He et al. / Thin Solid Films 517 (2009) 5120–5129

experiments concerning the relationship between emissivity andcoating thickness were explored in details [56]. A typical relationshipbetween them in particle embedded in Al2O3 coating is shown inFig. 9, revealing that initially the emissivity increases rapidly asthickness increases and becomes steady above thickness of 40 μm.Similar relationship was also found in other ceramic coatings, such asSiC, Si3N4 and A12O3 [5], anodic film on aluminum [57], Si3N4 and SiO2

on silicon substrates [58]. Therefore, it is concluded that the coatinghas a critical thickness over and above which emissivity almost doesnot change.

When the coating has more than two layers (called multilayercoating), the optical interference will happen. If the thickness ofmultilayered coatings falls into the same order of magnitude as theincident radiation, a considerable amount of optical interferencewithin the multilayer structure tends to occur [59]. The net surfacereflectivity and absorption often change significantly even thoughthere is any thickness change in any layer. The resultant temperatureprofiles and quality of the film are therefore affected by these changesin optical properties. This was evident in SiO2/Si system, as shown inFig. 10 [60]. The relationship between emissivity and thickness showsa converted “V” shape. The emissivity peaks at the SiO2 thickness ofaround 240 nm for a givenwavelength. When the coating is ultra thin,approaching zero, smaller than the wavelength of the radiation light,the increase of emissivity is ascribed to the coating thickness. Thereflections from the second interface are in phase with the one fromthe first interface, resulting in a high reflection and low emissivity. Asthe coating thickness increases, when the phase difference of reflectedlights from surface and interface is equal to the radiation lightwavelength, extinction happens where the emissivity reaches themaximum value. As the thickness further increases, this phasedifference deviates from the wavelength and causes the emissivitydecreasing. However, this critical thickness (from several tens to100 μm) where interference of radiation light occurs depends on thecoating composition [61].

2.6. Nanocomposite coatings

For normal incident radiation, reflection coefficient R is given bythe Fresnel's formula:

R =n−1ð Þ2 + k2

n + 1ð Þ2 + k2ð8Þ

where n is the refractive index, k the extinction coefficient. Accordingto Eq. (4), the extinction coefficient k becomes much larger than 1

Fig. 9. Comparison of theoretical and experimental data of emissivity as function ofoxide coating thickness [55].

when spectral emissivity ε is approaching the maximum value of 1.Eq. (8) states the reflection coefficient tends to reach a maximumvalue of 1. As a result, there is an abrupt decrease of ε while R has asudden increase as wavelength increases, which is clearly demon-strated in Fig. 11 for boron nitride [62]. Note that a peak of R in Fig. 11(a) accompanies with a sudden drop of emissivity and Fig. 11(b). Thisconcluded that single component material can hardly show highemissivity on the whole frequency range.

Besides pursuit of high emissivity, the other properties such asmechanical properties, thermal shock resistance and oxidationresistance, should be also taken into account when high emissivitycoatings are used at high temperature (N1000 °C) likeMTPS. However,single component coating could rarely possess these comprehensiveproperties. Designing nanocomposite coatings may help a lot. Ananocomposite coating comprises of at least two phases: a crystallinephase and an amorphous phase, or two different crystalline phases, inwhich at least one is capable of absorbing and re-radiating thermalenergy (emissivity agents), such as BN, SiC, SiB6, traditional metaloxides, refractory metals carbides [63]. The other phase is like bindersuch as colloidal silicon dioxide, which can provide high strength,thermal expansion characteristics similar to their intended substrates,and adequate bond strength with the substrate [64]. These compositecoatings represent a new class of materials, whose mechanical oremittance properties are not subjected to volume mixture rules, butdepends on synergetic interactions of the composite constituents.

Recently, our group deposited SiC/SiO2 composite coating on thepre-oxidized 316 SS substrate via EB-PVD technology [65,66]. It isinteresting to find that nanocrytalline SiC deposited at high tempera-ture was formed, as shown in Fig. 12, giving rise to the nanohardnessas high as 50 GPa. At the same time, it is well known that SiC exhibitsvery good emittance property. Combination of emittance andmechanical properties makes SiC promising applied in high tempera-ture environment or space, like MTPS [67]. In SiC/SiO2 system, SiO2

can act as an additive oxidation insulator because of low thermalconductivity and high oxidation resistance. The spectral directionalemissivity of SiC/SiO2 composite thin film almost keeps a constant ofaround 0.65 at a wide wave-number range from 400 to 4000 cm−1, asshown in Fig. 13. Unfortunately, the other properties like oxidationresistance, shock resistance, have not been investigated. But it can bepredicted that the nanocomposite coatings are a promising branch ofmaterials that could potentially possess both high emissivity andunique mechanical or thermal shock resistance properties.

Another composite coating structure, i.e. multilayer, is proposed inorder to meet comprehensive requirements, such as oxidation,emittance, and low catalysis characteristics in hypersonic structures.In the multilayered system, normally five different layers are realized(schematically shown in Fig. 14) [68]: (i) low catalysis layer (based onborosilicate or borophosphate glasses) to minimize aerothermody-namic heating; (ii) the high-emittance layer consisting of various

Fig. 11. Reflectance and emittance of boron nitride [62].

5126 X. He et al. / Thin Solid Films 517 (2009) 5120–5129

Fig. 12. HRTEM and cross-sectional images of SiC thin film prepared by EB-PVD at900 °C.

Fig. 14. Schematic diagram of the multilayer coating [68].

5127X. He et al. / Thin Solid Films 517 (2009) 5120–5129

constituents, such as SiB6, providing emittance enhancement to thetotal system; (iii and iv) the oxygen-resistance layer to block oxygendiffusion, comprising of a base layer and sealing layer; (v) reactionbarrier layer (like Al2O3, TiAl3) to prevent reaction between thesubstrate alloy and the other coating layers. It is recognized thatthermal expansion coefficient differences between layers are inherentin this multilayer design. Accordingly, we can accommodate thermalstrains by adjusting the layer thickness, and its self-healing char-acteristics can help avoid the cracking that may occur.

Bird et al. realized this multilayer design by sol–gel method [68].The emittance layer of the coating imparted an emittance in excess of0.8 during exposure at 980 °C for up to 100 h, which was at least 15%greater than that for the coated alloy without the emittance layer asshown in Fig. 15. Likewise, the low catalysis outer layer of the coatingreduced the catalytic efficiency for recombination of atomic oxygenand nitrogen in hypersonic flowing air. The catalytic efficiency ofcoated Inconel 617 was reduced by a factor of four, compared to theuncoated condition, for a 1-h exposure at 980 °C.

To summarize, nanocomposite coatings with multiple componentsor multilayered structure may be most attractive to reach highemissivity value within thewhole frequency range as well as excellentmechanical or thermal properties. The research of multilayer design

Fig. 13. The emissivity and reflectivity of SiC/SiO2 composite thin film as a function ofwavenumber [9].

has just begun, and it is expected to attract ever increasing attention inthe near future.

3. Emissivity characterization

In a classical definition, emissivity measures the ability of amaterial to radiate the heat through electromagnetic wavelength.Emissivity measurements are difficult or tedious in practice because itis a function of several variables and an extrinsic parameter of amaterial. Althoughmanymethods have been proposed and developedso far, a standard procedure or a universally accepted methodologydoes not exist yet. Qualitatively, the emissivity can be estimated byfour developed basic methods: Calorimetric method [69–71]; opticalreflectivity method [72], multispectral radiation thermometry [73]and radiation energy method [74]. Calorimetric method involvesmeasuring the heat power delivered to the test sample as well as thetemperatures of the emitting surface (the sample) and the surround-ings in a vacuum over time [70,71]. An energy balance is then used to

Fig. 15. Emittance (a) and weight-change (b) as a function of exposure in air at 1000 °Cfor coated Inconel 617 with emittance coating embedded within the oxidation coating[68].

5128 X. He et al. / Thin Solid Films 517 (2009) 5120–5129

determine the emissivity. The sample must be well insulated on theside and back, and parasitic heat losses must be accounted for toaccurately determine the heat transfer through the sample surface[75]. Reflectivity method involve illuminating a sample with infraredenergy and measuring the percentage of energy reflected from thesurface [4,76]. The absorptance is calculated from the reflectance andused to calculate normal emittance by Kirchhoff's law and the Stefan–Boltzmann equation. Multispectral radiation thermometry (MRT) [73]employs radiancemeasurements at three ormorewavelengths and anemissivity model [77,78]. This method could measure the spectralemissivity and temperature simultaneously based on the relationshipbetween temperature and spectral emissivity as follows:

eλ = exp a0ffiffiffiλ

p� �ð9Þ

This method provides many advantages, such as sample-prepara-tion fast test free, on site test, no limitation for temperature, but itsaccuracy is lower and there is no commonmodel to be applicable of allthe materials. The most widely method used in the assessment ofemissivity of coating is energy method which is based on themeasurement of radiation power of sample surface over the radiationpower of black body at the same temperature (Stefan–Boltzmannequation). The key problem is to select the suitable black body whichdetermines the measurement accuracy. More details for emissivitymeasurement via energymethod could be found in references [74,79–83]. Fourier spectroscopy based energy method for emissivity test hasbeen rapidly developed recently, which prides more accuracy (lessthan 3%) of emissivity measurement. The temperature range ofmeasurement is from−20 to 2000 °C and thewavelength ranges fromvisible light to above 25 um.

4. Prospect

Absorption coefficient α and scattering coefficient S are importantfactors to be considered when designing a high emissivity coating;increasing α and decreasing S can enhance the emissivity. In principle,all emissivity enhancement mechanisms can be applied to theenhancement of ceramic coatings, the real challenge lies in theimplementation. However, for real engineering applications, pursuinghigh emissivity is not enough, while a combination of high emissivityand other comprehensive properties (called multifunctional coatings)will be pursued more by the academics driven by the industrial needs.In-situ synthesis of nanocrystalline or nanocomposite coatings (such asnc-SiC/α-SiO2 or nc-SiC/α-C) or designing multilayered structurallayersmay hold the key (at least, two of the keys) to thismultifunctionalcoating. Another promisingbranchcalled variable emittance coating hasjust come up [84]. Small spacecraft, including micro and nanosatelliteswill require an alternativemeans to achieve thermal control due to theirsmall power andmass budgets. Variable emittance coatingswill be usedas spacecraft radiators for thermal balance.

In respect to the emissivity measurements, the lack of a universallyaccepted emissivity measurement methodology and instrumentationfor any shape or different size samples has effectively hindered theprogress of the research. It is hoped that in a near future, a universallyaccepted methodology will be born, and that this will accompany thebirth of an emissivity measurement apparatus for coatings, a stand-alone machine. To catalyze the birth of the universal emissivitymeasurement, regular international seminars should be organizedespecially on this topic. Fourier spectroscopy based energy methodmay become the promising technique and needs further development.

References

[1] B.V. Cockeram, D.P. Measures, A.J. Mueller, Thin Solid Films 355–356 (1999) 17.[2] S. Kawakita, M. Imaizumi, T. Sumita, K. Kushiya, T. Ohshima, M. Yamaguchi, S.

Matsuda, S. Yoda, T. Kamiya, K.Kurokawa, L.L. Kazmerski, B.McNelis,M.Yamaguchi, C.

Wronski, W.C. Sinke, Proceedings of 3rd World Conference on Photovoltaic EnergyConversion, Osaka, Japan, May 11–18, 2003, p. 8PB511.

[3] T. Sumita, M. Imaizumi, S. Kawakita, S. Matsuda, S. Kuwajima, in: J.L. Wert (Ed.), IEEERadiationEffectsDataWorkshopRecord,Monterey, U.S.A., July21–25, 2003,P11, 2003.

[4] T. Sumita,M. Imaizumi, S. Kawakita, S.Matsuda, S. Kuwajima, T. Ohshima, T. Kamiya, in:J. Black (Ed.), IEEE Radiation Effects DataWorkshop Record, Atlanta, U.S.A., July 19–23,2004, P130, 2004.

[5] M. FaIz, G. Leonhardt, Surf. Coat. Technol. 61 (1993) 97.[6] M.G. Hocking, V. Vasantasree, P.S. Sidky, Metallic Ceramic Coatings: Production,

High Temperature Properties Applications, Longman, London, 1989 P670.[7] C. Bartuli, T. Valente, M. Tului, Surf. Coat. Technol. 155 (2002) 260.[8] A. Ordine, C.A. Achete, O.R. Mattosa, I.C.P. Margarit, S.S. Camargo Jr., T. Hirsch, Surf.

Coat. Technol. 583 (2000) 133.[9] J. Yi, X.D. He, Y. Sun, Y. Li, Appl. Surf. Sci. 253 (2007) 4361.[10] M. Auslender, S. Hava, Infrared Phys. Technol. 36 (1995) 1077.[11] G.B. Rybicki, A.P. Lightman, Radiative Processes in Astrophysics, JohnWiley & Sons,

New York, 1979 22 pp.[12] N.M. Ravindra, W. Chen, F.M. Tong, Arun Nanda, in: N.M. Ravindra, R.K. Singh

(Eds.), Transient Thermal Processing Techniques in Electronic Materials, TMS Soc.,Warrendale, PA, Aug. 1996, pp. 159–164.

[13] C.R. Bamford, Phys. Chem. Glasses 3 (1962) 189.[14] J.A. Duffy, J. Am. Ceram. Soc. 60 (1977) 440.[15] B.M. Wang, D.Z. Su, G.Y. Zhang, China J. Infrared Res. 2 (1983) 55 (in Chinese).[16] W.G. Spitzer, D. Kleinman, D. Walsh, Phys. Rev. V113 (1959) 127.[17] F. Seronde, P. Echegut, J.P. Coutures, F. Gervais, Mater. Sci. Eng. B8 (1991) 315.[18] R.H. Pratt, Radiat. Phys. Chem. 70 (4–5) (2004) 595.[19] M.G. Krishna, M. Rajendran, D.R. Pyke, A.K. Bhattacharya, Sol. Energy Mater. Sol.

Cells 59 (1999) 337.[20] A. Lowe, D.L. Chubb, S.C. Farmer, B.S. Good, Appl. Phys. Lett. 64 (1994) 3551.[21] L.G. Ferguson, Fatih Dogan, Mater. Sci. Eng. B83 (2001) 35.[22] L.G. Ferguson, F. Dogan, J. Mater. Sci. 36 (2001) 137.[23] K. Nassau, The Physics and Chemistry of Color— The Fifteen Causes of Color, Wiley,

New York, 1983 104 pp.[24] C.D. Wen, I. Mudawar, Int. J. Heat Mass Transfer 49 (2006) 4279.[25] F. Ghmari, T. Ghbara, J. Appl. Phys. 96 (2004) 2656–2664.[26] J.F. Ony Sacadura, Int. J. Heat Mass Transfer 15 (1971) 1451.[27] B. Rousseau, M. Chabin, P. Echegut, A. Sin, F. Weiss, P. Odier, Appl. Phys. Letts. 79

(2001) 3633.[28] R. Siegel, J.R. Howell, Thermal Radiat. Heat Trans. McGraw-Hill, New York, NY, 1981.[29] M.J. Haugh, Radiation thermometry in the aluminum industry, in: D.P. DeWitt, G.D.

Nutter (Eds.), Theory and Practice of Radiation Thermometry, John Wiley,New York, NY, 1988.

[30] J. De Santo, G. Brown, in: E. Wolf (Ed.), Progress in Optics, Cambridge UniversityPress, Cambridge, 1991.

[31] R. Carminati, J.J. Greffet, A. Sentenac, in: J.S. Lee (Ed.), Proceedings of the 11thInternational Heat Transfer Conference, vol. 7, Korean Society of MechanicalEngineers, Kyongju, 1998, p. 427.

[32] R. Petit, Electromagnetic Theory of Gratings, Springer, Berlin, 1980.[33] P. Lalanne, G.M. Morris, J. Opt. Soc. Am. A 13 (1996) 779.[34] M.Q. Brewster, Thermal Radiative Transfer and Properties, John Wiley, New York,

NY, 1992.[35] R.A. Dimenna, R.O. Buckius, J. Thermophys. Heat Transf. 3 (1994) 393.[36] R.A. Dimenna, R.O. Buckius, Trans. ASME 116 (1994) 639.[37] K. Tang, R.A. Dimenna, R.O. Buckius, Int. J. Heat Mass Transfer 40 (1997) 49.[38] K. Tang, R.A. Dimenna, R.O. Buckius, Int. J. Heat Mass Transfer 41 (1997) 2037.[39] K. Tang, R.A. Dimenna, R.O. Buckius, Int. J. Heat Mass Transfer 44 (2001) 4059.[40] G.J. Dail, M.G. Fuhrman, D.P. DeWitt, Proceeding of 4th Int. Aluminum Extrusion

Technology Seminar, Chicago, IL, vol. 2, April, 11–14, 1988, p. 281.[41] K. Kanayama, Heat Transfer-Jpn. Res. 1 (1972) 11.[42] A. Abdulkadir, R.C. Birkebak, Rev. Sci. Instrum. 45 (1974) 1356.[43] W. Konz, J. Hildenbrand, M. Bauersfeld, S. Hartwig, A. Lambrecht, V. Lehmann, J.

Wöllenstein, Proceedings Microsystems for the New millennium, SPIE, vol. 5836,2005, doi:10.1117/12.608712.

[44] J.R. Aronson, A.G. Emslie, F.E. Ruccia, C.R. Smallman, E.M. Smith, P.F. Strong, Appl.Opt. 18 (1979) 2622.

[45] W.L. Xu, S.C. Shen, Appl. Opt. 36 (1997) 1644.[46] D. Rozenbaum1, De Sousa Meneses P. Echegut, Int. J. of Thermophys. 30 (2009)

580.[47] J.G. Fleming, S.Y. Lin, I. El-Kady, R. Biswas, K.M. Ho, Nature (London) 417 (2002) 52.[48] S.Y. Lin, J.G. Fleming, Z.Y. Li, I. El-Kady, R. Biswas, K.M. Ho, J. Opt. Soc. Am. B 20 (2003)

1538.[49] E. Yablonovitch, Phys. Rev. Lett. 58 (1987) 2059.[50] S. John, Phys. Rev. Lett. 58 (1987) 2486.[51] S.Y. Lin, J. Moreno, J.G. Fleming, Appl. Phys. Lett. 83 (2003) 380.[52] S.-Y. Lin, J.G. Fleming, I. El-Kady, Optics Lett. 28 (2003) 1909.[53] T. Trupke, P. Würfel, M.A. Green, Appl. Phys. Lett., 84 (2004) 1997.[54] M.A. Ordal, L.L. Long, R.J. Bell, S.E. Bell, R.R. Bell, R.W. Alexander Jr., C.A. Ward, Appl.

Opt. 22 (1983) 1099.[55] S.Y. Lin, J. Moreno, J.G. Fleming, Appl. Phys. Lett. 84 (2004) 1999.[56] J.W. Zhou, X.Q. Wang, J.C. Han, Acta Mater. Compositae Sin. 24 (2007) 86.[57] C.S. Kumar, A.K. Sharma, K.N. Mahendra, S.M. Mayanna, Sol. Energy Mater. Solar

Cells 60 (2000) 51.[58] P.L.A. Chr. M. van der Meer, L.J. Giling, S.G. Kroon, J. Appl. Phys. 47 (1976) 652.[59] P.Y. Wong, I.N. Miaouiis, P. Zavracky, Sens. Actuator 19 (1990) 175.[60] N.M. Ravindra, S. Abedrabbo,W. Chen, FeimingM. Tong, Arun K. Nanda, Anthony C.

Speranza, IEEE Trans. Semicond. Manuf. 11 (1998) 1.

5129X. He et al. / Thin Solid Films 517 (2009) 5120–5129

[61] M.H. Han, X.G. Liang, Y. Huang, Chin. Sci. Bull. 50 (2005) 588.[62] Y.S. Touloukian, D.P. DeWitt, Thermal Radiative Properties: Nonmetallic Solids, IFI/

Plenum, New York, 1972 P. 1042, 1047.[63] Y. Ozaki, R.H. Zee, Mater. Sci. Eng. A, 202 (1995) 134.[64] H.W. Babel, H.G. Le, Us patent No. 5,296,288 (1994).[65] X.D. He, J. Yi, Y. Sun, P. Xiao, X.F. Zhao, Appl. Phys. Lett. 91 (2007) 241914 1.[66] J. Yi, X.D. He, Y. Sun, Y. Li, M.W. Li, Appl. Surf. Sci. 253 (2007) 7100.[67] A.K. Costa, S.S. Camargo Jr., C.A. Achete, R. Carius, Thin Solid Films 377–378 (2000)

243.[68] R. Keith Bird, Terryl A. Wallace, J. Spacecr. Rockers 41 (2004) 213.[69] J. Hameury, B. Hay, J.R. Filtz, Int. J. Thermophys. 28 (2007) 1607.[70] R.D. Karam, in: P. Zarchan (Ed.), “Thermal Control Hardware,” Satellite Thermal

Control for Systems Engineers, AIAA, Reston, VA, 1998, pp. 160–164.[71] A. Mychkovsky, and R. Ponnappan, “Calorimetric Measurement of Emissivity in

Space Conditions,” AIAA 2005-961, Jan. 2005.[72] Dunkle R.V. In progress in International Research on Thermodynamics and

Transport Properties Asme[M], 1962, p. 100.[73] C.D. Wen, I. Mudawar, J. Commun. Heat Mass Transfer 33 (2006) 1063.

[74] J. Ishii, A. Ono, Meas. Sci. Technol. 12 (2001) 2103.[75] R.D. Karam, Thermal control hardware, in: P. Zarchan (Ed.), Satellite Thermal

Control for Systems Engineers, AIAA, Reston, VA, 1998, pp. 160–164.[76] D.A. Jaworske, T.J. Skowronski, in: M. El-Genk (Ed.), Proceedings of the Space

Technology and Applications International Forum (STAIF-2000), AIP ConferenceProceedings, No. 504, American Institute of Physics Press, New York, 2000, p. 791.

[77] P.B. Coates, Metrologia 17 (1981) 103.[78] C. Wen, I. Mudawar, J. Heat Mass Transfer 48 (2005) 1316.[79] B.M. Liu, Z.X. Chu, 10th European Conference on Thermophysical Properties, 1986,

p. 22.[80] S. Clausen, A. Morgenstjerne, O. Rathmann, Appl. Optics 35 (28) (1996) 5683.[81] A. RKorb, P. Dybwad,W.Wadsworth, J.W. Salisbury, Appl. Optics 35 (10) (1996) 1679.[82] E. Lindermeir, V. Tank, P. Haschberger, Proceedings of SPIE-The Int. Society for

Optical Eng., vol. 1682, 1992, pp. 354–364.[83] J.M. Dai, X.D. Liu, X.G. Sun, 7th Int. Workshop on Subsecond Thermophysics, 2004.[84] D. Farrar, W. Schneider, R. Osiander, J.L. Champion, A.G. Damn, Inter Society

Conference on Thermal Phenomena, 2002, p. 1020.