Journal of Alloys and Compounds 566 (2013) 125–130

Contents lists available at SciVerse ScienceDirect

Journal of Alloys and Compounds

journal homepage: www.elsevier.com/locate/jalcom

Ablation behavior of ZrB2–SiC sharp leading edges

Xinxin Jin, Rujie He ⇑, Xinghong Zhang, Ping Hu
Science and Technology on Advanced Composites in Special Environments Laboratory, Harbin Institute of Technology, Harbin 150001, PR China

a r t i c l e i n f o a b s t r a c t

Article history: The ablation behavior of ZrB2–SiC sharp leading edges with different radius (R) values of 0.15 (R1), 0.5
Received 18 January 2013 (R2), 1.0 (R3) and 1.5 mm (R4) was investigated using an oxy-acetylene torch. Under the same ablation
Received in revised form 7 March 2013 condition, sharp leading edge with a lower R value underwent a severer ablation. The surface tempera-
Accepted 8 March 2013
ture reached a maximum of 2100 °C for R1, 2040 °C for R2, 2000 °C for R3 and 1930 °C for R4,
Available online 20 March 2013
respectively, and subsequently decreased to a state value of about 1910–1935 °C for all models. R1
and R2 underwent intensive mass loss and linear shrinkage, whereas R3 had only a slight mass loss
Keywords:
and linear shrinkage, and R4 had a slight mass gain and linear expansion. The finite element analysis
Composite materials
Oxidation
(FEA) also gave the simulated temperature distributions in the ZrB2–SiC sharp leading edge models,
Microstructure which were in good accordance with the experimental results. Considering the surface temperature,
shape retention behavior and the lift to drag ratio, R3 (1.0 mm) was therefore selected as the optimal
radius value for ZrB2–SiC sharp leading edge. The microstructures of the cross-section and oxidation sur-
face of R3 were also investigated.
Ó 2013 Elsevier B.V. All rights reserved.

1. Introduction can undergo lower ablation. Nevertheless, to our best knowledge,

significantly less work has been related to ZrB2–SiC sharp leading
Zirconium diboride (ZrB2) and zirconium diboride based com- edges [7,13]. Systematic research on UHTC sharp leading edges,
posites, commonly referred as ultra high temperature ceramic especially the ablation behavior needs to be done and the optimal
(UHTC) composites, have drawn great interests due to their R value for UHTC sharp leading edge needs to be found out.
extremely high melting temperatures, retained strength at elevated In this present study, ZrB2–SiC sharp leading edges with differ-
temperatures, high thermal conductivity, excellent oxidation and ent radius (R) values were tested using an oxy-acetylene torch
ablation resistance [1–3]. Among these materials, ZrB2–SiC compos- flame under the same condition for comparison. The purpose of
ites are currently considered the baseline UHTCs and the most this paper is to investigate the ablation behavior of the ZrB2–SiC
promising candidates for use in extreme high temperature sharp leading edges, and to give the optimum R value. We believe
structural applications, such as sharp leading edges on the future this work can lay the foundation for the engineering applications of
generation of hypersonic aerospace vehicles and reusable atmo- UHTC components, especially for sharp leading edges.
spheric re-entry vehicles [4–7].
In contrast to traditional blunt capsules or shuttle-type vehicles,
vehicles with sharp leading edge have lower drag and higher lift to
2. Experimental procedure
drag ratios (L/D) than blunt-edged vehicles, but they also have to
endure higher surface temperature [8,9]. As for the application of The testing samples here for ablation were fabricated from commercial ZrB2
ZrB2–SiC sharp leading edges, the ablation resistance is one of (Northwest Institute for Non-ferrous Metal Research, China) and SiC (Weifang Kai-
hua Micro-powder Co. Ltd., China) powders. The ZrB2 and SiC powders had the same
the most important issues. Obviously, sharp leading edges with
purity of 99% and the mean particle size of them was 2 lm and 0.5 lm, respectively.
low radius (R) values can maintain high L/D [8]. Unfortunately, The powder mixture of ZrB2 plus 20 vol.% SiC was ball-milled in ethanol for 8 h with
however, lower radius leading edges are always more subject to zirconia ball media in a planetary mill (MCA-10B, Nanjing University Instrument
much more aerothermal heating than blunt edges, such as those Plant, China) at 240 rpm (revolutions per minute). After the ball-milling, ethanol
on the space shuttle orbiter, and these edges will reach tempera- was removed by a rotating evaporator (R-202, Shanghai Shensheng Biotech Co.
Ltd., China) at 80 °C to minimize segregation. The as-received ZrB2–SiC powder mix-
tures exceed 2000 °C during re-entry conditions [10–12]. There-
tures were sieved through a 200 mesh and then uniaxially hot-pressed in a BN
fore, there is an optimum R value existing for UHTC sharp coated graphite die at 1950 °C for 60 min under vacuum and 30 MPa applied pres-
leading edges, which cannot only maintain higher L/D, but also sure. The bulk density was measured by Archimedes method, and compared to the
theoretical value calculated through rule-of-mixture of the starting composition.
The hot-pressed ZrB2–SiC composite had a relative density higher than 99%. Micro-
⇑ Corresponding author. Tel./fax: +86 451 86403016. structure was determined using a SEM (FEI Quanta 200F, USA) with an X-ray EDS
E-mail address: herujie2003jci@163.com (R. He). (EDAX Inc., USA) analyzer attachment. Fig. 1 shows the microstructure of the

0925-8388/$ - see front matter Ó 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.jallcom.2013.03.067
126 X. Jin et al. / Journal of Alloys and Compounds 566 (2013) 125–130

Fig. 3. Oxy-acetylene testing system configuration: (1) sample; (2) graphite holder;
(3) two-color Raytek pyrometer; (4) computer; (5) oxygen (O2); (6) acetylene
(C2H2); (7) copper nozzle; (8) flowmeter.

Table 1
Experimental conditions for oxy-acetylene test.
Fig. 1. SEM image of the polished surface of ZrB2–SiC.
O2 gas pressure (kPa) 0.45
polished surface of the as-prepared ZrB2–SiC composite. The SiC particles, which O2 gas flux (L/s) 0.6
represented the majority of the dark features, were homogeneously dispersed in C2H2 gas pressure (kPa) 0.1
the ZrB2 matrix and no obvious agglomeration was detected. C2H2 gas flux (L/s) 0.4
ZrB2–SiC sharp leading edge models with different radius (R) values were cut Diameter of nozzle (mm) 2
from the hot-pressed billet based on the tips paralleling to the hot-pressing direc- Distance from sample surface to nozzle (mm) 20
tion, as shown in Fig. 2. Ablation testing was conducted using an oxy-acetylene
torch. The R value was 0.15, 0.5, 1.0, and 1.5 mm, respectively, and the correspond-
ing models were designated as R1, R2, R3, and R4 for simplicity. The model was
locked upon a graphite holder at a distance of 20 mm from the nozzle exit, the sharp leading edge models underwent different ablation behavior,
surface temperature was determined using a two-color Raytek pyrometer especially represented by the surface temperature state. Fig. 4
(RAYMR1SCSF, USA), as shown in Fig. 3. Ablation time was chosen as 300 s. Table
plots the surface temperature curves for the four ZrB2–SiC sharp
1 lists the specific experimental conditions for sharp leading edge models. The
microstructural observations of the model after ablation were carried out by SEM. leading edge models with the ablation time increased up to
Finite element analysis (FEA) was used to simulate the temperature distribution 300 s. Sharp increments in the surface temperature could be ob-
in the sharp leading edge during ablation. The room temperature (25 °C) thermal served for all models at the beginning of ablation. Hereafter, the
expansion coefficient of the ZrB2–SiC composite was 4.28  106/K, which was ob- temperature reached a maximum point (Tms, the maximum surface
tained using a thermal dilatometer (DIL-402C, NETZSCH, Germany). The thermal
diffusivity, thermal conductivity, specific heat were 0.27 cm2/s, 840 J/kg K, respec-
temperature) as 2100 °C for R1, 2040 °C for R2, 2000 °C for R3
tively, which were measured using a thermal constant analyzer (Flashline 5000, and 1930 °C for R4, respectively. Such differences in the maxi-
ANTER, USA). The Young0 s modulus and Poisson0 s ratio were 420 GPa and 0.25, mum surface temperature were attributed to the different radius
respectively, which were measured by a ultrasonic velocity pulse method (Model (R) values. According to the Kemp–Riddell formula, the heat flow
5077PR, Olympus, Japan) using cubic specimen 4 mm  4 mm  4 mm test bars.
has an inverse proportional relationship with the square root of
All the testing samples were cut paralleling to the hot-pressing direction.
the radius (R) value, as shown below [14],

3. Results and discussion 1

q / pffiffiffi ð1Þ
R
3.1. Oxy-acetylene testing
where q is the heat flow and R is the radius value. Sharp leading
Table 2 gives the characteristics of the four sharp leading edge edges with lower radius are subject to higher heat flow, and thus
models after oxy-acetylene test. During the testing, the ZrB2–SiC suffer much severer aerothermal heating and ablation. Therefore,

Fig. 2. ZrB2–SiC sharp leading edge models with different curvature radius (R): (a) sharp leading edge billet cut from the hot-pressing billet; (b) R1, 0.15 mm; (c) R2, 0.5 mm;
(d) R3, 1.0 mm; (e) R4, 1.5 mm.
 X. Jin et al. / Journal of Alloys and Compounds 566 (2013) 125–130 127

Table 2
The maximum and stable surface temperatures of ZrB2–SiC sharp leading edge
models during ablation.

Model R value (mm) Ablation time (s) Surface temperature (°C)

Tms Tss
R1 0.15 300 2100 1935
R2 0.5 300 2040 1920
R3 1.0 300 2000 1910
R4 1.5 300 1930 1910

Tms, the maximum surface temperature; Tss, the steady state surface temperature.
Fig. 5. Macrographs of the ZrB2–SiC sharp leading edge models after the oxy-
acetylene testing.

their smaller radius value. On one hand, the radius values for both
R1 and R2 became much higher after severe ablation, indicating the
aerothermal heating and ablation became slighter correspond-
ingly. On the other hand, the evaporation of the oxides former
during ablation would adsorbed and removed away some of heat.
Therefore, the surface temperature of R1 and R2 reached a maxi-
mum at first and then started to decrease once the radius value
increased by ablation. For R3 model, however, only a very slight
mass loss and linear shrinkage was observed, and the surface tem-
perature curve nearly kept constantly. In addition, R4 model
underwent the slightest ablation because of the slightest aerother-
mal heating and ablation due to its high radius value. For R4 model,
a slight mass gain and linear expansion was thus observed. All the
four ZrB2–SiC sharp leading edge models decreased to the steady
state surface temperature (Tss) of about 1910–1935 °C during
ablation.
Fig. 4. The surface temperature vs. ablation time for the ZrB2–SiC sharp leading
edge models. 3.2. Finite element analysis (FEA)

All the models reached the maximum surface temperature after

the surface temperature increasing rate and the maximum surface
about 50 s ablation, as indicated in Table 2 and Fig. 4. In this study,
temperature increased with the decreasing of the radius value,
finite element analysis (FEA) was used to simulate the temperature
and Tms(R1) > Tms(R2) > Tms(R3) > Tms(R4). In addition, the surface
distribution in the ZrB2–SiC sharp leading edge during ablation. The
temperatures for all the four sharp leading edges increased drasti-
simulated temperature distributions at the ablation time of 50 s are
cally in a very short time when the models exposed to the oxyacet-
shown in Fig. 6. During ablation, the marco-scale temperature fields
ylene torch flame. Specifically, the surface temperature for R1 rose
for all four models were not uniform at the edge and bottom of the
dramatically to about 1700 °C in less than 5 s, and then rose until
models. Namely, the temperatures of the edge were much higher
reaching the maximum surface temperature of about 2100 °C in less
than that of the bottom. It could be seen from Fig. 6a, the simulated
than 50 s. The surface temperature increasing rates for R2, R3 and
maximum surface temperature of the edge for R1 model was about
R4 models descended gradually with their R values increase gradu-
2107 °C, which was nearly the same as the result from Fig. 4
ally. Besides, R2, R3 and R4 models all reached the maximum sur-
(2100 °C). In addition, the FEA indicated the maximum surface
face temperature after ablation of about 50 s.
temperature of the edge for R2, R3 and R4 model was about 2005,
Table 3 lists the linear and mass changes of the ZrB2–SiC sharp
1988 and 1909 °C, respectively, which was in good accordance with
leading edge models after ablation for 300 s under the same condi-
the results of Fig. 4. The FEA also gave the temperature distributions
tion. The macrographs of the ZrB2–SiC sharp leading edge models
of these four models during testing. The FEA indicated that there
after the oxy-acetylene testing were given in Fig. 5. It was found
was a great temperature gradient of about 150 °C/cm from the edge
that R1 and R2 underwent intensive mass loss and linear shrinkage,
to the bottom existed for each model. Such great temperature gra-
whereas R3 has only a slight mass loss and linear shrinkage, and R4
dient would induce great stress gradient in the models, especially
even had a slight mass gain and linear expansion. For R1 and R2,
at the edge of the model. Although high thermal stresses occurred
owing to the intensive mass loss and linear shrinkage, the ablation
for all models during ablation as a result of the large temperature
on tips of the sharp leading edges was obviously. Both R1 and R2
gradient, however, no obvious cracks were observed in these mod-
underwent a severer aerothermal heating and ablation owing to
els (as shown in Fig. 5).

3.3. Optimal R value

Table 3
The linear and mass changes of the ZrB2–SiC sharp leading edge models after ablation.
During the test, oxidation of the ZrB2–SiC material could induce
Model Linear change (mm) Mass change (g) to the mass gain and linear expansion; however, ablation and ero-
R1 3.4 33 sion of the material could cause the mass loss and linear shrinkage
R2 1.2 17 [15–18]. As can be clearly seen, models R1 and R2 underwent the
R3 0.3 2.5
severest ablation, the tips of the samples were badly ablated and
R4 +0.4 +1.3
the mass loss and linear shrinkage caused by ablation and erosion
+, Linear expansion and mass gain; , linear shrinkage and mass loss. were higher than the weight gain and shrinkage expansion caused
128 X. Jin et al. / Journal of Alloys and Compounds 566 (2013) 125–130

Fig. 6. The temperature distributions in the sharp leading edge models simulated by FEA at the ablation time of 50 s: (a) R1; (b) R2; (c) R3; (d) R4.

by oxidation. Besides, owing to the badly ablation, the radius At high temperatures, B2O3 vaporized quickly due to its low
values or R1 and R2 became higher and the shapes of R1 and R2 melting point (450 °C) and high vapor pressure. The rapid evapora-
were changed, which were harmful for the hypersonic and re-entry tion of B2O3 resulted in increased oxidation of ZrB2 since ZrO2 was
flight. It was also observed from Fig. 5 and Table 2 that both model not a perfect protective oxide. By introducing SiC, the oxidation
R3 and model R4 had much better ablation resistance and shape resistance was remarkably improved due to the formation of silica
retention ability, due to their linear changes and mass changes glass during ablation (Reaction 3). The ZrB2–SiC composite had a
were smaller than those of others. Both R3 and R4 underwent significant improvement in oxidation resistance below 2000 °C be-
slight mass and linear changes, which would not change the shapes cause of the formation of a silica glass layer with low oxygen
of sharp leading edges obviously. However, the radius value of R4 permeability, which provided an efficacious protective oxidation
was higher than R3, resulting in the lift to drag ratio (L/D) of R3 was barrier. However, above 2200 °C, Reactions (4)–(6) happened, the
higher than R4 [8]. Considering the surface temperature, shape oxidation resistance ability of the ZrB2–SiC composite significantly
retention behavior and the lift to drag ratio, R3 (1.0 mm) was decreased because of the active oxidation (Reactions (4) and (5))
therefore selected as the optimal radius value for ZrB2–SiC sharp and rapid evaporation (Reaction 6) due to the high vapor pressure.
leading edge. The detailed ablation mechanism of ZrB2–SiC composite had been
described in our previous published papers [8,15–17].
3.4. Ablation mechanism and microstructure of R3 From the oxy-acetylene testing results, the maximum surface
temperature and stable surface temperature of the R3 model was
It was well known that oxides produced from the oxidation of about 2000 and 1910 °C, respectively, which was lower than
ZrB2 and SiC were able to improve the ablation resistance of the 2200 °C. Therefore, there was a silica glass existing after ablation.
ZrB2–SiC composite. During testing, there were severe reactions The cross-section of model R3 after oxy-acetylene testing is given
happened to the ZrB2–SiC composite. The main expected reactions in Fig. 7. Compositional analysis of EDS showed there were three
during the oxidation process were as follows: distinct oxidation layers after ablation. The outmost layer was
ZrB2 ðsÞ þ 5=2O2 ! ZrO2 ðsÞ þ B2 O3 ðlÞ ðReaction1Þ mainly composed of a lot of SiO2, a little ZrO2 and many voids
which were formed due to the formation and vaporization of
B2 O3 ðlÞ ! B2 O3 ðgÞ ðReaction2Þ B2O3 with a high partial pressure. Underlying this layer, there
was a SiC depletion layer which was mainly composed of a little
SiCðsÞ þ 3=2O2 ðgÞ ! SiO2 ðlÞ þ COðgÞ ðReaction3Þ SiO2, recrystallized ZrO2 and unaltered SiC. The bonding strength
at the interface between the outmost layer and the SiC depletion
SiCðsÞ þ O2 ðgÞ ! SiOðgÞ þ COðgÞ ðReaction4Þ layer was very weak due to the existence of numerous large voids
and the mismatch of the thermal expansion coefficient between
SiCðsÞ þ 2SiO2 ðlÞ ! 3SiOðgÞ þ COðgÞ ðReaction5Þ the oxide products. Underneath, there was the unaltered ZrB2–
SiC matrix layer. Furthermore, Fig. 7 also shows an obvious
SiO2 ðlÞ ! SiO2 ðgÞ ðReaction6Þ difference of the thickness of these three oxidation layers in
 X. Jin et al. / Journal of Alloys and Compounds 566 (2013) 125–130 129

Fig. 7. The microstructure for different positions and EDS of oxide layers in R3: (a) low magnification of R3; (b, c, d, and e) high magnification for different positions of R3; (f)
EDS for oxidation layers.

different region of R3. On the tip of R3, the thickness of the layers radius (R) values of 0.15, 0.5, 1.0 and 1.5 mm were ablated by an
was the highest and there was an apparent crack formed between oxyacetylene torch under the same condition. During testing, the
the outmost layer and the second layer owning to the large voids surface temperatures for all the four sharp leading edge models in-
and the mismatch of the thermal expansion coefficient. From the creased drastically in a very short time. Specifically, the surface
tip to bottom, the thickness of the layers became smaller. It was temperature for R1 rose dramatically to about 1700 °C in less than
indicated that there was a great temperature gradient existing at 5 s, and then rose until reaching a maximum surface temperature
the ZrB2–SiC sharp leading edge during ablation, which was in of about 2100 °C in less than 30 s. The surface temperature reached
accordance with the results of FEA research. a maximum of 2100 °C for R1, 2040 °C for R2, 2000 °C for R3
and 1930 °C for R4, respectively, and subsequently decreased to
a state value of about 1910–1935 °C for all models. The FEA gave
4. Conclusions the simulated temperature distributions in the ZrB2–SiC sharp
leading edges at the ablation time of 30 s, which were in good
This study provided a successful proof of the ablation behavior accordance with the experimental results. R3 (1.0 mm) was the
of ZrB2–SiC sharp leading edges by means of the oxyacetylene optimal radius value for ZrB2–SiC sharp leading edges. The micro-
torch flame. ZrB2–SiC sharp leading edge models with different structures of the cross-section and oxidation surface of model R3
130 X. Jin et al. / Journal of Alloys and Compounds 566 (2013) 125–130

were also investigated. Three oxidation layers were detected. The [3] D. Sciti, F. Monteverde, S. Guicciardi, G. Pezzotti, A. Bellosi, Mater. Sci. Eng.
A434 (2006) 303–309.
first layer was mainly composed of ZrO2 and a little SiO2, the sec-
[4] F. Monteverde, R. Savino, J. Eur. Ceram. Soc. 27 (2007) 4797–4805.
ond layer was SiC-depleted layer and the third layer was unaltered [5] S.B. Zhou, W.J. Li, P. Hu, C.Q. Hong, L. Weng, Corros. Sci. 51 (2009) 2071–2079.
ZrB2–SiC matrix layer. A great temperature gradient was observed [6] F. Monteverde, R. Savino, M.D.S. Fumo, A.D. Maso, J. Eur. Ceram. Soc. 30 (2010)
through the microstructure investigation. 2313–2321.
[7] F. Monteverde, R. Savino, J. Am. Ceram. Soc. 95 (2012) 2282–2289.
This work demonstrated R3 (1.0 mm) was the optimal radius [8] X.H. Zhang, P. Hu, J.C. Han, S.H. Meng, Comp. Sci. Technol. 68 (2008) 1718–
value for ZrB2–SiC sharp leading edge. And we believe this work 1726.
can lay the foundation for the engineering applications of ultra [9] R. Savino, M.D.S. Fumo, D.M. Paterna, A.D. Maso, F. Monteverde, Aerosp. Sci.
Technol. 14 (2010) 178–187.
high temperature ceramic sharp leading edges. [10] R. Monti, D.M. Paterna, Aerosp. Sci. Technol. 10 (2006) 156–167.
[11] S.P. Walker, B.J. Sullivan, Sharp refractory composite leading edges on
Acknowledgements hypersonic vehicles, in: Twelfth AIAA International Space Planes and
Hypersonic Systems and Technologies, Virginia, 2003.
[12] I.D. Boyd, J.F. Padilla, Simulation of sharp leading edge aerothermodynamics,
The authors would like to express their gratitude to financial in: Twelfth AIAA International Space Planes and Hypersonic Systems and
supports from the National Natural Science Foundation of China Technologies, Virginia, 2003.
[13] R.J. He, X.H. Zhang, P. Hu, Key Eng. Mater. 512–515 (2012) 710–714.
(Nos. 510702042 and 50972029). [14] R.W. Detra, N.H. Kemp, F.R. Riddell, Jet Propulsion 27 (1957) 1256–1257.
[15] J.C. Han, P. Hu, X.H. Zhang, S.H. Meng, W.B. Han, Comp. Sci. Tech. 68 (2008)
References 799–806.
[16] J.C. Han, P. Hu, X.H. Zhang, Scripta Mater. 57 (2007) 825–828.
[17] P. Hu, G.L. Wang, Z. Wang, Corros. Sci. 51 (2009) 2724–2732.
[1] F. Monteverde, A. Bellosi, S. Guicciardi, J. Eur. Ceram. Soc. 22 (2002) 279–288.
[18] P. Hu, Z. Wang, J. Eur. Ceram. Soc. 30 (2010) 1021–1026.
[2] F. Monteverde, Corros. Sci. 47 (2005) 2020–2033.