j_002

Refractory Diborides of Zirconium and Hafnium

William G. Fahrenholtz*,w and Gregory E. Hilmas*

Materials Science and Engineering Department, University of Missouri-Rolla, Rolla, Missouri 65409

Inna G. Talmy* and James A. Zaykoski*

Naval Surface Warfare Center, Carderock Division, West Bethesda, Maryland 20817

This paper reviews the crystal chemistry, synthesis, densica-tion, microstructure, mechanical properties, and oxidation be-havior of zirconium diboride (ZrB2) and hafnium diboride(HfB2) ceramics. The refractory diborides exhibit partial orcomplete solid solution with other transition metal diborides,which allows compositional tailoring of properties such as ther-mal expansion coefcient and hardness. Carbothermal reductionis the typical synthesis route, but reactive processes, solutionmethods, and pre-ceramic polymers can also be used. Typically,diborides are densied by hot pressing, but recently solid stateand liquid phase sintering routes have been developed. Fine-grained ZrB2 and HfB2 have strengths of a few hundred MPa,which can increase to over 1 GPa with the addition of SiC. Purediborides exhibit parabolic oxidation kinetics at temperaturesbelow 11001C, but B2O3 volatility leads to rapid, linear oxida-tion kinetics above that temperature. The addition of silica scaleformers such as SiC or MoSi2 improves the oxidation behaviorabove 11001C. Based on their unique combination of properties,ZrB2 and HfB2 ceramics are candidates for use in the extremeenvironments associated with hypersonic ight, atmospheric re-entry, and rocket propulsion.

I. Introduction

ZIRCONIUM diboride (ZrB2) and hafnium diboride (HfB2) aremembers of a family of materials known as ultra high-tem-perature ceramics (UHTCs). Several carbides and nitrides of thegroup IVB and VB transition metals are also considered UHT-Cs based on melting temperatures in excess of 30001C and otherproperties. Very few elements or compounds from any class ofceramic materials have melting temperatures approaching30001C (Fig. 1).1 From the broader family of UHTCs, this

paper focuses on ZrB2 and HfB2. Discussion of TiB2 has beenintentionally minimized due to its more widespread use as armorand cutting tools.24

Some structural, physical, transport, and thermodynamicproperties of ZrB2 and HfB2 are summarized in Table I.

512

Applications that take advantage of these properties include re-fractory linings,1315 electrodes,1618 microelectronics,19 and cut-ting tools.20 In addition to high melting temperatures, ZrB2 andHfB2 have a unique combination of chemical stability, highelectrical and thermal conductivities, and resistance to erosion/corrosion that makes them suitable for the extreme chemical andthermal environments associated with hypersonic ight, atmos-pheric re-entry, and rocket propulsion.2124 Because of recentefforts to develop hypersonic aerospace vehicles and re-usableatmospheric re-entry vehicles, interest in UHTCs has increasedsignificantly in the past few years. As a result, groups in theUnited States, Italy, Japan, India, and China are investigatingdiborides. Many of these studies draw inspiration from the lateProfessor G. V. Samsonov, whose work is noted for both itshigh quality and enormous quantity.This article provides a critical evaluation of historic (1970s or

earlier) and recent (1990 or later) studies of ZrB2 and HfB2 cer-amics as well as presenting some recent results. Fundamentalaspects of crystal structure and bonding, synthesis, densication,microstructures, properties, and oxidation behavior are exam-ined.

II. Crystal Structure and Bonding

Crystal chemistry and crystal structure determine the chemical,physical, and thermal properties of materials. For the transitionmetal diborides, fundamental aspects of crystal structure andcrystal chemistry have been discussed extensively dating to theearly 1950s.2546 This section reviews the bonding, structure,structure-property relations, and solid solutions of the diborides.

(1) Bonding and Structure

Borides have a wide range of compositions with boron:metal(B:M) ratios ranging from 1:4 to 12:1. The B:M ratio affectsboth properties and electronic structure. Changing B:M changesthe electronic structure of boron, which leads to the formationof different structural complexes containing one-, two-, andthree-dimensional (3D) B networks. Increasing the number of

Feature

D. Greencontributing editor

At UMR, portions of this work were funded by the Air Force Ofce of ScientificResearch (F49620-03-1-0072 and FA9550-06-1-0125), the National Science Foundation(DMR-0346800), and the Air Force Research Laboratory (FA8650-04-C-5704). AtNSWCCD the work was funded by the Ofce of Naval Research on several contractsmonitored by Dr. Steve Fishman.

*Member, American Ceramic Society.wAuthor to whom correspondence should be addressed. e-mail: [email protected]

Manuscript No. 22451. Received November 6, 2006; approved January 10, 2007.

Journal

J. Am. Ceram. Soc., 90 [5] 13471364 (2007)

DOI: 10.1111/j.1551-2916.2007.01583.x

r 2007 The American Ceramic Society

B atoms in a structural complex leads to increases in the BBbond strength and an increase in the stiffness of the crystal lat-tice along with increases in melting temperature (Tm), hardness(HV), strength (s), and chemical stability.TheMB bond strength in diborides depends on the degree of

electron localization around the M atoms. The valence electronconguration in isolated B atoms is 2s22p. In metal borides, theouter electron congurations are sp2 and sp3, which promotestrong covalent bonding. In diborides, B atoms are electron ac-ceptors, while the M atoms are electron donors. Each M atomdonates two electrons (one to each B), which converts M to adoubly charged cation, while B atoms become singly chargedanions. So, theMB2 formula can be expressedM

21(B)2.31,32,4045

The electron congurations vary depending on the donor prop-erties of M, which produces a diversity of crystal structure typesand properties. The MB bonds have ionic characteristics as aresult of the donoracceptor interactions, but they also havecovalent characteristics due to partial excitation of d electronsand the formation of spd hybrid congurations. The tendencyfor B atoms to form sp2 and sp3 hybrids also affects properties.

However, hardness and brittleness are lower than the corre-sponding carbides because the B structural complexes combinesp3 hybridization with the lower-strength sp2 (and even lower-strength s2p and sp) congurations, whereas the carbon atoms incarbides exhibit only sp3 hybridization.As shown in Fig. 2, the crystal structure of Group IVVI

transition metal diborides is primitive hexagonal (AlB2-type,P6/mmm space group). The unit cell contains one MB2 formulaunit. The structure is composed of layers of B atoms in 2Dgraphite-like rings or nets, which alternate with hexagonallyclose-packed M layers. Each M atom is surrounded by six equi-distant M neighbors in its plane and 12 equidistant B neighbors(six above and six below the M layer). Each B is surrounded bythree B neighbors in its plane and by six M atoms (three aboveand three below the B layer).The unit-cell parameters and interatomic distances for dibor-

ides are summarized in Table II. In general, BB separationcontrols the a-axis length. However, the a-axis length is also af-fected by MB contact. For ZrB2, which has the largest M atomof the diborides, the BB distance is 1.83 A (a/O3), which ex-ceeds the normal BB distance (1.74 A) by 0.09 A due tostretching of the BB bonds by ZrZr contact. Likewise, thesmallest M atoms (Cr, V) lead to reductions in the BB distance.From crystal chemical considerations, the length of the a-axis isa balance between two forces: (1) repulsion between atoms in theM layers and (2) attraction between atoms in the B nets.26,29,30

As a result, stable AlB2-type diborides do not form for M atomssmaller than Cr or larger than Zr. The BB bond length forminimum strain in the boron nets has been estimated to beB1.75A, which is the value for TiB2.

29

The MB distance in diborides increases linearly with theM:B radius ratio, increasing from 2.30 A for CrB2 to 2.54 A forZrB2. The MB separation is equal to a2=3 c2=41=2, which islarger than the sum of the M and B radii (Table II). Generally,the structural data indicate that bonding in the B nets controlsthe lattice parameters in the AlB2-type structure. Because the BB bonds are strong relative to the other bonds, increases in the a-axis with increasing M size are minimal. In contrast, no sucheffect is observed for the c-axis. Hence, the c:a ratio increaseswith increasing M atom size. Owing to separation of the close-packed M planes by B nets, the c-axis is always substantiallylarger than 2RM. In summary, differences among diborides re-sult from different M radii, which lead to variations in the in-teratomic bond lengths. Larger changes are observed for the c-axis than for the a-axis due to the relative bond strengths.

(2) Structure-Property Relations

Hardness, bulk modulus, Debye temperature (YD), Tm, coef-cient of thermal expansion (CTE), thermal conductivity (k), andenthalpy of formation DHof are some properties that are re-lated to bond strength (cohesive energy). Generally, the combi-nation of bonds (MM, BB, and MB) inuences the materialproperties. However, in some cases, a specific type of bond con-trols a property.31 For example, BB and MB bonds in dibor-ides control hardness and thermal stability. Hardness is,therefore, a qualitative indicator of bond strength.47

The thermal and elastic properties of diborides are summar-ized in Table III.4854 The data indicate that the Group IV di-borides have lower CTE and higher Youngs modulus andthermal conductivity than Group V diborides.26 The propertychanges suggest that BB bonds are strongest for Group IV at-oms (Ti, Zr, and Hf) and the bonds weaken as atomic numberincreases across a period of the Periodic Table. Grimvall andGuillermet46 found it useful to correlate cohesive energy and Tmtrends to the average number of valence electrons per atom forisostructural diborides. For example, the series ScB2, TiB2, VB2,CrB2, MnB2, and FeB2 have 9, 10, 11, 12, 13, and 14 valenceelectrons, respectively. With its 10 valence electrons (3.3 elec-trons per atom), TiB2 possesses the highest melting temperatureof the group. Higher or lower numbers of valence electrons re-sult in lower melting temperatures.

Fig. 1. A comparison of the melting temperatures of the most refrac-tory members of several classes of materials. Several borides, carbides,and nitrides have melting temperatures above 30001C and are consideredultra high-temperature ceramics. For comparison, the melting tempera-ture of Zr is B18501C and the melting temperature of Hf isB22271C.

Table I. Summary of Some Structural, Physical, Transport,and Thermodynamic Properties of ZrB2 and HfB2

Property ZrB2 HfB2

Crystal system space groupprototype structure

Hexagonal5

P6/mmm AlB2

Hexagonal6

P6/mmm AlB2a (A) 3.17 3.139c (A) 3.53 3.473Density (g/cm3) 6.1195 11.2126

Melting temperature (1C) 32451 33801

Youngs modulus (GPa) 4899 48011

Bulk modulus (GPa) 215 212Hardness (GPa) 239 287

Coefcient of thermalexpansion (K1)

5.9 106 7 6.3 106 7

Heat capacity at 251C(J (mol K)1)

48.28 49.57

Electrical conductivity (S/m) 1.0 107 7 9.1 106 7Thermal conductivity(W (m K)1)

607 1049

Enthalpy of formation at251C (kJ)

322.68 358.17

Free energy of formation at251C (kJ)

318.28 332.27

1348 Journal of the American Ceramic SocietyFahrenholtz et al. Vol. 90, No. 5

Hardness can also be related to electronic structure despitewide differences in published data. The hardness of diboridesdecreases as the atomic number of the metal increases forGroups IIIVI and Periods 46 of the Periodic Table. TheFermi level for ZrB2 is located in a pseudogap between thecompletely occupied bonding states and free antibonding states.As a result, ZrB2 has the maximum stability and microhardnessfor its period. Likewise, TiB2 and HfB2, which are isoelectronicand isostructural, have the highest hardness among their peri-ods. For Group V diborides, a considerable number of valenceelectrons enter antibonding states, leading to decreases in bondstrength and microhardness.35 Hardness for TiB2, ZrB2, andHfB2 decrease as temperature increases.

55

The bonding in diborides also inuences the anisotropy inproperties. Microhardness measurements for TiB2 conducted byVehldiek52 did not reveal any significant anisotropy in hardnessalong the a- or c-axes. In comparison, the Young modulusshowed a high degree of anisotropy in some diborides.53 Esti-

mation of stresses from strain-induced broadening of X-ray dif-fraction (XRD) peaks revealed that the Youngs moduli of TiB2and ZrB2 were reasonably isotropic, while NbB2 was extremelyanisotropic, with lower values observed for the [00l] direction.56

This behavior is observed in structures where the bonds areweaker in one direction than for other directions. Based on linebroadening studies, it was concluded that TiB2 had strong bond-ing in all directions, bonding in ZrB2 was isotropic, but weakerthan in TiB2, and that NbB2 was highly anisotropic with weakbonding between (00l) planes.Post et al.30 were rst to show that diborides with metals of

large radii had the highest melting temperatures. Because the BB separations are the highest and the BB bonds are the weakestfor the highest melting temperature diborides, the authors cameto the conclusion that the MB bond strength was responsiblefor melting temperatures.The CTEs of various diborides have been measured.5759 Us-

ing XRD analysis up to 16001C, Keihn and Keplin57 showedthat ZrB2 and HfB2 had lower CTE anisotropy than the dibor-ides of Ti, Nb, and Ta. The authors concluded that polycrys-talline ZrB2 and HfB2 ceramics would develop lower internalstresses during rapid cooling than other diborides. With the ex-ception of CrB2, the lattice parameters and CTEs of the dibor-ides had similar temperature dependences for the a- and c-axes,which indicates little anisotropy in the bond strength in the twodirections. The CTE in the c-direction decreases with increasingM radius, which can be correlated with an increase in the MBbond strength with increasing M size. The CTE in the a-direc-tion does not change significantly with increasing M radius.Taken together, the results can be interpreted to indicate that theBB bonds determine the cohesive forces in the a-direction,while weaker MB bonds control cohesive behavior in the c-direction.59

Fig. 2. Projections of the AlB2-type structure. Pictures are taken from the Crystal Lattice Structures Web page, provided by the Center for Com-putational Materials Science of the United States Naval Research Laboratory.

Table II. Unit Cell Parameters and Interatomic Distances inDiborides26,30

Metal

boride

Size or length (A)

c/a c a R0M BB MB 2R0M RB

ZrB2 1.114 3.530 3.169 1.61 1.830 2.54 2.48HfB2 1.105 3.470 3.141 1.58 1.813 2.51 2.46TaB2 1.041 3.225 3.097 1.49 1.788 2.41 2.36NbB2 1.05 3.27 3.11 1.48 1.79 2.43 2.36TiB2 1.064 3.228 3.028 1.47 1.748 2.38 2.34CrB2 1.030 3.066 2.969 1.30 1.714 2.30 2.18

R0M is the metal radius and RB is the radius of boron.

May 2007 Diborides of Zirconium and Hafnium 1349

(3) Solid Solubility

Mutual continuous solid solutions form among Group IV and Vdiborides. Because their radii differ by o10%, HfB2 and ZrB2form continuous solid solutions with TiB2, NbB2, and TaB2. Alarger radius difference leads to limited solubility between thesediborides and CrB2.

25,30 The activation energy for the formationof diboride solid solutions is essentially the same as the activa-tion energy forM diffusion. For example, the activation energiesfor the formation of solid solutions in the ZrB2TiB2 and TiB2NbB2 systems are 175 and 112 kJ/mol, respectively, which cor-respond to the activation energies for metal diffusion in thesesystems. In comparison, the NbB2CrB2 system exhibits highervalues for activation energy (400 kJ/mol), which can be attrib-uted to the large disparity in atomic radii between Nb and Cr(13%) compared with Zr and Ti (9%) or Ti and Nb (1%). Basedon activation energy data, solid solution formation appears todepend on diffusion of M atoms, not the strongly bonded Batoms.Some diboride solid solutions show deviations from rule of

mixture predictions for properties such as hardness and elec-trical conductivity, although some changes are only observedover limited composition ranges that are specific to each system.For example, TiB2ZrB2 solid solutions exhibited a maximum inhardness between 30 and 70 mol% ZrB2. The hardness at 50mol% ZrB2 was 24 GPa compared with values of 18 GPa forTiB2 and 12 GPa for ZrB2 as reported in the study.

60 For TiB2ZrB2 ceramics synthesized by a combustion synthesis-micro-wave process followed by hot pressing (HP), the highest valuesof relative density (99%), exure strength (680 MPa), fracturetoughness (7.3 MPa m1/2), and hardness (27 GPa) were ob-served at 20% ZrB2.

61 In another study, TiB2ZrB2 ceramicscontaining 40 mol% ZrB2 had the highest hardness (30.4 GPa),exure strength (354 MPa), fracture toughness (3.87 MPa m1/2), Youngs modulus (449 GPa), and fracture energy (16.5 J/m2).62 Local maxima in hardness have also been observed inother systems such as TiB2NbB2 (at B38 mol% NbB2),

63

NbB2CrB2 (30 GPa at 80 mol% CrB2),64,65 and TiB2CrB2

(at B20 mol%CrB2).66

The lattice constants of diboride solutions show a linear de-pendence on composition in many systems. However, based onbonding considerations discussed above, c-axis parameters aremore strongly affected by changes in composition than a-axisparameters. For TiB2WB2 solid solutions with WB2 contents

between 40 and 60 mol%, the a parameter decreased only slight-ly as WB2 content increased, whereas the c parameter exhibiteda significant decrease.67 For TiB2CrB2 and TiB2ZrB2 solu-tions, Zdaniewski68 found that the degrees of lattice contraction(for Cr) or expansion (for Zr) were smaller than expected basedon the sizes and concentrations of the M atoms. This suggestsstronger interactions between the M and B atoms than would beexpected from the M size alone and is likely due to differencesin electron congurations and bond strength.35,40 Changes inlattice anisotropy and elastic strain energy in solid solutionsmay also affect grain-boundary cohesion in polycrystalline cer-amics.68 If true, then knowledge of lattice parameter changeswith composition could be used to improve the mechanicalbehavior of diboride ceramics by using compositional changesto induce benecial stresses (for example, surface compressionstresses).Fendler et al.69 studied thermal expansion of TiB2CrB2 and

TiB2WB2 solid solutions using high-temperature XRD analy-sis. The CTEs of TiB2CrB2 did not follow predictions from ruleof mixtures calculations. The c-parameter exhibited an unex-pected increase with temperature. For TiB2WB2 solid solu-tions, the CTE was lower than TiB2 at low temperatures.Using experimental hardness values obtained from solid so-

lutions as a function of composition,61,64,66 NSWCCD research-ers calculated the BB bond lengths for compositions showingthe maximum hardness in TiB2CrB2, NbB2CrB2, and TiB2ZrB2 solid solutions. The BB bond lengths were calculated tobe 1.74, 1.73, and 1.75 A, respectively. The values were similar to1.75 A, which Hughes and Hoard29 suggested as the BB bondlength for minimum stress in the boron net. The result is alsoconsistent with hardness values from pure diborides, which werehigher for borides with the largest M atoms.

III. Synthesis

The diborides can be synthesized by a variety of routes. Thissection focuses on three main synthesis routes: (1) reductionprocesses, (2) chemical routes, and (3) reactive processes.

(1) Reduction Processes

Several reduction processes are used to synthesize diborides.Carbon and boron are the most common reducing agents, butboron carbide (B4C) or aluminum (Al) can also be used as wellas combinations of reducing agents.70 Examples of the reactionsused to synthesize diborides are shown in Table IV (Reactions(1)(5)). These reactions were written for ZrB2, but analogousprocesses produce HfB2 or other diborides.Carbothermal reduction is used to produce ZrB2 and HfB2

commercially.7173 The process is strongly endothermic; Reac-tion (1) has an enthalpy of reaction DHorxn of 1475.6 kJ at 298K. As a result, Reaction (1) is only thermodynamically favorable(i.e., DGorxn < 0 aboveB15001C (Table IV). Based on reactionfavorability and the heat absorbed, temperatures of B20001Care generally employed to synthesize diborides.71,74 The reac-tions also produce significant volumes of various gases, whichmust be removed for the reactions to proceed to completion.Excess B2O3 is often added to promote formation of diboridesover carbides. Along with B2O3 and carbides, carbon is also acommon impurity in the nal powder.70

Table III. Thermal and Elastic Parameters of TransitionMetal Borides26

Boride

CTE

(ppm/K)

YD (K)

E

(GPa)

HV(GPa)

k W (m K)1

Calc.

from aCalc

from TM

TiB2 4.8 1100 970 530 13 20.6ZrB2 6.2 765 730 420 15 18.9HfB2 6.6 550 580 16.6VB2 8.0 880 880 340 13.1NbB2 8.2 701 720 6.9TaB2 8.5 545 570 280 6.0CrB2 10.5 726 780 220 13 31.8

Table IV. Examples of Reduction Reactions that can be Used to Synthesize Borides and the Change in Free Energy of ReactionDGorxn as a Function of Temperature

Reactions Category Example DGorxn (kJ)

(1) Carbothermal ZrO2c B2O3l 5Cg ! ZrB2cr 5COg 14310.803T(2) Borothermal ZrO2c 4Bc ! ZrB2c B2O2g 3010.178T(3) Aluminothermal 3ZrO2c 3B2O3l 10All ! 3ZrB2c 5Al2O3c 241210.495T(4) Boron Carbide 7ZrO2c 5B4Cc ! 7ZrB2c 3B2O3g 5COg 13780.924T(5) Combined 2ZrO2c B4Cc 3Cg ! 2ZrB2c 4COg 11340.668T


Reduction with B4C have also been used to produce diboridesand diboride composites.7073,75 Like other reduction processes,the reactions between oxides and B4C are endothermic. How-ever, these reactions become favorable at temperatures lowerthan the corresponding reactions with carbon. For example,Reaction 4 becomes favorable around 12001C whereas Reac-tions (1), (2), and (5) are only favorable above 14001C (TableIV). Processes similar to Reaction (4) offer the additional benetof reduced levels of carbon, B2O3, and oxide impurities in thereaction products. Reaction (4) has also been exploited to pro-mote densication of B4C and ZrB2 based on the ability of B4Cto react with oxides and thereby reduce the amount of oxideimpurities in non-oxide ceramics.76,77

(2) Chemical Routes

Chemical routes used to produce diborides include solutions,reactions with boron-containing polymers, and pre-ceramicpolymers. Nanocrystalline ZrB2 and HfB2 have been synthe-sized by reacting anhydrous chlorides with sodium borohydride(NaBH4) above 5001C under pressure.

78,79 The overall process isdescribed by Reaction (6), but the reaction may proceed by avapor phase mechanism that involves the decomposition ofNaBH4 to borane (BH3) and its subsequent reaction with a gas-eous chloride. Diborides prepared by these routes can have crys-tallite sizes as small as 1020 nm78,79:

HfCl4g NaBH4g ! HfB2c 2NaClc 2HClg 3H2g (6)

Titanium diboride has also been prepared by thermal decom-position of titanium borohydride (Reaction (7)). Powders pre-pared by this method were also nano-sized with diameters of100200 nm80:

2TiBH43c ! 2TiB2c B2H6g 9H2g (7)

Nano-crystalline oxides produced by precipitation could beconverted to diborides using reduction processes (Table IV).Sacks et al.81 synthesized nano-crystalline ZrC and HfC by car-bothermal reduction of ZrO2 and HfO2 produced by a solgelprocess. Similar processes should also work for nano-crystallinediborides, but, to date, only conventional micro-sized oxideshave been converted to diborides.82 A pre-ceramic polymer hasbeen directly converted to a refractory diboride,83 but the pol-ymer was used as a binder for commercial ZrB2 powder, not toproduce a stand-alone coating or a monolithic ceramic. Finally,diboride-based composites have been prepared by combiningconventional diboride powders with SiC-bearing pre-ceramicpolymers such as polycarbosilane.84,85 From these reports, a va-riety of chemical synthesis methods are available, but the proc-essing of these powders into coatings or ceramics has not beenfully explored.

(3) Reactive Processes

The simplest reaction synthesis method for producing diboridesis the reaction of elemental precursor powders:

Zr 2B! ZrB2 (8)

Hf 2B! HfB2 (9)

The use of direct reactions dates back over 100 years, al-though pure diborides were not produced until nearly 50 yearslater due to the difculty in obtaining high-purity boron.70 Thehigh oxygen afnity of both Zr and Hf means that the reactionsmust be carried out in inert or reducing atmospheres to preventthe formation of oxide impurities.Reactions (8) and (9) are both extremely exothermic

(DHorxn 323 and 328kJ, respectively, at 298 K) and favor-

able at all temperatures.10,86 If allowed to proceed in an uncon-trolled manner, the heat can ignite a self-propagating reactionand generate temperatures high enough to promote local melt-ing of the Zr (Tm 5 B18501C) or Hf (Tm5B22251C).

87 Thus,Reactions (8) and (9) can be used to synthesize ZrB2 and HfB2by self-propagating high-temperature synthesis (SHS), which isalso known as combustion synthesis.8890 The high heating andcooling rates associated with SHS produce high defect concen-trations in the resulting diborides, which have been linked toimproved sinterability of SHS-derived powders.91 The SHSmethod also lends itself to preparation of diboride-containingcomposites or diborides doped with additives such as sinteringaids.92,93 In addition, a great deal of work has been done on SHSof TiB2 ceramics, all of which is directly applicable to ZrB2 andHfB2.In contrast to SHS, recent research at UMR has employed

slow heating (B11C/min) and extended isothermal holds (6 h at6001C) to react ne powders of Zr and B without the ignition ofSHS reactions.94 Under controlled heating, Zr and B react toform phase-pure ZrB2 at temperatures as low as 6001C. TheZrB2 crystallites produced by controlled reactions were the samesize as the Zr precursor particles. Hence, reduction of the start-ing Zr particle size before reaction reduced the size of the re-sulting ZrB2. Analysis showed that the crystallite size of ZrB2formed from Zr attrition milled for 240 min wasB10 nm. Par-ticles coarsened significantly as temperature increased and grewto more than 1 mm at 16501C (Table V). Coarsening in non-oxides ceramics, including diborides, is promoted by oxygenpresent as oxide impurities on the particle surfaces.3,77,95,96 Asignificant amount of oxygen (2.4 wt%) was detected in ZrB2produced by Reaction (8), which was higher than reported forconventional powders (0.9% for as received H. C. Starck GradeB powder, 2.0 wt% after attrition milling for 2 h in hexane).

IV. Densication

The need for dense diboride ceramics was initially driven bynuclear applications that required high neutron absorption crosssections. Later, research was driven by aerospace applications.71

Without sintering additives, full density has, historically, onlybeen achieved by HP. From the early 1950s to the 1970s, con-siderable effort was directed at densication of commerciallyavailable, non-reactive diboride powders. Research focused onways to activate sintering so that larger components with morecomplex shapes could be produced as HP has limits related toboth size and geometry. Metallic additives and liquid-phase sin-tering aids were used to enhance densication. Recent studieshave again focused on additives that promote densication,along with some modern densication techniques. As a result,progress has been made in understanding the sintering of di-borides while minimizing second phases that may be detrimentalto their high-temperature properties. Reports of fully dense di-borides, produced by reactive HP date back to 1951,70 indicatingthe viability of this approach as well.This section discusses four densication methods: (1) HP; (2)

pressureless sintering; (3) reactive routes; and (4) spark plasma

Table V. Crystallite/Grain Size and Density for ZrB2 andZrB2SiC Produced by Reactive Hot Pressing of Zr-B and

ZrBSiC Powder Mixtures

Processing

temperature (1C)

ZrB2 ZrB2SiC

Crystallite

size (nm)

Relative

density (%)

Crystallite

size (nm)

Relative

density (%)

600 10 36 10 371000 50 35 50 371450 600 34 300 401650 1000 40 500 95


sintering. Induction zone melting has also been used to densifyZrB2 and HfB2.

55,97 However, this process was not reviewedbecause the specimen size (B0.250.6 cm) only allowed forlimited microstructural analysis (grain size of B300 mm) andproperty measurements (microhardness).

(1) HP

Because of strong covalent bonding and low self-diffusion, hightemperatures and external pressures are required to densify di-borides.2 Results of early HP studies on commercially availableZrB2 and HfB2 powders are summarized in Table VI,

7,12,98104

which includes details of the starting materials, HP conditions,and nal densities. Most of the diborides were presumed to bestoichiometric (B:M5 2.0). However, Kalish et al.99 studiedboron-decient compositions (e.g., ZrB1.98 and Hf1.82). TheZrB1.98 composition was nominally in the single-phase ZrB2eld and, thus, should not have affected the HP conditions orthe resulting microstructure. Metal additions used to produceHfB1.82 should place the composition in the HfB2HfB two-phase eld. It is not clear how this change in composition mayhave affected densication and the resulting microstructure, butthe HP temperature (18001C) was well below the solidus tem-perature for the HfB2HfB binary (B21001C). Because no li-quid should form, densication would still have to proceed by adiffusion-controlled (as opposed to a liquid phase) mechanismand would be expected to occur at a temperature similar to thestoichiometric diboride. HP of both ZrB1.98 and HfB1.82 at18001C required the application of a high pressure (B827MPa) to reach full density.In historic studies, nominally stoichiometric ZrB2 and HfB2,

without additives, have only been densied by HP at 20001C orhigher with pressures of 2030 MPa,105 or at reduced tempera-tures (179018401C) with much higher pressures (8001500MPa).99,102 Recent studies (also included in Table VI) have pro-duced similar results. Nearly phase pure ZrB2 and HfB2, withaverage starting particle sizes of 510 mm, require HP above20001C to achieve full density.104 For example, HfB2 with anominal starting particle size of 10 mm was o95% dense afterHP at 21601C and 27.3 MPa for 180 min.12

Recent research at UMR showed that commercial ZrB2 withaB2 mm starting particle size (Grade B, H. C. Starck, Newton,MA, 499% purity with 0.89% O) can be further comminutedby standard techniques (e.g., attrition milling) to enhance den-sication. Particle size reduction reduced the HP temperaturenecessary to achieve full density to 19001C (pressed for 45 min at32 MPa).7 Analysis revealed no porosity, but B1.9 vol% WCwas present. The WC was incorporated during attrition millingdue to wear of the WC-based milling media. The microstructure

contained both equiaxed and slightly elongated ZrB2 grains withan average grain size ofB6 mm. Minimization of grain growthwas attributed to lower HP temperatures and reduced startingparticle sizes. This result can be contrasted with 1020 mm av-erage grain sizes that are common in diborides processed fromcoarser ZrB2 and HfB2 powders with higher HP tempera-tures.12,100,101,106

Research in the late 1960s on metal (Zr, Hf, Cr, Y, and Al)and ceramic (SiC, MoSi2, and TaB2) additions to ZrB2 and HfB2revealed the superior densication, and more importantly theimproved oxidation and thermal stress resistance (TSR), of di-borides containing SiC, C, or SiC plus C.107,108 Fenter100 re-viewed many compositions, the most important of which wereZrB2 and HfB2 with 20 vol% SiC or 18 vol% SiC and 10 vol%C. Subsequent to these key studies, recent research using com-mercial ZrB2 and HfB2 powders has typically included non-re-active additives (e.g., SiC) or liquid phase forming additions(e.g., Ni, Si3N4, or MoSi2) to improve densication. The highHP temperatures and pressures of historic studies99,102 are notnecessary with the ner starting powders and additives thatminimize grain growth, produce a liquid phase, or result in solidsolution formation. In recent studies, both ZrB2 and HfB2 havebeen hot pressed to near full density by adding Ni,109,110

Si3N4,110,111 AlN,110,112 MoSi2,

113115 SiC,7,116118 Ta5Si3,119

and other UHTCs, or combinations of additives.109,120123

Many of these additives reduced the densication temperatureof ZrB2, with Ni reducing it to as low as 16001C.

120 However,even with these additions, pressures of 2050 MPa were re-quired.

(2) Pressureless Sintering (PS)

PS of diborides would enable the fabrication of components tonear-net shape using standard powder processing methods.Compared with HP, sintering is more efcient and reducescosts, potentially opening the door for other applications. Thetwo approaches typically used to enhance densication are re-duction of starting particle size and the use of sintering aids.Potential sintering aids can be classied into categories includingliquid phase formers, solid solution formers, and reactive agents.Some discrepancy exists for the onset temperature for densi-

cation of diborides. Temperatures as low as 1300115001Cwere reported in historic studies,98,102 while more realistic tem-peratures of B17501C have been reported recently.106 The dis-crepancy is likely due to the purity of the powders. Densicationof pure diborides is due to grain boundary and volume diffu-sion, which are not appreciable until 18001C or higher.2 Thepresence of oxide or metallic impurities increases the drivingforce for densication, but also enhances grain and pore growth,

Table VI. Hot Pressing Conditions and Densities of Hot-Pressed ZrB2 and HfB2

Boride Particle size (mm)

Hot pressing conditions

Final density (%) ReferencesTemperature (1C) Pressure (MPa) Time (min)

Historic studiesZrB2 7 2100 B28 80 B95 Chown

99

ZrB1.89 1800 B827 r100 Kalish et al.100ZrB2 2000 B17 260 99.2 Fenter

101

ZrB2 22002300 2939 2030 B9572 Andrievskii et al.102

HfB2 5 1840 B793 10 r100 Kalish and Clougherty103HfB2 5 1790 B1,551 10 r100 Kalish and Clougherty103HfB1.82 5 1800 B827 r100 Kalish et al.100HfB2 2200 B23 200 98.0 Fenter

101

Recent studiesZrB2 6 1900 30 30 86.5 Bellosi and colleagues

104,105

ZrB2 o2 1900 32 45 99.8 Cutler9HfB2 B10

2160 B27 180 9095 Opeka et al.12

[], information not disclosed in the reference. 325 mesh (nominally B10 mm on average) HfB2 powder from Cerac Inc., Milwaukee, WI.


which limit the maximum achievable density to 95% or less. Inparticular, B2O3 enhances grain growth and inhibits densica-tion by promoting coarsening at temperatures below18001C.124,125

Historic studies implied that phase-pure diborides could notbe pressurelessly sintered to full density since the hexagonalcrystal structure allowed for anisotropic grain growth and en-trapped porosity (i.e., coarsening was more favorable than den-sication). However, Baumgartner and Steiger126 successfullysintered TiB2 to 499% density at 2000121001C using submi-crometer starting powders. While ne particle size enhanceddensication, Baik and Becher3 later showed that the other keyto pressureless densication of TiB2 was the total oxygen con-tent. Reducing the oxygen content to r0.5 wt% minimizedevaporationcondensation, which allowed densication to pro-ceed without significant grain or pore growth.Historic studies on PS of ZrB2 and HfB2 are limited, and

many were carried out nearly 40 years ago.127 However, in-creased densication rates were demonstrated through additionsof transition and refractory metals.127 Meerson and Gorbu-nov128 used 0.7 wt% C, along with ZrO2 and B, to producenearly fully dense ZrB2. They hypothesized that the C reducedZrO2, which subsequently reacted with the B to produce activeZrB2 nuclei on the surface of larger ZrB2 grains. Cech et al.

129

added metals to accelerate densication. They reportedB100%density for ZrB2 sintered at 22001C for 1 h in argon when adding1 wt% Re, or 0.5 wt% Cr 10.5 wt% Ti. Lattice parametermeasurements after sintering showed metal atom substitutiononto Zr sites in the ZrB2 lattice. They hypothesized that the localcontraction of the lattice affected the surface and volume freeenergies, effectively increasing the driving force for sintering.Coble and Hobbs124 showed similar results for TiB2. Kisliyet al.130 used W additions to ZrB2. They reported significantdensication, along with evidence for substitution of some Winto the ZrB2 lattice.Gropyanov131 also noted that intensive mechanical milling,

vibratory milling in this case, enhanced densication of ZrB2.He hypothesized that the milling not only increased the surfacearea, but also increased the number of point defects near thesurface of the particles, enhancing the driving force for densi-cation by grain boundary diffusion.In a recent study of pressureless sintering, Kida and Seg-

awa132,133 sintered ZrB2 to 95% relative density without externalpressure. However, densication was only possible with signif-icant amounts of BN (5 wt%), AlN (15 wt%), and SiC (5 wt%).Quabdesselam and Munir134 found that ne TiB2 produced bySHS did not have improved densication compared with pow-ders produced by carbothermal reduction. Their results demon-strated that powder processing alone does not activatedensication. Mishra et al.88 sintered ZrB2 with a nominal par-ticle size of 2.8 mm, produced by SHS, to a density of 94% at18001C for 1 h in owing argon. They noted a small amount ofZrO2 during microstructural analysis, but concluded that Feand Cr impurities promoted local melting, which improved den-sication. In a subsequent study, Mishra et al.93 focused on Feand Cr additions to ZrB2. Although Cr resulted in swelling andcracking, the addition of up to 10 wt% Fe improved density toB93% at 18001C. While transition metal additives reduce thedensication temperatures for diborides, they also cause markedreductions in melting temperature, hardness, and high-tempera-ture strength.109,128,129

Modern ZrB2 can be sintered to near full density withoutmetal additions. Sciti et al.135 sintered ZrB2 (Grade B, H. C.Starck, Germany) to full density at 18501C by adding 20 vol%MoSi2, which led to liquid phase formation. Research at UMRhas shown that the same ZrB2 can be densied using reactiveadditives.77,136 With small amounts of WC (B2 vol%), ZrB2was sintered to498% density at 21501C for 540 min. Analysisby XRD indicated that W and C were incorporated into theZrB2 lattice after reaction. Similar to the work of Baik and Be-cher3 in TiB2, densication of ZrB2 was activated when the ox-ide impurities (B2O3 and ZrO2) were removed from particle

surfaces, which minimized coarsening. Thermodynamic predic-tions followed by XRD analysis conrmed that ZrO2 could beremoved with either WC or B4C. However, B4C reacted withZrO2 at lower temperatures (B12001C) than WC (B18501C),which reduced the densication temperature (Fig. 3). The mi-crostructures of ZrB2 sintered with B4C (Fig. 4) conrm that

Fig. 3. Relative density as a function of sintering temperature for ZrB2(as-received and attrition-milled powders), with and without 4 wt% B4Cadditions. The addition of B4C led to higher relative densities at all sin-tering temperatures for both attrition-milled powder, achieving 100% ofthe theoretical density for milled ZrB2 with 4 wt% B4C sintered at18501C.

Fig. 4. Microstructure of ZrB2 sintered at 18501C for 1 h in vacuum.Specimens were prepared from attrition-milled powder with (a) 2 wt%B4C and (b) 4 wt% B4C additions.


dense ceramics can be prepared at temperatures as low as18501C.

(3) Reactive Routes to Densication

Synthesis and densication can be combined into single-step,in situ reaction and densication processes. Either elemental re-actions (Reactions (8) and (9)) or reduction processes (Reactions(1)(5)) can be used to form diborides by reactive HP (RHP). Anextremely ne crystallite size (B10 nm) was achieved for ZrB2by reacting Zr and B that were attrition milled.94 The ne crys-tallite size should enhance the driving force for sintering, asdensication is driven by minimization of surface free energy.However, HP at 21001C was required to achieve a relative den-sity of 99% for ZrB2, despite its extremely ne crystallite size,possibly due to grain coarsening below the nal densicationtemperature.94 For comparison, ZrB2 ceramics produced fromcommercially available powders with submicrometer-sized par-ticles could be hot pressed to499% density at 19001C.137 Anal-ysis of scanning electron microscopy (SEM) images revealed anaverage grain size of B12 mm for ZrB2 produced by RHP at21001C, compared with B6 mm for ZrB2 produced by conven-tional HP at 19001C.137 The crystallite size and density data inTable V conrm significant coarsening in the RHP material. Ashas been reported for other non-oxide ceramics such as TiB2,

3

B4C,138 and SiC,139 oxygen impurities in ZrB2 inhibited densi-

cation by promoting coarsening at temperatures below whichdensication could occur.Combined in situ synthesis and densication processes

have also been developed for ZrB2SiC94,140 and HfB2SiC.

141

These processes employ reactions of elemental precursors (Zr, B,and SiC)142 as well as more complex displacement reac-tions140,141,143,144:

2Zr Si B4C! 2ZrB2 SiC (10)

2 x Hf 1 x Si B4C! 2HfB2 1 x SiC xHfC (11)

The presence of SiC had a dramatic effect on both the den-sication temperature and the resulting microstructure. UsingZr, B, and SiC as precursors, a relative density of B95% wasachieved by RHP at 16501C (Table V). Not only was the den-sication temperature (16501C) reduced significantly comparedwith ZrB2 produced by RHP (21001C) or conventional HP(19001C), but the ZrB2 grain size was also much smaller. Anal-ysis of the microstructure (Fig. 5) revealed an average grain sizeof B0.5 mm for ZrB2SiC produced by RHP compared withB12 mm for dense RHP ZrB2. The reduction in processing tem-perature for ZrB2SiC by RHP has two likely causes: (1) mini-

mization of oxide impurities (B2O3, ZrO2) and (2) necrystallite/particle sizes.

(4) Spark Plasma Sintering (SPS)

SPS has recently been applied to densify UHTCs114,115,122 andother ceramics.145147 The SPS process consists of simultaneous-ly applying a uniaxial load and a direct or pulsed electric currentto a powder compact. SPS is similar to HP. However, in place ofindirect heating, the applied electrical eld heats the die and thepowder compact (if the powder is electrically conductive). Rapidheating rates (hundreds of 1C/min) facilitate rapid densication,which minimizes grain growth. As an example, Monteverdeet al.148 used SPS to achieve full density for HfB2130 vol%SiC by heating at 1001C/min to 21001C with a 2 min hold and at30 MPa. The entire SPS cycle tookB30 min, including heatingand cooling, which resulted in a uniform microstructure with amean grain size of B2 mm.Medri et al.122 compared ZrB2 containing 30 vol% ZrC and

10 vol% SiC prepared by HP and SPS. Despite the addition of3.7 vol% Si3N4 as a sintering aid, the HP specimen had a max-imum densication rate of only 1.3 105 s1 at 18701C, ul-timately reaching only B90% density. In contrast, the SPSspecimen had a nearly constant densication rate of 2 104 s1over a broad temperature range (1750120501C), resulting innear full density at 21001C ino60 min. Similar results were alsoobserved in HP versus SPS comparison studies for HfB2115vol% MoSi2,

114 ZrB2115 vol% MoSi2,115 and HfB2130 vol%

SiC12 vol% TaSi2.149

V. Microstructure and Mechanical Properties

Besides melting temperature, the properties that make ZrB2 andHfB2 attractive for high-temperature structural applications arestrength at elevated temperatures and TSR. This section focuseson ZrB2 and HfB2, as well as diborides with SiC or MoSi2 ad-ditions, from recent studies. The subsections that follow discussthe strength, modulus, hardness, toughness, and TSR of dibor-ides from recent studies.Results from historic studies are not discussed in this section

because the large starting particle size of the powders, higherimpurity levels (40.25 wt% C or Fe), or densication tempera-tures (420001C), which resulted in exaggerated grain growth.2

Large grains reduced mechanical strength, due, in part, to re-sidual thermal stresses that result from thermal expansion ani-sotropy.150 For anisotropic materials, a critical grain size existsbelow which microcracking can be avoided.151 While valueshave not been established for ZrB2 or HfB2, the values should besimilar to the critical grain size for TiB2, which is B15 mm.

152

Hence, microcracking may inuence the observed properties oflarge grained (i.e.,415 mm) diborides. In addition, many studiesused additives such as transition metals that reduce strength atelevated temperature (s1000 o 0.3sRT).103,104,109,110

(1) Strength of Diborides

Like most ceramics, higher strengths are reported for diborideswith ner grain sizes.7,114,121,153 The dependence of exurestrength on grain size (Fig. 6) demonstrates that strength in-creases as grain size decreases for ZrB2 and HfB2 based ceram-ics, with and without MoSi2 or SiC.

7,11,99101,113116,118,121,

135,140,141,148,154 As indicated in the legend, the plot includes di-borides produced by HP, RHP, SPS, and PS. Most strengthswere measured in four-point bending, but some authors usedthree-point bending. Further, specimen size varied, mostly be-tween U.S. and European standards, and, thus, the strengths arenot directly comparable. However, strengths, in general, showan inverse square root relation with grain size (GS1/2) as wouldbe expected for ceramics free from other, larger aws. A line inFig. 7 is intended to highlight an approximate GS1/2 relation-ship, not as a t to the data.

Fig. 5. A polished, thermally etched cross section of a ZrB2SiCceramic prepared by the reactive hot pressing of Zr, B, and SiC at16501C. SiC is the darker phase.


Fabrication of diborides with ne grain size and high strengthrequires processing below 20001C (e.g., HP, RHP) or limitedtime at higher temperatures (e.g., SPS). Second phases such asSiC, MoSi2, ZrC, or HfN that improve densication may alsoinhibit grain growth and improve strength.2,100,113,121 Researchat UMR has shown that hot-pressed ZrB2 (no additive) has astrength of 565 MPa (Fig. 7).7 The addition of 10, 20, or 30vol% SiC or MoSi2 reduced the average grain size toB23 mm,which increased strength to 7001000 MPa.7,113 Consideringthat the ZrB2110 vol% SiC composition was only B93.2%dense while the other compositions were fully dense,7 the resultsindicate that 10 vol% of a second phase may be enough to limitgrain growth and maximize strength at room temperature.Room temperature strengths for diborides produced by HP,SPS, and RHP support this conclusion (Fig. 8), particularly thework of Monteverde.100,103,114,115,117,118,121,123,124,141,148,149,155

Although ZrB2 grain size is a factor, the room temperaturestrength of ne-grained ZrB2SiC is not controlled by the size ofthe ZrB2 grains.

153,156 Rezaie et al.153 used linear elastic fracturemechanics (Grifth criteria) to show that the critical aw sizecorrelates strongly to the SiC particle size in ZrB230 vol% SiCceramics. This conclusion is consistent with analysis of the CTEmismatch between SiC (aavg. B4 106 K1) and HfB2 (aavg.B8 106 K1) that results in GPa-level residual stresses thatmay produce microcracks at HfB2SiC interfaces.

141 The CTEs

of ZrB2 single crystals were reported to be 6.9 106 K1 alongthe c-axis and 6.7 106 K1 along the a-axis.150 The differencein CTE values between the two axes within the hexagonal crystalstructure results in a B95 MPa thermal stress within polycrys-talline ZrB2 when cooled from a typical processing temperature(19001C).157 The addition of a-SiC (6H) particles results inlarger thermal stresses in ZrB2SiC than in polycrystallineZrB2 due to the lower CTE of SiC (4.7 106 and 4.3 10 6K1 along the c- and a-axis, respectively).158 An Eshelby anal-ysis (Eq. (12)) of ZrB2 containing SiC-particulates, assumingspherical SiC inclusions, predicts radial compressive stresseswithin the ZrB2 matrix to be as high as 2.1 GPa and tangentialtensile stresses of 4.2 GPa at ZrB2SiC interfaces

159:

sradial 2 stan

am aiDT1 2ni Ei 1 nm2EmR

r R 3 (12)

where m denotes the matrix, i denotes the inclusion, a is theCTE, n is Poissons ratio and E is Youngs modulus. The stressesdecrease as distance from the inclusion (r) increases, and in-crease as the inclusion size (R) increases.Based on this analysis, smaller SiC particles reduce the ten-

dency for microcracking and the nucleation of larger, criticalaws.151 This conclusion is also supported by the strong corre-lation between the room temperature exure strength of ZrB230 vol% SiC ceramics and the average SiC grain size, whereasstrength does not correlate to average ZrB2 grain size (Fig. 9).The data for Fig. 9 were obtained from specimens hot pressed atvarious temperatures and times using a single SiC starting par-ticle size,153 and from specimens hot pressed using SiC with dif-ferent particles sizes at the same HP temperature.156 Analysis ofthe critical aw size along with maximum ZrB2 and SiC grainsizes, conrmed that the maximum SiC grain size in ZrB2SiCceramics is the strength limiting factor.156

Second phase additions reduce densication temperature andgrain size, but the resulting ceramics may experience mechanicaldegradation at elevated temperature.117 Modern, hot-presseddiboride-based composites show a significant loss of strengthbelow 15001C (s1500 o 0.5sRT),115,118,121,123,141 which is wellbelow temperatures expected in proposed applications(420001C). Typically, oxide impurities form grain boundaryphases or become localized at triple grain junctions, which re-duce strength.121

In contrast, SPS has been used to produce HfB2SiC andHfB2MoSi2 ceramics that maintain their room temperature

Fig. 8. Room temperature exure strength as a function of the volumepercent additive (e.g., SiC, MoSi2, plus others). The shaded area is in-tended to indicate that all levels of additives increase the mechanicalstrength, but that roughly 10 vol%may be enough tomaximize strength.

Fig. 6. Room temperature exure strength as a function of grain sizefrom the diboride literature.

Fig. 7. Room temperature exure strength as a function of volumepercent SiC and MoSi2 in zirconium diboride -based ceramics producedby hot pressing at 19001C.7,14


exure strength up to 15001C.114,115,148 The electrical dischargeused to heat the specimen may promote breakdown of the non-conductive oxides on the surface of the particles, perhaps re-moving them before sintering,160 which could explain the im-proved high-temperature mechanical performance reported inrecent SPS studies. The HfB2-based ceramics produced by SPSappear to contain fewer oxide impurities (SiHfCO) than hadbeen found in ceramics produced from the same powders usingconventional HP.115,148

Figure 10 is a plot of strength as a function of test tempera-ture for HfB2SiC produced in historic studies,

100 and three re-cent HfB2-based ceramics produced by HP,

149 RHP,141 andSPS.115 Strengths were obtained in four-point bending on sim-ilar-sized test bars (B4.95 mm2 in cross-sectional area). How-ever, more strength testing is needed at elevated temperatures.Reporting strength at two temperatures does not provide com-pelling evidence for the effect of temperature (the dashed lines inFig. 10 are only included to show general trends). Second, SPS

appears to produce materials with superior strengths at elevatedtemperature (s15001C is equivalent to sRT), which merits furtherinvestigation. To utilize these materials in the proposed appli-cations, the mechanical response must be understood for tem-peratures from 15001 to 25001C (or higher).

(2) Elastic Modulus, Hardness, and Toughness

Room temperature elastic modulus (E), Vickers hardness (HV),and fracture toughness values for ZrB2 and HfB2, pure or withSiC or MoSi2 additions, are summarized in Table VII. Elasticmodulus for ZrB2 (B489 GPa)

7 from recent studies is consistentwith values from historic studies using similar sonic resonancetechniques on polycrystalline samples.161 Elastic modulus valuesfor HfB2 range from 480 to 510 GPa,

8,162 although a value ofB445 GPa was reported for porous HfB2.

12 Information onelastic constants as a function of temperature has been report-ed.162 In general, elastic moduli of diboride composites scaleswith the volume fraction of additive (ESiC 475 GPa162 andEMoSi2 440 GPa163).Hardness shows a similar trend, generally following mixing

rules in relation to the amount and type of phases included inthe particulate composites. Reported hardness values were 2123 GPa2,7 for polycrystalline ZrB2 and B28 GPa

9 for HfB2.Therefore, additions of SiC (HVB28 GPa

164) to ZrB2 result in aslight increase in hardness,7 while SiC additions to HfB2 result inlittle or no change in hardness.115 On the other hand, MoSi2 hasa low hardness (B9 GPa165) and its addition decreases the hard-ness of either ZrB2

113 or HfB2114-based ceramics.

The fracture toughness of ZrB2 and HfB2, with and withoutadditives, is generally in the range of 3.54.5 MPa m1/2. Frac-ture toughness values (Table VII) have been obtained using avariety of measurement techniques, so direct comparisons aredifcult. In a systematic study of the effect of additive content,Chamberlain et al.7 reported that fracture toughness increasedfrom 3.5 MPa m1/2 for pure ZrB2 to 5.3 MPa m1/2 for ZrB2with 30 vol% SiC. Specimens exhibited crack deection andcrack bridging (Fig. 11). The ZrB2 grains typically failed in atransgranular manner and cracks deected at or near ZrB2SiCinterfaces, leaving SiC particles in the wake of the advancingcracks.153 This result is consistent with the residual stresses pre-dicted at ZrB2SiC interfaces due to the mismatch in thermaland mechanical properties between the dispersed SiC particu-lates and the ZrB2 matrix (discussed earlier). Further increasesin fracture toughness for diboride-based composites will likelyrequire second phase additions with a higher aspect ratio (e.g.,SiC platelets or rods/whiskers)166,167 or the fabrication of lamin-ate-type architectures.168,169

(3) Thermal Stress Resistance

The use of diboride-based ceramics in applications such asrocket nozzles and leading edges may be limited by thermalstress resistance (TSR), which is poor for most ceramics. Ther-mal stresses, which develop due to the rapid heating or cooling,are controlled by the thermal and mechanical properties of thematerial.170,171 Although ve thermal shock parameters arecommonly used to describe TSR (R, R0, R0 0, R0 0 0, and R0 0 0 0),172

the R parameter (Eq. (13)) is typically used as the worst casescenario where heating or cooling occurs under conditions ofinnite heat transfer:

R DTmax s1 naE in units ofC (13)

where DTmax is the maximum allowable temperature changeunder instantaneous heating or cooling that will not initiatefracture, s is the strength, a is the CTE, n is Poissons ratio, andE is Youngs modulus. The R0 and R0 0 parameters are the prod-uct of R and the thermal conductivity (R0 in W/m) and thermaldiffusivity (R0 0, in cm2 1C/s). The R, R0, and R0 0 parameterspredict resistance to crack initiation and are maximized by in-creasing strength and decreasing Youngs modulus. The R0 0 0 and

Fig. 10. Flexure strength as a function of test temperature for hafniumdiboride (HfB2) containing 20 vol% SiC produced in the 1960s (

100) andthree recent HfB2-based ceramics with 22.130 vol% SiC additions pro-duced by hot pressing (HP) (Monteverde149), RHP (Monteverde141), andSPS (Bellosi et al.115). The dashed lines for the recent HP, reactive HP ,and Spark plasma sintering studies are only meant to draw the eye to thesuggested trend in the data.

Fig. 9. Room temperature exure strength as a function of the averagezirconium diboride (ZrB2) grain and SiC particulate size in ZrB230vol% SiC ceramics (data obtained fromRezaie et al.153 and Zhu et al.156).The linear curve t indicates a strong correlation between strength andthe size of the SiC particulates.


R0 00 0 parameters predict crack propagation resistance and can beincreased by increasing Youngs modulus and decreasingstrength.173 However, resistance to crack initiation is arguablymore important for the applications proposed for diborides be-cause of the high heating rates and steady-state heat uxes.In historic studies, R parameters of 6011201C were calculat-

ed for ZrB2 and HfB2 from room temperature up toB12001C,while the R0 parameters were 26004500 W/m.99 The diboridesoffered a clear advantage over other structural ceramics avail-able at the time (BeO, Al2O3, SiC, ZrC, and MgO), which wasdue, largely, to their strength at elevated temperatures. Thebenecial role of SiC or SiC1C additions was quantied interms of improved R0 performance, with ZrB2-based ceramicsachieving B9600 W/m under steady-state heat ux.100,174

No subsequent studies have been published that support thepotential advantages of diborides or their composites in termsof their TSR, although a recent study at UMR has focused onexperimental measurement of R using water quench tests(see Sidebar 1). The results show that modern diborides haveR parameters that are three to four times higher than historicalvalues due to increased strength. Even so, the measured DTmax

(B4001C) is well below what may be encountered in proposedapplications.One method for improving TSR is to tailor structure on mul-

tiple length scales to produce architectures that are engineered toenhance TSR while maintaining load-bearing capability such asbrous monoliths.177 Koh measured an R of 10001C for mon-olithic Si3N4, and 14001C for a brous monolithic ceramic con-sisting of Si3N4 cells and BN cell boundaries. The improvedperformance was attributed to increased resistance to crackpropagation (increased R0 0 0 0 parameter). Research at UMR onbrous monoliths with ZrB2SiC cells and CZrB2 cell bound-aries showed an improvement of B10001C in the R parametercompared with conventional materials. However, analysis didnot necessarily support the same mechanism for improved per-formance in ZrB2-based materials.

VI. Oxidation

Many proposed applications for ZrB2 and HfB2 involve oxidiz-ing conditions and elevated temperatures. While other reactiveenvironments (e.g., molten metals, propellants from rocket mo-tors) and erosive conditions are also encountered, this sectionfocuses on static oxidation. The role of additives on oxidationbehavior is also discussed. In addition to conventional furnaceoxidation and thermal gravimetry (TG) in owing air, some re-sults from arc heater testing are reviewed.

(1) Pure Diborides

The diborides undergo stoichiometric oxidation when exposedto air at elevated temperatures178:

ZrB2c 52O2g ! ZrO2c B2O3l (14)

HfB2c 52O2g ! HfO2c B2O3l (15)

Both reactions are favorable at all temperatures withDGorxn 1977 0:361T (kJ) for Reaction (14) and 2003 10.374 (kJ) for Reaction (15). Analysis by TG shows negligiblemass gain belowB7001C (Fig. 12). A similar response has beenobserved for HfB2, except that its mass gain has been reportedto be significantly lower than ZrB2 at all temperatures.

179 Attemperatures below B11001C, ZrO2 (or HfO2) and B2O3

Table VII. Room Temperature Elastic Modulus, Hardness, and Fracture Toughness Values for ZrB2- and HfB2-Based Ceramics

Composition (vol%) Modulus (GPa) Hardness (GPa) Toughness (MPa-m1/2) Reference

ZrB2 489493 2123 3.54.22,7,100

ZrB2110 SiC 450507 24 4.14.87,118

ZrB2120 SiC 466531 24 4.47,100

ZrB2130 SiC 484 24 5.37

ZrB2110 MoSi2 516 20.4 4.1113

ZrB2115 MoSi2 479 14.916.2 3.34.4115

ZrB2120 MoSi2 489523 1618.5 2.33113,135

ZrB2130 MoSi2 494 17.7 4113

ZrB2115 a-SiC14.5 ZrN 467 15.6 5123

ZrB2135 HfB2110 a-SiC14.5 ZrN 494 16.7 4.8123

ZrB2137.5 HfB2119.5 a-SiC13HfN 497 22 121

HfB2 445 12

HfB2120 SiC 544 17.25 4.15116

HfB2130 SiC 512 26 3.9115,148

HfB2115 MoSi2 519530 15.720.9 3.773.82114

HfB2120 SiC13HfN 506 22.3 3.8115

HfB2122.1 SiC15.9 HfC 520 19 141

HfB2130 SiC12 TaSi2 489506 3.64.65149

Fig. 11. A thermally etched cross section of zirconium diboride (ZrB2)30 vol% SiC hot pressed at 20501C for 90 min in argon. The imageshows a Vickers indentation crack path with inset arrows indicatingpredominantly transgranular fracture for the ZrB2 grains and crack de-ection near the ZrB2SiC interfaces.


form a continuous layer that provides passive oxidation protec-tion.179184 Analysis has concluded that the rate limiting step foroxidation is the transport of oxygen through B2O3, which resultsin parabolic (diffusion-limited growth) kinetics for mass gainand the oxide layer thickness, whether the reaction rate wasmeasured by mass gain, oxygen uptake from the atmosphere, orreaction layer thickness.180,185

Between B11001 andB14001C, the weight change reects acombination of mass loss due to B2O3 evaporation and mass

gain due to the formation of condensed oxides.182,185 Specimenscontinue to gain mass as the mass of oxide (ZrO2 or HfO2)formed is greater than the mass of diboride reacted plus themass of B2O3 lost. As B2O3 evaporates, a porous ZrO2, layerremains, although a small amount of B2O3 may be retained.

143

Above B14001C, the oxide layer is not protective and rapid,linear mass gain kinetics have been reported.180 The TG curvefor ZrB2 in Fig. 12 has an increasing slope, reecting morerapid mass gain at higher temperatures. In addition to the

Sidebar 1. Thermal Shock Behavior of ZrB2-Based Ceramics

A major challenge for ceramics in thermostructural applica-tions is their poor resistance to thermal shock. The primaryfailure mechanism for pristine monolithic ZrB2 during ther-mal shock is crack initiation when the stresses imposed by athermal gradient exceed the strength of the material. Themaximum temperature change (i.e., the R thermal shockparameter) can be calculated using Eq. (S1) and measuredproperties.

R s 1 n E a

c S1

where r is the fracture strength, n is Poissons ratio, E is theYoung modulus, a is the coefcient of thermal expansion, andW is a stress reduction factor. Based on the properties listed inTable S1, the calculated R parameters for ZrB2 and ZrB230% SiC were 1401 and 2001C, respectively. The values can becompared with a calculated R-parameter ofB811C for mon-olithic ZrB2 from historic studies where the fracture strengthswere considerably lower (B300 MPa vs. B823 MPa for thecurrent study). The lower strength of the historic material wasdue to larger grain sizes from higher HP temperatures.175

The experimental thermal shock temperature (R) has beendetermined for ZrB2 and ZrB230 vol% SiC using waterquench thermal shock testing (ASTM C1525) on size B(3 mm 4 mm 45 mm) test bars. The thermal shock tem-peratures for ZrB2 and ZrB230% SiC were 3851 and 3951C(Figure S1). A stress reduction factor (W) of 0.36 was requiredto equate the experimental R for ZrB2 to the calculated value.If the sameW were assumed for ZrB2SiC, the rened R forZrB2SiC would be 1441C, which is about the same as ZrB2,compared with a calculated value of 1901C. The discrepancybetween the calculated and rened R values for ZrB2SiC maybe due to microcracking within the particulate composite,which has been reported for a similar diboride material.176

In addition to conventional materials, R values were meas-ured for ZrB2-based brous monoliths. As shown in FigureS2, a brous monolith was fabricated with cells composed ofZrB230 vol% SiC and cell boundaries composed of graphite15 vol% ZrB2. In contrast to the conventional materials, thebrous monolithic material had an experimental R parameterofB14001C (Figure S1), which is B350% greater than ZrB2or ZrB2SiC. The strength of the brous monolithic materialremained constant for T values up toB14001C, which repre-sents a marked improvement over the conventional ceramicsthat lost strength for T values ofB4001C. Previously, Si3N4/BN brous monoliths have also demonstrated a significantimprovement in thermal shock behavior, although the Si3N4/BN material had a gradual loss in strength as quench tem-perature increased that was not observed in the ZrB2-basedbrous monolith.177 The improved thermal shock resistanceof the ZrB2-based brous monolith compared with conven-tional ZrB2 and ZrB2SiC has been attributed to the brousmonolithic structure, which limits the initiation and propaga-tion of cracks to the cell boundaries, allowing the strong cellsto remain pristine.

Table S1. Material Properties and Thermal ShockParameters for ZrB2 and ZrB230 vol% SiC

Material property ZrB2 ZrB230 vol% SiC

Density (g/cm3) 6.27 5.33Flexure strength (MPa) 568 823Youngs modulus (GPa) 507 503Assumed Poissons ratio 0.16 0.16Calculated R (1C) 140 200Experimental R (1C) B385 B395Stress reduction factor 0.364

(calculated)0.364

(assumed)Rened R 140 144

Fig. S1. Flexure strength as a function of change in temperature forwater quench thermal testing (ASTM C1525) for ZrB2, ZrB230 vol%SiC and ZrB230 vol% SiC/graphite15 vol%ZrB2 brousmonoliths.

Fig. S2. Scanning electron microscopy image showing the macro-structure of a ZrB2SiC/graphiteZrB2 brous monolith. The archi-tecture consists of ZrB230 vol% SiC cells with a graphite15 vol%ZrB2 cell boundary.


experimental studies, the oxidation behavior of ZrB2 has beendescribed using various thermodynamic models that focus onoxidation in air at 15001C or below.21,186,187

(2) The Effect of Additives

Additives that alter the composition of the oxide scale improvethe oxidation resistance of ZrB2 and HfB2. By far, the mostcommon additive is SiC, which reduces the oxidation rate forboth ZrB2 and HfB2 by forming a silica-rich scale.

177,188191

Analysis by TG (Fig. 12) shows that ZrB2SiC had a normalizedmass gain of B0.02 mg/mm2 when heated to 15001C in aircompared with a mass gain ofB0.12 mg/mm2 for ZrB2. A SiCcontent of 20 vol% has been studied extensively based on his-toric studies indicating that this composition had the best com-bination of oxidation resistance and mechanical behavior.107

Below B11001C, the addition of SiC does not alter the oxida-tion behavior of the diborides.182 In this temperature regime, theoxidation rate of SiC is several orders of magnitude slower thanthat of the diborides.190 Consequently, the oxide scale on ZrB2SiC below 11001C consists of ZrO2 and B2O3, as it did for pureZrB2.

189 Above B11001C, two factors affect oxidation. First,the rate of SiC oxidation increases and the SiC particles areconverted to SiO2 plus CO or CO2. Second, the rate of B2O3evaporation becomes significant. As shown in Fig. 12, ZrB2SiCshows a mass loss between 12001 and 13001C due to B2O3 evap-oration. The silica-rich layer provides protective behavior, whichresults in mass gain with parabolic kinetics from room tempera-ture up to at least 16001C.21 Analysis of the outer oxide layerformed at 15001C has detected less than 1 wt% B, indicatingthat nearly all of the B2O3 has evaporated by this tempera-ture.192 The addition of SiC not only extends the temperaturerange of protective behavior, but it also imparts the ability torapidly regain protective behavior after the loss of protectiondue to excessive temperature, removal of oxide by shear forces,or other causes.Both ZrB2SiC and HfB2SiC ceramics exhibit passive oxi-

dation protection with parabolic mass gain kinetics over a widerange of temperatures. The oxidation rate is controlled by dif-fusion of oxygen through the outer oxide scale. Most authorsalso report the formation of a layer containing both ZrO2/HfO2and SiO2 beneath the outer layer. Often, this region is thin com-pared with the outer silica-rich scale, but thicker layers havebeen reported after thermal cycling.153 Below the ZrO2SiO2(HfO2SiO2) layer, some authors report the formation of a po-rous region from which SiC has been depleted.153,190,193195 Thisregion has been reported to form at temperatures of 15001C orabove and it contains ZrO2 (HfO2), ZrB2 (HfB2), or both. In dryair, thermodynamic modeling suggested that SiC was removedby active oxidation at the low oxygen partial pressures that arethought to exist under the scale.196 Subsequent experimentalstudies conrmed that a porous ZrO2 layer was formed whenZrB2SiC was oxidized at 15001C in an oxygen partial pressureof B1010 Pa. Other compounds such as SiC, SiO2, or B2O3were not detected in the layer.197

Besides SiC, additives such as MoSi2,113,198 tantalum com-

pounds,119,199 ZrSi2,195 and other diborides21,119,200 improve the

Fig. 13. Cross-sectional scanning electron microscopy image of zirconium diboride (ZrB2)-SiC after arc heater testing at 350 W/cm2 showing (a) the

mixed oxide layer, (b) the SiC-depleted layer, and (c) the underlying ZrB2SiC.

Fig. 12. Weight change measured by thermal gravimetric analysis(TGA) for (a) zirconium diboride (ZrB2) and (b) ZrB2 containing 20vol% SiC heated at 101C/min to 15001C in air.


oxidation resistance of diborides either alone or in combinationwith SiC. In particular, the addition of Ta compounds has beenshown to improve the oxidation resistance of ZrB2SiC.119,195,201 These additions alter the composition of the outerglassy layer, which can lead to phase separation of the liquid/glass due to the high cation eld strength of the transition met-als.119,201 In the case of Ta additions, the phase-separated glassylayers resulted in reduced oxidation rates in TG studies withZrB2SiC containing TaB2 having the lowest oxidation rate ofany of the materials evaluated.201

(3) Arc Heater Testing

Although less common than TG or furnace oxidation studies,arc heaterz testing is widely recognized as more representative ofthe environment that will be encountered in hypersonic ight oratmospheric re-entry because of its high heat uxes, low pressuredissociated atmosphere (i.e., O instead of O2), and high gasvelocities.116,196,195,201 Despite the drastic differences in theexperimental conditions, the response of boride-based ceramicsto arc heater testing appears to be similar to conventional stud-ies. After testing at a heat ux of 350 W/cm2 for 10 min (surfacemodied temperature B18001C), ZrB2SiC formed a surfacelayer of ZrO2, a mixed oxide layer, and a SiC-depleted layer(Fig. 13).196 HfB2SiC showed a similar response under thesame test conditions.116 A solid solution modied composition,(Zr0.9,Cr0.1)B2SiC, tested by NSWCCD at a heat ux of 326W/cm2 showed a reduced erosion rate compared with othermaterials. For both ZrB2-SiC and HfB2SiC, the denseoxide layer (ZrO2 or HfO2) that formed on the surfacecontained a small volume fraction of holes thought to allowgaseous reaction products (i.e., CO, CO2, boron and siliconoxides) to escape without disrupting large sections of thescale.116,196 Based on these arc heater evaluations, ZrB2SiCand HfB2SiC exhibit protective behavior under a wide range oftesting conditions.

VII. Summary

Historic and modern studies of the crystal chemistry, synthesis,densication, microstructure, mechanical properties, and oxida-tion of ZrB2 and HfB2 ceramics were reviewed. Zirconium andhafnium diborides have the AlB2 structure (P6/mmm) that iscomposed of alternating planes of hexagonally close-packed Zratoms and graphite-like B atoms. These compounds exhibit par-tial or complete solubility with other transition metal diborides,which allows properties such as hardness and CTE to be con-trolled through the application of crystal chemical principles.Commercial diboride powders, with particle sizes as ne as 2mm, are commonly synthesized by carbothermal reduction.Nano-crystalline powders can be synthesized using solution pre-cursors, exothermic reactions, or pre-ceramic polymers. WhileHP is the dominant densication technique, recent advanceshave led to pressureless densication processes through solid-state or liquid-phase sintering. Generally, the strength of the di-borides increases as grain size decreases with strengths as high asB500 MPa reported for materials with grain sizes of B5 mm.The addition of second phases such as SiC can increase strengthto over 1 GPa by further reducing grain size and promoting fa-vorable residual thermal stresses. Further, SPS has been used toproduce diboride-based materials that retain their room tem-perature strength to at least 15001C. Progress has also beenmade with regard to understanding the behavior of diboride-based ceramics in static oxidation as well as arc heater testing,which more accurately reproduces environments encounteredduring hypersonic ight and atmospheric re-entry.Although significant gains have been made, several critical

areas have been identied for future research. In particular, re-search is needed to understand oxidation mechanisms and fun-damental structure-property relations at elevated temperatures,

if these materials are to be used in hypersonic, atmospheric re-entry, or rocket propulsion applications.

Acknowledgments

The effort to write this paper was catalyzed by the NSF-AFOSR Joint Work-shop on Future Ultra High Temperature Materials held at the National ScienceFoundation in January 2004. Dr. Lynnette Madsen, Ceramics Program Directorat NSF, and Dr. Joan Fuller, Program Manager for Ceramics and Non-MetallicMaterials at AFOSR, provided the vision for the workshop, which was funded onNSF grant DMR-0403004. G. E. H. and W. G. F. gratefully acknowledge con-tributions of current and former members of the UMR UHTC research group.

References1A. E. McHale (ed.) Data Collected from Phase Diagrams for Ceramists, Vol. X.

American Ceramic Society, Westerville, OH, 1994.2R. Telle, L. S. Sigl, and K. Takagi, Boride-Based Hard Materials; pp. 802

945 in Handbook of Ceramic Hard Materials, Edited by R. Riedel. Wiley-VCH,Weinheim, Germany, 2000.

3S. Baik and P. F. Becher, Effect of Oxygen on the Densication of TiB2,J. Am. Ceram. Soc., 70 [8] 52730 (1987).

4R. G. Munro, Material Properties of Titanium Diboride, J. Res. Natl. Inst.Stand. Technol., 105 [5] 70920 (2000).

5Card number 75-1050, Powder Diffraction File, International Centre forDiffraction Data.

6Card number 89-3651, Powder Diffraction File, International Centre forDiffraction Data.

7A. L. Chamberlain, W. G. Fahrenholtz, G. E. Hilmas, and D. T. Ellerby,High Strength ZrB2-Based Ceramics, J. Am. Ceram. Soc., 87 [6] 11702 (2004).

8P. Vajeeston, P. Ravindran, C. Ravi, and R. Asokamani, Electronic Structure,Bonding, and Ground State Properties of AlB2-Type TransitionMetal Diborides,Phys. Rev. B, 63 [4] 04115111 (2001).

9R. A. Cutler, Engineering Properties of Borides; pp. 787803 inCeramics andGlasses: Engineered Materials Handbook, Vol. 4, Edited by S. J. Schneider Jr.ASM International, Materials Park, OH, 1991.

10M. W. Jr. Chase, NIST-JANAF Thermochemical Tables. American ChemicalSociety and the American Institute of Physics, Woodbury, NY, 1998.

11E. Wuchina, M. Opeka, S. Causey, K. Buesking, J. Spain, A. Cull, J. Rout-bort, and F. Guitierrez-Mora, Designing for Ultrahigh-Temperature Applica-tions: The Mechanical and Thermal Properties of HfB2, HfCx, and a-Hf(N),J. Mater. Sci., 39 [19] 593949 (2004).

12M. M. Opeka, I. G. Talmy, E. J. Wuchina, J. A. Zaykoski, and S. J. Causey,Mechanical, Thermal, and Oxidation Properties of Hafnium and ZirconiumCompounds, J. Eur. Ceram. Soc., 19 [1314] 240514 (1999).

13K. Kuwabara, S. Sakamoto, O. Kida, T. Ishino, T. Kodama, H. Nakajima,T. Ito, and Y. Hirakawa, Corrosion Resistance and Electrical Resistivity of ZrB2Monolithic Refractories; pp. 3025 in The Proceedings of UNITECR 2003. The8th Biennial Worldwide Conference on Refractories, Osaka, Japan. Edited by K.Asano. Technical Association of Refractories, Japan, 2003.

14O. Kida, Monolithic Refractory Material and Waste Melting Furnace Usingthe Same, Japanese Patent JP2000335969, May 12, 2000.

15N. Kaji, H. Shikano, and I. Tanaka, Development of ZrB2-Graphite Pro-tective Sleeve for Submerged Nozzle, Taikabutsu Overseas, 14 [2] 3943 (1992).

16B. Stucker, W. Bradley, P. T. Eubank, S. Norasetthekul, and B. Bozkurt,Zirconium Diboride/Copper EDM Electrodes From Selective Laser Sintering,Solid Freeform Fabric. Symp. Proc., 1, 25765 (1997).

17Z. J. Jin, M. Zhang, D. M. Guo, and R. K. Kang, Electroforming of Copper/ZrB2 Composite Coating and Its Performance as Electro-Discharge MachiningElectrodes, Key Eng. Mater., 291292, 53742 (2005).

18B. Stucker, M. Malhorta, and Z. Qu, Recent Developments in the SelectiveLaser Sintering of Zykron EDMElectrodes; in Proceedings of the Laser MaterialsProcessing Conference 1999, San Diego, CA. 87(2) F/158-F/161. Laser Institute ofAmerica, 2000.

19J. Sung, D. M. Goedde, G. S. Girolami, and J. R. Abelson, Remote-PlasmaChemical Vapor Deposition of Conformal ZrB2 Films at Low Temperature: APromising Diffusion Barrier for Ultralarge Scale Integrated Electronics, J. Appl.Phys., 91 [6] 390411 (2002).

20Y. Murata, Cutting Tool Tips and Ceramics Containing Hafnium Nitrideand Zirconium Diboride, U.S. Patent, 3, 487594 (1970).

21M. M. Opeka, I. G. Talmy, and J. A. Zaykoski, Oxidation-Based MaterialsSelection for 20001C1 Hypersonic Aerosurfaces: Theoretical Considerations andHistorical Experience, J. Mater. Sci., 39 [19] 5887904 (2004).

22D. M. Van Wie, D. G. Jr. Drewry, D. E. King, and C. M. Hudson, TheHypersonic Environment: Required Operating Conditions and Design Challeng-es, J. Mater. Sci., 39 [19] 591524 (2004).

23M. E. White and W. R. Price, Affordable Hypersonic Missiles forLong-Range Precision Strike, J. Hopkins APL Tech. Digest, 20 [3] 41523(1999).

24T. A. Jackson, D. R. Eklund, and A. J. Fink, High Speed Propulsion: Per-formance Advantage of Advanced Materials, J. Mater. Sci., 39 [19] 590513(2004).

25G. V. Samsonov, L. Ya. Markovsky, A. F. Zhigach, and M. G. Valyashko,Boron, Its Compounds and Alloys, Akad. Nauk Ukrain. SSR. Kiev., (1960)(in Russian).

26G. V. Samsonov, T. I. Serebryakova, and V. A. Neronov, Borides. Atomizdat,Moscow, 1975.

zA general description of arc heater testing can be found at http://thermo-physics.arc.nasa.gov/arcjet.htm


27G. V. Samsonov and K. I. Portnoi. Alloys Based on High-TemperatureCompounds (Splavy na Osnove Tugoplavkikh Soedineniy), Oboronnoi Prom.,Moscow, (1961) (English Translation).

28G. V. Samsonov and B. A. Kovenskaya, The Nature of the Chemical Bondin Borides; pp. 1930 in Boron and Refractory Borides, Edited by V. I. Marko-vich. Springer-Verlag, Berlin, 1977.

29J. L. Hoard and R. E. Hughes, Elemental Boron and Compounds of HighBoron Content: Structure, Properties, and Polymorphism; pp. 13141 in TheChemistry of Boron and Its Compounds, Edited by E. L. Muetterties. John Wi-ley&Sons Inc., New York, 1967.

30B. Post, F. W. Glaser, and D.Moskowitz, TransitionMetal Diborides, ActaMetall., 2 [1] 205 (1954).

31Y. B. Kuzma and S. I. Mikhalenko, Crystal Chemistry of RefractoryCompounds, Zh. Vses. Khim. Obshch. Mendeleeva, 30 [6] 50914 (1985).

32Y. B. Kuzma Crystal Chemistry of Borides (Kristallokhimia boridov),Vishcha Skola, Lvov, USSR, 1983.

33A. L. Ivanovskii, N. E. Medvedeva, G. P. Shveikin, Y. E. Medvedeva, and A.E. Nikiforov, First Principles Studies of the Stability and Electronic Properties ofMetal Borides. 1. Diborides of 3d Metals, Metall. Nov. Takhnol., 20 [11] 419(1998).

34A. L. Ivanovskii, N. E. Medvedeva, and Y. Y. Medvedeva, Ab Initio Studiesof the Stability and Electronic Properties of Metal Borides. II. Diborides of 4d and5d Metals, Metall. Nov. Tekhnol., 21 [12] 1933 (1999).

35I. R. Shein and A. L. Ivanovskii, The Band Structure of Hexagonal Dibor-ides ZrB2, VB2, NbB2 and TaB2 in Comparison With Superconducting MgB2,Phys. Solid State (Translation of Fizika Tverdogo Tela (Sankt-Peterburg)), 44 [10]18339 (2002).

36R. Kiessling, Manganese Borides, Acta Chem. Scand., 4, 14659 (1950).37G. Haigg and R. Kiessling, Distribution Equilibria in some Ternary Systems

M1M2B and the Relative Strength of the Transition MetalBoron Bonds,J. Inst. Metals, 81, 5760 (1952).

38K. E. Spear, Chemical Bonding in AlB2-type Borides, J. Less-CommonMetals, 47, 195201 (1976).

39V. Grigorovich, Sintering From Aspect of Interatomic Bonds, Sci.Sintering, 24 [3] 15560 (1992).

40W. N. Lipscomb and D. Britton, Valence Structure of Higher Borides,J. Chem. Phys., 33, 27580 (1960).

41I. I. Lyakhovskaya, T. M. Zimkina, and V. A. Fomichev, K-spectra of Boronin Transition Metal Diborides and in Compounds LaB6, BaB6, and AsB, Fyz.Tverd. Tela (Solid State Phys.), 12 [1] 17480 (1970).

42S. A. Nemnonov and K. M. Kolobova, X-Ray Spectra, Electron Structure,and Properties of Titanium Compounds, Fiz. Metall. Metalloved. (Phys. MetalsSci. Metals), 22 [5] 68092 (1966).

43D. W. Fischer and W. L. Baun, Band Structure and the Titanium L (sub II,III) X-ray Emission and Absorption Spectra from Pure Metal, Oxides, Nitride,Carbide, and Boride, J. Appl. Phys., 39 [10] 475776 (1968).

44S. A. Nemnonov, A. Z. Menshikov, K.M. Kolobova, E. Z. Kurmayev, and V.A. Trapeznikov, Study of the Electronic Structure and Interatomic Bonds inSome Compounds and Binary Alloys by the Method of X-Ray Spectroscopy,Trans. Metall. Soc. AIME, 245 [6] 11918 (1969).

45A. H. Silver and P. J. Bray, NMR Study of Bonding in Some Solid BoronCompounds, J. Chem. Phys., 32 [1] 28892 (1960).

46G. Grimvall and A. F. Guillermet, Phase Stability Properties of TransitionMetal Diborides; pp. 42330 in Boron-Rich Solids, AIP Conference Proceedings,Vol. 231, Edited by D. Emin, T. L. Aselage, A. C. Switendick, B. Morosin, andC. L. Beckel. American Institute of Physics, New York, 1990.

47T. Lundstroem, Structure, Defects, and Properties of Some Refractory Bor-ides, Pure Appl. Chem., 57 [10] 138390 (1985).

48E. M. Savitskii, N. I. Timofeeva, and V. I. Bakarinova, Chemical Stability ofHafnium Diboride, Izv. Akad. Nauk SSSR Neorganich. Mater., 6 [1] 1201(1970).

49A. A. Ivanko, Handbook of Hardness Data. Naukova Dumka, Kiev, 1968.50E. V. Clougherty and R. L. Pober, Physical and Mechanical Properties of

Transition Metal Diborides, IMD Spec. Rep. Ser., 10 [13] 42344 (1964).51R. Kiessling, The Crystal Structure of Molybdenum and Tungsten Borides,

Acta Chem. Scand., 1, 893916 (1947).52F. W. Vahldiek, Electrical Resistivity, Elastic Modulus and Debye

Temperature of Titanium Diboride, J. Less-Common Metal, 12 [3] 2029(1967).

53J. J. Gilman and B. W. Roberts, Elastic Constants of TiC and TiB2, J. Appl.Phys., 32 [7] 1405 (1961).

54G. V. Samsonov and V. S. Neshpor, Some Physical Properties ofCermets, Vop. Poroshkovoi Met. I Prochnosti Mater. Akad. Nauk Ukr. SSR, 5,334 (1958).

55R. D. Koester and D. P. Moak, Hot Hardness of Selected Borides, Oxidesand Carbides to 19001C, J. Am. Ceram. Soc., 50 [6] 2906 (1967).

56D. C. Gillies and D. Lewis, Bond Strength in Diborides of Some Group IVand V Metals, J. Less-Common Metals, 16 [2] 1623 (1968).

57F. G. Keihn and E. J. Keplin, High-Temperature Thermal Expansion ofCertain Group IV and Group V Diborides, J. Am. Cer. Soc., 50 [2] 814(1967).

58G. V. Samsonov, Thermal Expansion of Diborides of Group IV and VTransition Metals, Tepl. Vysok. Temp., 9 [1] 1957 (1971).

59B. Lonnberg, Thermal Expansion Studies on the Group IVVII TransitionMetal Diborides, J. Less-Common Metals, 141 [1] 14556 (1988).

60C. Mroz, Processing TiZrC and TiZrB2, Am. Ceram. Soc. Bull., 73 [4] 7881(1994).

61H. B. Lee, Y. J. Yoon, Y. K. Oh, and Z. S. Ahn, Study on Synthesis andCharacterization of TiZrB2 Composite by Microwave SHS, J. Kor. Ceram. Soc,36 [1] 714 (1999).

62M. Moriyama, H. Aoki, and Y. Kobayashi, Fabrication and MechanicalProperties of Hot-Pressed TiB2ZrB2 Ceramic System, J. Ceram. Soc. Jpn., 106[12] 1196200 (1998).

63G. V. Samsonov and V. S. Neshpor, Investigation of the Laws of Formationof Isomorphous Boride Alloys, Zh. Fiz. Khim., 29, 83945 (1955).

64I. G. Talmy, E. J. Wuchina, J. A. Zaykoski, and M. M. Opeka, Properties ofCeramics in the NbB2CrB2 System, Ceram. Eng. Sci. Proc., 17 [3] 12835 (1996).

65I. G. Talmy, E. J. Wuchina, J. A. Zaykoski, and M. M. Opeka, Ceramics inthe NbB2CrB2 System; pp. 817 in Ceramic Matrix Composites- Advanced HighTemperature Structural Materials, Materials Research Society Symposium Pro-ceedings, Vol. 365, Edited by R. A. Lowden, M. K. Ferber, J. R. Hellmann, K. K.Chawla, and S. P. DiPietro. Materials Research Society, Warrendale, PA, 1995.

66G. A. Meerson, G. V. Samsonov, R. B. Kotelnikov, M. S. Voinova, I. P.Evteeva, and S. D. Krasnenkova, Certain Properties of Alloys in the SystemsTiB2CrB, TiB2W2B5, and ZrBCrB2, Sbornik Nauch. Trudov. Nauch.-Tekh.Obshchestva Tsvetnoi Met. Moskov. Inst. Tsvetn. Metal. i Zolota, 29, 3238(1959).

67M. Shibuya, M. Kawata, M. Ohyanagi, and Z. A. Munir, Titanium Dibo-rideTungsten Diboride Solid Solutions Formed by Induction-Field-ActivatedCombustion Synthesis, J. Am. Ceram. Soc., 86 [4] 70610 (2003).

68W. A. Zdaniewski, Solid Solubility Effect on Properties of Titanium Dibo-ride, J. Am. Ceram. Soc., 70 [11] 7937 (1987).

69E. Fendler, O. Babushkin, T. Lindbaeck, R. Telle, and G. Petzow, ThermalExpansion of Diboride Solid Solutions; pp. 204212 in Proceedings of the 4thInternational Symposium on Ceramic Materials and Components for Engines, June1012, 1991, Goteborg. Edited by R. Carlsson, T. Johannson, and L. Kahlman.Elsevier Applied Science, London, 1992.

70P. Schwarzkopf and R. Kieffer, Refractory Hard Metals. Chapter 26. Mac-Millan Co., New York, 1953.

71R. Thomson, Production, Fabrication, and Uses of Borides; pp. 11320 inThe Physics and Chemistry of Carbides, Nitrides and Borides, Edited by R. Freer.Kluwer Academic Publishers, Dordrecht, 1990.

72T. Lundstrom, Transition Metal Borides; pp. 35176 in Boron andRefractory Borides, Edited by V. I. Markovich. Springer-Verlag, Berlin, 1977.

73C. Mroz, Zirconium Diboride, Am. Ceram. Soc. Bull., 73 [6] 1412 (1994).74C. Mroz, Processing and Properties of Microcomposite TiZrC and TiZrB2

Materials, Ceram. Eng. Sci. Proc., 14 [910] 72535 (1993).75H. Zhao, Y. He, and Z. Jin, Preparation of ZrB2 Powder, J. Am. Ceram.

Soc., 78 [9] 25346 (1995).76A. Goldstein, Y. Geffen, and A. Goldenberg, Boron Carbide-Zirconium

Diboride In-Situ Composites by the Reactive Pressureless Sintering of BoronCarbide-Zirconia Mixtures, J. Am. Ceram. Soc., 84 [3] 6424 (2001).

77S. C. Zhang, G. E. Hilmas, and W. G. Fahrenholtz, Pressureless Densica-tion of Zirconium Diboride With Boron Carbide Additions, J. Am. Ceram. Soc.,89 [5] 154450 (2006).

78L. Chen, Y. Gu, L. Shi, Z. Yang, J. Ma, and Y. Qian, Synthesis andOxidation of Nanocrystalline HfB2, J. Alloys Comp., 368 [12] 3536 (2004).

79L. Chen, Y. Gu, Z. Yang, L. Shi, J. Ma, and Y. Qian, Preparation and SomeProperties of Nanocrystalline ZrB2 Powders, Script. Mater., 50 [7] 95961 (2004).

80D. L. Segal, Chemical Routes for the Preparation of Powders; pp. 311 inThe Physics and Chemistry of Carbides, Nitrides, and Borides, Edited

j_002

Documents