Author’s Accepted Manuscript

Theoretical prediction of temperature-dependent

fracture strength for ultra-high temperature ceramic
composites considering the evolution of damage
and thermal residual stress

Xin Zhang, Weiguo Li, Yong Deng, Jiaxing Shao,

Xuyao Zhang, Xianhe Zhang, Haibo Kou, Yong
Tao, Zhaoliang Qu www.elsevier.com/locate/ceri

PII: S0272-8842(18)32504-5
DOI: https://doi.org/10.1016/j.ceramint.2018.09.043
Reference: CERI19442
To appear in: Ceramics International
Received date: 30 August 2018
Revised date: 4 September 2018
Accepted date: 5 September 2018
Cite this article as: Xin Zhang, Weiguo Li, Yong Deng, Jiaxing Shao, Xuyao
Zhang, Xianhe Zhang, Haibo Kou, Yong Tao and Zhaoliang Qu, Theoretical
prediction of temperature-dependent fracture strength for ultra-high temperature
ceramic composites considering the evolution of damage and thermal residual
stress, Ceramics International, https://doi.org/10.1016/j.ceramint.2018.09.043
This is a PDF file of an unedited manuscript that has been accepted for
publication. As a service to our customers we are providing this early version of
the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting galley proof before it is published in its final citable form.
Please note that during the production process errors may be discovered which
could affect the content, and all legal disclaimers that apply to the journal pertain.
 Theoretical prediction of temperature-dependent fracture strength for ultra-high

temperature ceramic composites considering the evolution of damage and thermal

residual stress

Xin Zhanga,b, Weiguo Lia,b, Yong Denga, Jiaxing Shaoa, Xuyao Zhanga, Xianhe Zhanga, Haibo

Koua, Yong Taoa, Zhaoliang Quc

a
College of Aerospace Engineering, Chongqing University, Chongqing, 400044, China

b
State Key Laboratory of Coal Mine Disaster Dynamics and Control, Chongqing, 400044, China

c
Institute of Advanced Structure Technology, Beijing Institute of Technology, Beijing, 100081, China

*Corresponding author: E-mail address: wgli@cqu.edu.cn

Abstract

Based on the maximum storage energy density criterion of material fracture, a model of

temperature-dependent fracture strength for ultra-high temperature ceramic composites is

established. The combined impacts of the evolution of damage and thermal residual stress with

temperature are considered. The model predictions are highly consistent with available

experimental values. Besides, the critical crack sizes of ZrB2-30vol%SiC in air from 1400 to

1600oC are predicted using the proposed model, which agree well with the total oxidation

thickness of the reported literature at 1400 and 1500oC, and a more reasonable definition of

critical crack size at 1600 oC are given. Moreover, the quantitative effect of crack size on the

fracture strength is analyzed under different environment temperature, and a useful conclusion is

obtained that decreasing crack size is more effective to improve the fracture strength of the

composites at low temperatures. This study not only provides a feasible and convenient method to

predict the fracture strengths at different temperatures, but also offers a theoretical support for the

design of ultra-high temperature ceramic composites.

1
Keywords: Fracture strength; Ultra-high temperature ceramic composites; Temperature-dependent;

Damage; Thermal residual stress

1. Introduction

Ultra-high temperature ceramic (UHTC) composites have been widely applied in thermal

protection systems and propulsion systems in aerospace applications, owing to their attractive

intrinsic properties, such as high strength, good chemical and physical stability, superb thermal

shock resistance and oxidation resistance [1-4]. The ceramics usually are subjected to high

temperature environment and sharp temperature variation in high-temperature applications [5, 6].

Fracture strength, as a vital mechanical property of the ceramics, determines the reliability and

safety of high-temperature structural components in service [7-9]. Besides, fracture mechanisms of

ceramics between room temperature and high temperature have remarkable differences, which has

great effects on the fracture strength at different temperatures [10, 11]. Acquiring the strengths and

studying the control mechanisms of the strength under different environment temperature have

always been a great concern in the ceramics field [12-18].

In the past, researchers mainly focused on using experimental methods to obtain fracture

strengths of UHTC composites at different temperatures [6-9, 14, 16-18]. Justin and Jankowiak [6]

reported that the strength of ZrB2-20vol%SiC composites is 451 MPa at normal temperature, and

it reduces to 331 MPa and 286 MPa at 1000oC and 1150oC, respectively. Hu and Wang[14]

reported the strengths of ZrB2-15vol%SiC composites, and the reported values are 500 MPa at

normal temperature and 217 MPa at 1800oC, respectively. Additionally, Neuman et al. tested that

the strengths of ZrB2 based composites up to 1600oC in air [7]. The fracture strength was

determined by the SiC cluster size from normal temperature to 1000oC, and oxidation damage

2
above 1200oC. And Neuman et al. also reported that the fracture strength of ZrB2-30vol%ZrB2 in

argon was controlled by the SiC cluster size up to 1800oC, and the formation of liquid phases,

precipitation and inclusions above 1800oC [17]. These results indicated that the crack size which

controls the fracture strength of the materials would evolve with the increase of temperature. Thus,

it is essential to take into account the effect of damage evolution on fracture strength of materials

when characterizing the temperature dependent fracture strength. Besides, residual thermal stress,

caused by the thermal expansion mismatch between the matrix and the adding phase, is a function

of temperature and would be completely relaxed at a certain high temperature [19-21]. Thus, the

fracture strength of UHTC composites is strongly dependent on temperature. Experiments can be

regarded as a powerful and useful tool to research and reveal the control mechanisms of fracture

strength with increasing temperature. However, the ultra-high temperature fracture strength testing

is difficult to be carried out. On the other hand, it is extremely expensive and time-consuming to

research and distinguish the quantitative influence of each affecting factor with temperature on the

fracture strength of UHTC composites only through experimental ways as they will consume lots

of specimens and time.

To better characterize the predominant affecting factors determining fracture of UHTC

composites and predict their fracture strengths at different temperatures, a theoretical model, thus,

needed to be proposed. In previous literature, many fracture strength models at normal

temperature have been established [3, 22-24], but those which can be applied to high temperatures

are still lacking . Recently, Deng et al. [11] developed a temperature dependence of fracture

strength model of ceramics without considering the influence of heat of fusion on fracture strength.

Ceramics still resist tensile deformation at the melting point only when it absorbs the heat of

3
fusion totally and becomes liquid [25]. Thus, it is essential to take the impact of heat of fusion into

consideration in the modeling process. And thermal residual stress is an important factor of

determining the fracture strength, which should be also taken into account when predicting the

fracture strengths of UTHCCs [26, 27]. Besides, to our knowledge, the evolution of damage with

temperature has a great influence on the fracture strength [17]. Therefore, it is necessary to

establish a temperature-damage dependence of fracture strength model. Wang et al. [28] presented

a fracture strength model of UHTC composites under different environment temperature based on

the fracture toughness of matrix materials. In addition, Shen et al. [29] proposed a temperature

dependence of fracture strength model for ZrB2-SiC ceramic composites, which considered the

impact of oxidation damage on fracture strength. However, those models are only applied to

particulate-reinforced UHTC composites as they assume that annular cracks emanate from the

particles. Moreover, those models need to acquire the specific heat capacity of ceramics, which

makes them difficult and inconvenient to predict fracture strength, especially for ceramic matrix

composites.

In this study, so as to take the influence of temperature on the fracture strength of UHTC

materials into consideration, a temperature dependence of fracture strength (TDFS) model taking

into account the impact of the heat of fusion firstly is derived, which is based on the maximum

storage energy density criterion. It relates the fracture strength under diverse temperature

environment of UHTC materials to that at arbitrary reference temperature. The TDFS model is

free of adjustable parameters. Then, on the basis of the proposed model, a new

temperature-damage dependence of fracture strength (TDDFS) model for UHTC composites is

developed, which further considers the combined influences of the evolution of damage and

4
thermal residual stress with temperature. The predicted values achieve good consistency with

available experimental data. Using the proposed model, the critical crack sizes of UHTC

composites at different temperatures are also predicted and the quantitative influence of crack size

on fracture strength is analyzed.

2. Temperature dependence of fracture strength model

2.1 Modeling

According to an equivalent relation between strain energy and heat energy because of their

contribution to the onset of fracture of materials, a method is proposed to study the temperature

dependence of the fracture strength of ceramics. For a particular material with fracture, there

exists a maximum storage energy density, which includes strain energy density and heat energy

density. The start of material failure occurs at a certain temperature when the maximum storage

energy density reaches a critical value. Based on the thermodynamics theory, the heat energy

density of a system includes the potential energy density between atoms and the kinetic energy

density of atomic motion. Ceramics will start to melt when temperature reaches the melting point.

Then, after continually being heated, it absorbs the heat energy (the heat of fusion per unit mass

H m ) and totally converts into liquid state [25]. Based on the discussions above, the maximum

storage energy density is formulated as:

Wtotal  A( E p  Ek  H m )  W th (T ) (1)

where Wtotal denotes a constant value related to the variety and structure of materials; W th (T )

denotes the strain energy density at different temperatures with material fracture; The critical

strain energy density per unit mass for UHTC materials, assumed as linear elastic solid, takes the

following equation [10]:

5
  th (T )
2

W th (T )  (2)
2 E (T )  (T )

where E (T ) denotes the temperature dependence of Young’s modulus;  th (T ) denotes the

fracture strength with temperature, respectively;  (T ) denotes the density at temperature T ;

A , assumed constant, denotes the conversion coefficient between the critical strain energy and

the heat energy. When the temperature is exactly the melting point, H m is expressed as:

Hm  H M / M (3)

where H M and M denote the heat of fusion per mole and molar mass, respectively. When T

is below the melting point， H M  0 ；the case when T exceeds the melting point is not

considered in this study. E（

p
T） denotes the potential energy density of atoms per unit mass at

different temperatures, which can be expressed as:

3 3
E（ = kN aT  kN 0T / M
p T） (4)
2 2

where k and N 0 denote Boltzmann constant ( 1.3811023 J K-1 ) and Avogadro’s constant

( 6.023 1023 mol-1 ), respectively; N a denotes the number of atoms per unit volume; E（
k
T）

denotes the kinetic energy density per unit mass at different temperatures. The kinetic energy and

the potential energy transfer periodically in the material because of the vibrating atoms, and their

average values are equal [12]. Thus, the temperature dependence of potential energy density of

atomic motion per unit mass E（

k
T） also has the form:
3
E（
k T）
= kN0T / M (5)
2

Combining Eqs. (1)-(5), one can get that:

 th2 (T ) A
Wtotal   (3kN 0T  H M ) (6)
2 E (T )  (T ) M

Substituting T  T0 into Eq.(6), where T0 is an arbitrary reference temperature,

6
H M  0 .Thus, one can get that:

 th2 (T0 ) A
Wtotal   (3kN 0T0 ) (7)
2 E (T0 )  (T0 ) M

It is worth noting that the material cannot sustain deformation at the melting point Tm when

it absorbs heat energy of H m and melts totally, thus  th (Tm )  0 .

So substituting T =Tm into Eq. (6), and thus

A
Wtotal = （3kN 0Tm +H M） (8)
M

According to Eq. (7) and Eq. (8), one can conclude that:

 2 (T0 ) M
A= th
 (9)
2 E (T0 )  (T0 ) 3kN 0  Tm -T0  +H M

Combining Eqs. (6) with (9), Wtotal can be obtained as:

 2 (T )  2 (T0 ) 3kN 0T
Wtotal = th
 th
 (10)
2 E (T )  (T ) 2 E (T0 )  (T0 ) 3kN 0 Tm -T0  +H M

And combining Eqs. (8) with (9), Wtotal can be obtained as:

 2 (T0 ) 3kN 0Tm +H M

Wtotal = th
 (11)
2 E (T0 )  (T0 ) 3kN 0 Tm -T0  +H M

Finally, combining Eq. (10) with Eq. (11), a model of fracture strength at different

temperatures can be presented as:

 E (T )  (T ) Tm -T  H M / 3kN 0 
1/ 2

 th (T )   th (T0 )  
 (12)
 E (T0 )  (T ) Tm -T0  H M / 3kN 0 
0

Moreover, according to the relationship between the density and the volume for the same

quality of the same material, one can get that:

 E (T )V (T0 ) Tm -T  H 
1/ 2
/ 3kN 0
 th (T )   th (T0 )  
M

 (13)
 E (T0 )V (T ) Tm -T  H
0 M
/ 3kN 0 

The temperature dependent volume V (T ) of materials can be taking the follow expression

7
 [30]:

T 

V (T )  V (T0 )exp   V (T )dT  (14)
T0 

where  V denotes the coefficient of the volume thermal expansion. Finally, substituting Eq.

(14) into Eq. (13), a temperature dependence of fracture strength (TDFS) model for ceramics can

be obtained:

 
1/ 2

 
 E (T ) Tm -T  H M / 3kN 0 
 th (T )   th (T0 )  (15)
 T  Tm -T0  H M / 3kN 0 
 E (T0 )exp    V (T )dT  
 0T  

The TDFS model establishes a quantitative relationship between fracture strength under

different environment temperature and that at the reference temperature. It is worth noting that

there are no fitting parameters in this model. All material property parameters can be obtained

easily from material handbooks or literatures. The evolution of damage and thermal residual stress

with temperature is ignored in the present model. The TDFS model provides a convenient method

of predicting fracture strength of ceramic materials conveniently at extremely high temperature.

Table 1. Temperature dependence of Young’s modulus (in GPa) of ZrB2-20vol%SiC [6],

2.5DSi3N4f-BN [31], Cf/TiC [32] and ZrB2-30vol%SiC [17].
Temperature/oC ZrB2-20vol%SiC 2.5DSi3N4f-BN Cf/TiC ZrB2-30vol%SiC
Room temperature 194 28.9 398 513
700 - - 364 -
900 - - 338 -
1000 137 25.7 - 501
1100 - - 296 -
1150 101 - - -
1200 - 21.9 - 431
1300 - 13.5 253 -
1400 - - - 427
1600 - - - 399
1800 - - - 360
2000 - - - 283
2200 - - - 164

2.2 Model verification

8
 In this section, ZrB2-20vol%SiC[6], ZrB2-30vol%SiC[17], 2.5DSi3N4f-BN[31] and

Cf/TiC[32] are used to verify the TDFS model, and the predictions are compared with the

experimental results. Because the fracture strength at normal temperature is tested more easily

than that at higher temperature, it is convenient to taken room temperature as the reference

temperature. In calculations, as most of the magnitude of V is lower than 105 and

maximum temperature is about 3000oC for UHTC materials, thus the calculation value of

T T
T 
0
V
(T )dT is quite small. Therefore, the value of exp   V (T )dT is set as 1. Temperature
T 0

dependence of Young’s modulus of the calculated ceramics are showed in Table 1.

2.2.1 ZrB2 -20vol%SiC

Justin et al. tested the temperature-dependent fracture strength of ZrB2 -20vol%SiC, and the

detailed testing process is shown in Ref.6. The average dimension of test bars is

35.40  5.20 1.75 mm3 for fracture strength testing with a speed of 0.3 mm  min 1 [6].

Three to four test bars were used for each temperature [6]. The material property parameters

needed in Eq. (15) are listed as follows:  th (T0 )  451MPa [6] ， Tm  3040℃ [33],

H M  104.6kJ/mol [34]. In Fig. 1, the theoretical predictions both by the TDFS model and Deng’s
-1

model achieve good consistency with the experimental data.

700

600
Fracture strength th(MPa)

500

400

300
Experimental values [6]
200 Theoretical predictions by TDFS model
Theoretical predictions by Deng's model [11]
100

0
0 200 400 600 800 1000 1200
Temperature T ( ℃ )
9
 Fig. 1 Experimental values and theoretical predictions of TDFS of ZrB2 -20vol%SiC

2.2.2 2.5DSi3N4f-BN

Zou et al. measured the fracture strength of 2.5DSi3N4f-BN at different temperatures, and

the detailed testing process is shown in Ref.31. The size of the test bars for the fracture strength

testing was 55  6  4 mm3 , and the loading speed was 0.5 mm  min 1 [31]. The material

property parameters needed in Eq. (15) are listed as follows:  th (T0 ) =132.6MPa [31],

Tm  3000℃ [31], H M  81kJ/mol [34]. In Fig.2, the theoretical predictions by the TDFS model
-1

and Deng’s model are highly consistent with the experimental values.

2.2.3 Cf/TiC

Song et al. tested the fracture strength of Cf/TiC with temperature, and the detailed testing

process is shown in Ref.32. The specimen dimension was 30  3  4 mm3 , and the fracture

strength was measured with 20 mm span at a crosshead speed of 0.5 mm  min 1 [32]. The

material property parameters needed in Eq. (15) are listed as follows:  th (T0 )  593MPa [32]，

Tm  3016℃ [32], H M  71kJ/mol [34]. As one can see from Fig.3 that good agreement is
-1

obtained between the theoretical predictions by the TDFS model and the experimental values. The

fracture strength at 1400oC is not predicted as the Young’s modulus at that temperature is not

reported in Ref.32. Deng’s theoretical predictions are lower than experimental results at high

temperature because the effect of heat of fusion was not involved in that model.

10
 200

160

Fracture strength  th(MPa)

120

80
Experimental data [31]
Theoretical predictions by TDFS model
40 Theoretical predictions by Deng's model [11]

0
0 200 400 600 800 1000 1200 1400
Temperature T ( )
℃

Fig. 2 Experimental values and theoretical predictions of TDFS of 2.5DSi3N4f-BN

800

700

600
Fracture strength  th(MPa)

500

400

300 Experimental values [32]

Theoretical predictions by TDFS model
200 Theoretical predictions by Deng's model [11]

100

0
0 200 400 600 800 1000 1200 1400
Temperature T ( ) ℃

Fig. 3 Experimental values and theoretical predictions of TDFS of Cf/TiC

2.2.4 ZrB2-30vol%SiC

Neuman et al. reported the fracture strength of ZrB2-30vol%SiC under different environment

temperature in argon, and the detailed testing process is shown in Ref.17. Twenty specimens

(45  4  3 mm3 ) were measured at room temperature and a minimum of five specimens

(45  4  3 mm3 ) were tested at each elevated temperature (1000, 1200, 1400, 1600, 1800,

2000 and 2200oC) [17]. The material property parameters needed in Eq. (15) are listed as follows:

 th (T0 )  595MPa [17], Tm  3040℃ [33], H M  104.6kJ/mol [34]. In Fig.4, the theoretical
-1

11
predictions achieve good consistency with the experimental values expect at 2200oC. Neuman et

al.[17] reported that the evolution of damage occurred and the strength-limiting crack increased

sharply to 47μm at 2200oC. An increase in the crack size would lead to a decrease in fracture

strength based on the Griffith theory, thus the fracture strength decreased sharply at 2200oC. The

evolution of damage with temperature is not considered in the TDFS model. Therefore, the model

prediction is larger than the experimental result at 2200oC. Deng’s theoretical predictions are all

lower than experimental values at high temperature as the effect of heat of fusion is not included.

900

800

700
Fracture strength  th(MPa)

600

500

400
Experimental values [17]
300
Theoretical predictions by TDFS model
200 Theoretical predictions by Deng's model [11]

100

0
0 400 800 1200 1600 2000
Temperature T ( ) ℃

Fig. 4 Experimental values and theoretical predictions of TDFS of ZrB2 -30vol%SiC in argon

According to the discussion above, after comparing with Deng’s model, it is necessary to take

heat of fusion into consideration when predicting the facture strength of the composites at different

temperatures. Influences of initial damage and thermal residual stress have been included in the

TDFS model by the means of the fracture strength at reference temperature. However, the

evolution of damage and thermal residual stress with temperature will lead to discrepancies

between the fracture strength predicted values and experimental data of ceramics at high

temperatures. Besides, thermal residual stress is another important factor that cannot be ignored

12
for predicting fracture strength of ceramic composite materials. Thus, to predict temperature

dependence of fracture strength of ceramics more accurately, damage, thermal residual stress and

their evolution are further considered in the next modeling process.

3. Temperature-damage dependence of fracture strength model

3.1 Modeling

Based on the same idea and calculation process of the TDFS model proposed above, the

formula of the critical strain energy W (T ) is replaced by [35]:

th

(1   (T ) ) K C (T )
2 2

W (T )  (16)
th E (T )

where  (T ) , KC (T ) and E (T ) denote Poisson’s ratio, fracture toughness and Young’s

modulus at different temperatures, respectively. Similar with the deducing process in Section 2.1,

according to Eqs. (1), (3), (14) and (16), the temperature dependent fracture toughness can be

obtained:

1/ 2
 
 
 E (T ) 1   (T0 ) 2 Tm -T  H M / 3kN 0 
K C (T )  K C (T0 )    (17)
T  1  (T ) 2 Tm -T0  H M / 3kN 0 
 E (T0 )exp   V (T )dT  
   
 T 0 

On the basis of the Griffith fracture criterion [36], fracture toughness K C and fracture

strength  th have the following relationship,

KC =Y th c (18)

where Y and c denote a geometric constant related to the fracture origin and the critical crack
size in the materials, respectively. The criterion of brittle strength was presented to describe the

extension of a crack in a homogeneous elastic and infinite body. The essential condition for brittle

fracture under static condition, is that the energy released from the crack tip is equal to the energy
13
required to form the crack area. Combining Eqs. (17) and (18), one can get that：

1/2
 
 
KC (T0 )  E (T )[1  (T0 ) 2 ] Tm -T  H M / 3kN 0 
 th
（T）=   
(19)
Y c(T )  T  Tm -T0  H M / 3kN 0 
 E (T0 )exp    V (T )dT  [1  (T ) ]
2

  T0  

Besides, thermal residual stress is a function of temperature and is an important factor that

cannot be ignored when predicting fracture strength of ceramic composite materials [37]. Thus,

the temperature-damage dependence of fracture strength (TDDFS) model of UHTC composites is

obtained:

1/2
 
 
KC (T0 )  E (T )[1  (T0 ) 2 ] Tm -T  H M / 3kN 0 
 th（T）=     BP(T ) (20)
Y c(T )  T  Tm -T0  H M / 3kN 0 
 E (T0 )exp    V (T )dT  [1  (T ) ]
2

  T0  

where B is a constant, which can be obtained using the fracture strength at reference

temperature [38], P(T ) denotes the temperature dependence of thermal residual stress. For

different second adding phase-reinforced UHTC composites, the corresponding temperature

-dependent thermal residual stress model or expression can be substituted [39, 40] into Eq. (20).

The combined effects of the evolution of damage and thermal residual stress with temperature are

included in the TDDFS model. The temperature dependence of Young’s modulus and critical

crack size, and fracture toughness at reference temperature can be obtained from experiments or

existing literatures. Besides, melting point, sintering temperature and heat of fusion can easily be

obtained from materials handbooks or literatures.

3.2 Model verification

In this section, particulate-reinforced UHTC composites are taken as examples to validate the

14
TDDFS model. On account of the thermal expansion mismatch between the matrix and particle,

there will exist residual stress within particulate-reinforced UHTC composites. The thermal

residual stress with temperature is formulated as [40]：

1
1   m (T ) 1  2 p (T ) 
P(T )   m (T )- p (T )  (Ts -T )    (21)
 2 Em (T ) Ep (T ) 

where Ts denotes sintering temperature; m and p denote particle and matrix, respectively;

 (T ) denotes the coefficients of temperature-dependent thermal expansion;  (T ) denotes the

Poisson’s ratio under different environment temperature; E (T ) denotes Young’s modulus of at

different temperatures.

The temperature-damage dependence of fracture strength of ZrB2-30vol%SiC in an argon

atmosphere is re-predicted. The fracture strengths of ZrB2-20vol%SiC, 2.5DSi3N4f-BN and Cf/TiC

are not re-calculated as the critical crack size with temperature is not reported in Ref.6, Ref.31 and

Ref.32, respectively. During calculations, the impact of temperature on Poisson’s ratio is ignored

due to its weak dependence on temperature. Relative material parameters used in Eq. (20) are

listed in Table1 and Table2. The other parameters needed in Eq. (20) are listed as follows:

KC (T0 )  4.9MPa  m1/2 [17], Tm  3040℃ [33], H M  104.6kJ/mol-1 [34], Y   [17],

Ts  1950℃[17]. The thermal expansion coefficients and Young’s modulus of SiC and ZrB2 under

different environment temperature can be obtained by the following expression [41-43]:

 p (T )  1.8276  0.0178T  1.5544  10 T  4.5246  10 T

5 9 6
(10 /K) (125  T <1273K) ;
2 3

 p (T )  5.0 (10 / K)
6
(T  1273K)

 m (T )  2.33  0.006  T  0.2  10  T (10 /K)

5 2 6

Ep (T )  410  0.04T exp( 962 / T )

Em (T )  450.0  2.54T exp( Tm / T )  1.9(T  0.363Tm  T  0.363 Tm ) exp( Tm / T )

Neuman et al. [17] pointed out that the fracture strengths of ZrB2-SiC was determined by the
15
largest SiC cluster sizes from room temperature to 1800oC in argon, and the largest SiC cluster

sizes were defined as the critical crack size in this temperature regime. Thus, the reported

definition of critical crack sizes is used to predict the fracture strengths of ZrB2-30vol%SiC in

argon up to 1800oC. The measured largest SiC cluster sizes are listed in Table2. As one can see

from Fig. 5, the theoretical predictions by TDDFS model agree well with experimental values up

to 1800 oC. However, the predictions are still highly consistent with the experimental data when

the largest SiC cluster sizes of the material are identified as the critical crack sizes above 1800 oC.

Wang et al. [28] also reported that the crack sizes of ZrB2-SiC were identified as the largest SiC

cluster sizes above 1800 oC, and their predictions showed good agreement with experimental data.

900

800

700
Fracture strength th(MPa)

600

500

400
Experimental values [17]
300 Theoretical predictions by Wang's model [28]
Theoretical predictions by TDDFS model
200

100

0
0 200 400 600 800 1000 1200 1400 1600 1800 2000 2200
Temperature T ( ) ℃

Fig. 5 Experimental values and theoretical predictions of TDDFS of ZrB2-30vol%SiC in argon

Table 2 The largest SiC cluster size in argon and total oxidation thickness of the ZrB2-30vol%SiC
in air.
Temperature(oC) Largest SiC cluster size(μm) [7] Total oxidation thickness (μm) [17]
25 32.8 -
1000 28.4 -
1200 32.6 -
1400 31.0 16.7
1500 - 23.2
1600 33.2 27.4
1800 32.5 -
2000 38.4 -
2200 46.5 -

16
 Moreover, the temperature-damage dependence of fracture strengths of ZrB2-30vol%SiC in

air up to 1600oC are also predicted. Neuman et al. reported the fracture strength of

ZrB2-30vol%SiC in air at different temperatures, and the detailed testing process is shown in Ref.

7. Nine specimens (45  4  3 mm3 ) were measured at room temperature and a minimum of

five specimens (45  4  3 mm3 ) were tested at each elevated temperature (1000, 1200, 1400,

1500 and 1600oC) [7]. The speed of loading is 0.5 mm  min 1 up to 1200 oC, 1.5 mm  min 1 at

1400 oC, 2.0 mm  min 1 at 1500 oC and 2.5 mm  min 1 at 1600 oC, respectively [7]. In this study,

the largest SiC cluster size from normal temperature to 1200 oC and the total oxidation thickness

above 1400 oC are defined as the critical crack size of ZrB2-30vol%SiC in air. The measured

largest SiC cluster size and total oxidation thickness are listed in Table2. However, at 1600oC, the

major difference was full penetration of the glassy phase into the ZrO2 layer and the layer had a

higher density at 1600oC [7], which would not decrease the strength of the composites as the layer

was extremely strong. Thus, the critical crack size ought to equal to the oxidation thickness above

the layer at 1600oC. With this consideration, a comparison between the experimental values and the

theoretical predictions of ZrB2-30vol%SiC from normal temperature to 1600oC is showed in Fig. 6.

Predicted results by the TDDFS model are well consistent with the report of experimental data. The

Wang’s model and Shen’s model predictions are also in good agreement with the experimental

values of ZrB2-30vol%SiC in air. However, Wang’s model and Shen’s model are only applied to

particulate-reinforced UHTC composites as they assume that annular cracks emanate from the

particles, so the application range of TDDFS model is wider than the previous methods. And those

models need to acquire the specific heat capacity of ceramics, which makes them difficult and

inconvenient to predict fracture strength, especially for ceramic matrix composites. Besides,

17
compared with Wang’s model, TDDFS model establishes a simpler relationship among temperature,

critical crack size, thermal residual stress and fracture strength. Shen et al. only consider the effect

of damage evolution because of crack healing on fracture strength in air. In other words, other

modes of damage evolution are not included in Shen’s model, which makes the prediction at

1500oC deviate from the experimental value. Thus, Shen’s model cannot calculate the fracture

strength of ZrB2-30vol%SiC when damage evolution exists in argon.

900

800

700
Fracture strength  th(MPa)

600

500

400
Experimental values [7]
300 Theoretical predictions by TDDFS model
Theoretical predictions by Wang's model [28]
200 Theoretical predictions by Shen's model [29]
100

0
0 200 400 600 800 1000 1200 1400 1600
Temperature T ( ℃ )

Fig. 6 Experimental values and theoretical predictions of TDDFS of ZrB2 -30vol%SiC in air

With Eq. (20), the critical crack size of ZrB2-30vol%SiC is also predicted. In Fig.7, the

predicted values by the TDDFS model of ZrB2-30vol%SiC in an argon atmosphere are compared

with the experimental values, the calculated results by Neuman et al. and those by Wang et al.. As

one can see, both the predictions of TDDFS model and Wang’s model are good consistency with

the experimental values of the maximum SiC cluster size except it at 2200oC. At 2200oC, the

prediction by TDDFS model is larger than the largest SiC cluster size because there exists the

formation of liquid phases, and precipitation of large BN and B-O-C-N inclusions, which would

lead to the decrease in fracture strength [17]. These influencing factors are not considered in the

TDDFS model. However, the calculated values by Neuman et al. using the normal temperature
18
model were much larger than the experiments above 1400oC as the combined effects the evolution

of crack size and thermal residual stress with temperature were not considered in their

calculations.

When ZrB2-30vol%SiC in an air atmosphere, measured total oxidation thickness and

ZrO2+ZrB2+SiC+Porosity* layer as well as the predicted half of critical crack sizes of the between

1400oC and 1600oC are shown in Fig. 8. The fracture strength was determined by oxidation

damage at this temperature range, and the critical crack sizes equal to total oxidation thickness [7].

However, as one can see that the predicted results are highly consistent with the total oxidation

thickness from 1400oC to 1500oC, but lower than the total oxidation thickness at 1600oC. As

mentioned above, ZrO2 layer was extremely strong at 1600oC, which would not decrease the

fracture strength of the composites. Thus, the critical crack size ought to equal to the thickness

above the ZrO2 layer at 1600oC. And the predicted critical crack size is exactly equivalent to the

reported ZrO2 + ZrB2 + SiC + Porosity* layer at 1600oC, which demonstrated the reasonability of

the critical crack size.

120
110
100 The largest SiC cluster sizes [17]
Critical flaw size c (m)

90 Predicted values by the TDDFS model

Calculated results by Wang et al. [28]
80 Calculated results by Numan et al. [17]

70
60

50
40
30

0 400 800 1200 1600 2000

Temperature T ( ) ℃

Fig. 7 The critical crack size of the ZrB2-30vol%SiC in argon up to 2200oC

It can be seen from the above results that the critical crack size varies with temperature and
19
has a great influence on the fracture strength of materials. To quantitatively analyze the impact of

crack size with temperature on fracture strength,  / c and  / c as a function of temperature

are plotted. In Fig. 9(a), the fracture strength reduces with the incensement of crack size at a given

temperature; when the crack size keeps a constant value, the fracture strength decreases with

increasing temperature. In Fig. 9(b), the variation of fracture strength causes by the same change

of crack size become smaller at higher temperature. It indicates that the fracture strength of UHTC

composites becomes less sensitive to the crack size with increasing temperature. In other words,

decreasing the crack size is more effective to improve the fracture strength of the composite at low

temperatures. Thus, reducing the maximum SiC cluster size could lead to improving the strength

of ZrB2-SiC ceramics at room temperature and retention of strength at high temperatures.

Generally, the possibility of forming SiC clusters could be decreased by improving the dispersion

of SiC powder, controlling the volume fraction of SiC lower than the percolation threshold [44]

and using a larger SiC power than the ZrB2 powder. These methods can improve the fracture

strength of ZrB2-SiC significantly under different environment temperature, especially at low

temperatures.

45 45
Total oxidation thickness [7]
40 40
Predicted half of critical crack sizes

ZrO2+ZrB2+SiC+Porosity* [7]
35 Predicted half of critical flaw sizes by the TDDFS model 35

30 30
Thickness (m)

25 25

20 20

15 15

10 10

5 5

0 0
1400 1450 1500 1550 1600

. Temperature T( )
℃

20
Fig. 8 The total oxidation thickness, ZrO2+ZrB2+SiC layers and the predicted half of critical crack

size of ZrB2-30vol%SiC in air between 1400oC to 1600oC. *The combination of ZrO2+ZrB2+SiC

and ZrO2-SiO2+B2O3 layers at 1600oC.

(a) 1000 (b) 250

o o
T= 25 C T=1000 C T= 25oC
900 o o
200
T=1600 C T=2000 C o
T=1000 C
Fracture strength th(MPa)

o T
800 T=2200 C 150 o
T=1600 C
100 T=2000oC
700 o

th(MPa)
T=2200 C
50
600
0
500
T -50
400 Initial crack size
-100 T
c0=30 m
300
-150
16 20 24 28 32 36 40 44 -14 -12 -10 -8 -6 -4 -2 0 2 4 6 8 10 12 14
Flaw size c(m) c (m)

Fig. 9 The sensitivity of TDDFS of ZrB2-30vol%SiC in argon to crack size at different

temperatures

4. Conclusions

In this study, based on the maximum storage energy density criterion of material fracture, a

temperature dependence of fracture strength (TDFS) model for ultra-high temperature ceramic

(UHTC) materials is establish firstly, which considers the impact of the heat of fusion. The model

is free of adjustable parameters. Then, on the basis of the TDFS model, a new temperature-damage

dependence of fracture strength (TDDFS) model for UHTC composites is proposed, which further

takes into account the influences of the evolution of damage and thermal residual stress with

temperature. The new model predictions are obtained good agreement with the experimental

values of ZrB2-30vol%SiC in different atmospheres. The TDDFS model, based on the

thermodynamics theory, is not needed to acquire the specific heat capacity of ceramics any more,

and all material parameters needed in the model can be obtained easily. Thus, it is convenient to

predict the temperature-damage dependence of fracture strength, especially for UTHC composites.

21
Besides, using the TDDFS model, the critical crack size of ZrB2-30vol%SiC is predicted, which

achieves good agreement with the maximum SiC cluster size up to 2200oC in argon. And the

predicted critical crack size are highly consistent with the total oxidation thickness of the reported

literature at 1400 and 1500oC, but the ZrO2+ZrB2+SiC+Porosity* layer at 1600oC in an air

atmosphere. So the definition of the critical crack size at 1600oC is modified as the

ZrO2+ZrB2+SiC+Porosity* layer.

The TDDFS model provides a new theoretical method to predict the temperature-damage

dependence of fracture strengths of UHTC composites and helps to decrease the dependence on

carrying out strength experiments at ultra-high temperatures. Besides, this study could help to

quantitatively characterize the influence of the crack size on the fracture strength under different

environment temperature, and a useful conclusion is obtained that the fracture strength of ceramics

becomes less sensitive to crack size with increasing temperature. Moreover, the methods of

reducing the largest SiC cluster size to improve the temperature-damage dependence of fracture

strength of ZrB2-SiC are proposed.

Acknowledgments

This work was supported by the National Natural Science Foundation of China [Nos.

11727802，11672050, 11472066]; the Fundamental Research Funds for the Central Universities

[grant numbers 106112017CDJXYY0001 and 106112017CDJQJ328840]; the graduate research

and innovation foundation of Chongqing, China [grant No. CYB18066].

22
Reference

[1] B. L. Wang, J. C. Han, and S. Du, Analysis of fiber breaking in fiber/matrix composites under

torsional loads, Mech. Compos. Mater. 46 (2010) 365-374.

[2] J.J. Meléndez-Martı́Nez, A. Domı́nguez-Rodrıǵ uez, F. Monteverde, C. Melandri, G.D. Portu,

Characterisation and high temperature mechanical properties of zirconium boride-based materials,

J. Eur. Ceram. Soc. 22 (2002) 2543-2549.

[3] A. Nisar, S. Ariharan, K. Balani, Establishing microstructure-mechanical property correlation

in ZrB2-based ultra-high temperature ceramic composites, Ceram. Int. 43 (2017) 13493-13492.

[4] Y. J. Cui, B. L. Wang, K. F. Wang, Fracture Mechanics analysis of buckling of a ceramic foam

coating from elastic substrates, Ceram. Int. (2018) https://doi.org/10.1016/j.ceramint.2018.06.276

[5] X. Xu, S. Sheng, et al., Evolution Mechanisms of Thermal Shock Cracks in Ceramic Sheet, J

Appl. Mech. 83 (2016)

[6] J. F. Justin, A. Jankowiak, Ultra High Temperature Ceramics: Densification, Properties and

Thermal Stability, Aerospace Lab 3 (2011) 1-11.

[7] E.W. Neuman, G.E. Hilmas, W.G. Fahrenholtz, Mechanical behavior of zirconium

diboride-silicon carbide ceramics at elevated temperature in air, J. Eur. Ceram. Soc 33 (2013)

2889-2899.

[8] T. Kusunose, T. Sekino, Improvement in fracture strength in electrically conductive AlN

ceramics with high thermal conductivity, Ceram. Int. 42 (2016) 13183-13189.

[9] Z. Qu, R. He, X. Cheng, D. Fang, Fabrication and characterization of B 4C-ZrB2-SiC ceramics

with simultaneously improved high temperature strength and oxidation resistance up to 1600 °C,

Ceram. Int. 42 (2016) 8000-8004.

[10] R.Z Wang, W.G Li, Characterization models for thermal shock resistance and fracture

strength of ultra-high temperature ceramics at high temperatures, Theor. Appl. Fract. Mec. 90

(2017) 1-13

[11] Y. Deng, W.G Li, J.X. Shao, X.H. Zhang, H.B. Kou, et al., A novel theoretical model to

predict the temperature-dependent fracture strength of ceramic materials, J. Eur. Ceram. Soc. 37

(2017) 5071-5077
23
[12] A. Rezaie, W.G. Fahrenholtz, G.E. Hilmas, Effect of hot pressing time and temperature on the

microstructure and mechanical properties of ZrB2-SiC, J. Mater. Sci. 42 (2007) 2735-2744.

[13] Li W, Yang F, Fang D, The temperature-dependent fracture strength model for ultra-high

temperature ceramics. Acta Mech. Sinica 26 (2010) 235-239.

[14] P. Hu, Z. Wang, Flexural strength and fracture behavior of ZrB2-SiC ultra-high temperature

ceramic composites at 1800 oC, J. Eur. Ceram. Soc. 30 (2010) 1021-1026.

[15] W.G Li, D.Y Li, et al. Modelling the effect of temperature and damage on the fracture

strength of ultra-high temperature ceramics. Int. J. Fracture 176 (2012) 181-188.

[16] E.W. Neuman, G.E. Hilmas, W.G. Fahrenholtz, Strength of Zirconium Diboride to 2300°C, J.

Am. Ceram. Soc. 96 (2013) 47-50.

[17] E.W. Neuman, G.E. Hilmas, W.G. Fahrenholtz, Mechanical behavior of zirconium

diboride-silicon carbide-boron carbide ceramics up to 2200°C, J. Eur. Ceram. Soc. 35 (2015)

463-476.

[18] E.W. Neuman, G.E. Hilmas, W.G. Fahrenholtz, Ultra-high temperature mechanical properties

of a zirconium diboride-zirconium carbide ceramic, J. Am. Ceram. Soc. 99 (2016) 597-603.

[19] D. Sciti, F. Guicciardi, et al., Microstructure and mechanical properties of ZrB2-MoSi2

ceramic composites produced by different sintering techniques, Mat. Sci. Eng. A 434 (2006)

303-309.

[20] L. Silvestroni, D. Sciti, Effects of MoSi2 additions on the properties of Hf-and ZrB2

composites produced by pressureless sintering, Scripta Mater. 57 (2007) 165-168.

[21] D. Kalish, E. V. Clougherty, and K. Kreder, Strength, fracture mode, and thermal stress

resistance of HfB2 and ZrB2, J. Am. Ceram. Soc. 52 (1969) 30-36.

[22] D.J. Green, An Introduction to the Mechanical Properties of Ceramics, 348 (1998).

[23] G.A. Francfort, J.J. Marigo, Revisiting brittle fracture as an energy minimization problem, J.

Mech. Phys. Solids 46 (1998) 1319-1342.

[24] V.D. Krstic, Effect of microstructure on fracture of brittle materials: Unified approach, Theor.

Appl. Fract. Mec. 45 (2006) 212-226.

[25] T.B. Cheng, W.G Li, and D.N Fang, Modeling of the temperature-dependent ideal tensile

strength of solids, Phys. Scripta 89 (2014).

24
[26] V. D. Krstic, and M.D. Vlajic, Conditions for spontaneous cracking of a brittle matrix due to

the presence of thermoelastic stresses, Acta Metall. 31 (1983) 139-144.

[27] J. Watts, G. Hilmas, W.G. Fahrenholtz, D. Brown, B. Clausen, Measurement of thermal

residual stresses in ZrB2-SiC composites. J. Eur. Ceram. Soc. 31 (2011) 1811-1820.

[28] R. Wang, W.G Li, B. Ji, D.N Fang, Fracture strength of the particulate-reinforced ultra-high

temperature ceramics based on a temperature dependent fracture toughness model, J. Mech. Phys.

Solids 107 (2017) 365-378.

[29] H. Wang, J. Chen, P. Yu, et al., Temperature-dependent fracture strength and the effect of

oxidation for ZrB2-SiC ceramics. J. Eur. Ceram. Soc. 38 (2017) 1112-111.

[30] O.L. Anderson, D.G. Isaak, Elastic constants of mantle minerals at high temperature, Mineral

Physics and Crystallography: A Handbook of Physical Constants 2 (1995) 64-97.

[31] C. Zou, B. Li, S. Wang, et al., Fabrication and high-temperature mechanical properties of

2.5DSi3N4f/BN fiber-reinforced ceramic matrix composite, Mater. Design 92 (2016) 335-344.

[32] G.M Song, Y. Zhou, S.J.L. Kang. Experimental description of thermomechanical properties

of carbon fiber-reinforced TiC matrix composites, Mater. Design 24 (2003) 639-646.

[33] G.J. Zhang, Z.Y. Deng, N. Kondo, J.F. Yang, T. Ohji, Reactive hot pressing of ZrB 2-SiC

composites, J. Am. Ceram. Soc. 83 (2000) 2330-2332.

[34] J.G.Speight, Lange’s handbook of chemistry, New York: McGrawHill, (2005) .

[35] J.H. Gong. Fracture Mechanics of Ceramics. Beijing: Tsinghua University Press, (2001).

[36] B. Lawn, Fracture of Brittle Solids, Cambridge University Press, Cambridge, U.K., (1993)

30-33.

[37] J. Watts, G. Hilmas, W.G. Fahrenholtz, et al., Stress measurements in ZrB2-SiC composites

using Raman spectroscopy and neutron diffraction. J. Eur. Ceram. Soc. 30 (2010) 2165-2171.

[38] Y. Li, W.G. Li, et al., Temperature-dependent longitudinal tensile strength model for

short-fiber-reinforced polymer composites considering fiber orientation and fiber length

distribution, J. Mater. Sci. 53 (2018) 12190-12202.

[39] Y. Deng, W.G. Li, R.Z. Wang, et al., 2017b. Temperature dependent first matrix cracking

stress model for the unidirectional fiber reinforced ceramic composites. J. Eur. Ceram. Soc. 37

(2016) 1305-1310.

25
[40] J. Selsing, Internal Stresses in Ceramics. J. Am. Ceram. Soc. 44 (1961) 419-419.

[41] L.L. Snead, et al., Handbook of SiC properties for fuel performance modeling, J. Nucl. Mater.

371 (2007) 329-377.

[42] W.G. Li, R.Z. Wang , D.Y. Li, D.N. Fang, A model of temperature-dependent Young’s

modulus for ultra-high temperature ceramics. Phys. Res. Int. 2011 (2010) 13-14.

[43] J.C. Han, B. L. Wang, Thermal shock resistance of ceramics with temperature dependent

material properties at elevated temperature. Acta Mater. 59 (2011) 1373-1382 .

[44] D. He, N.N. Ekere, Effect of particle size ratio on the conducting percolation threshold of

granular conductive-insulating composites. J. Phys. D Appl. Phys. 37 (2004) 1848-1852.

26