Thermal Decomposition of MAX Phases

by Professor Jim Low

Introduction

MAX phases exhibit a unique combination of characteristics of both ceramics and metals with unusual mechanical, electrical and thermal properties^1-3. These materials are nano-layered ceramics with the general formula M_n+1AX_n (n = 1 - 3), where M is an early transition metal, A is a group A element, and X is either carbon and/or nitrogen. The unique combination of these interesting properties enables these ceramics to be promising candidate materials for use in diverse fields which include automobile engine components, heating elements, rocket engine nozzles, aircraft brakes, racing car brake pads and low-density armour .

However, the high-temperature stability in MAX phases has hitherto generated much controversy among researchers. For instance, several researchers have reported that Ti₃SiC₂ became unstable at temperatures greater than 1400ºC in an inert atmosphere by dissociating into Si, TiC_x and/or Ti₅Si₃C_x^4-8. A similar phenomenon has also been observed for Ti₃AlC₂ whereby it decomposes in vacuum to form TiC and Ti₂AlC^9-12.

In other studies, Zhang et al.¹³ reported Ti₃SiC₂ to be thermally stable up to 1300ºC in nitrogen, but above this temperature drastic degradation and damage occurred due to surface decomposition. Feng et al.¹⁴ annealed the Ti₃SiC₂-based bulk samples at 1600ºC for 2h and 2000ºC for 0.5 h in vacuum (10^-2 Pa) and found that TiC_x was the only phase remaining on the surface. According to Gao et al.¹⁵ the propensity of decomposition of Ti₃SiC₂ to TiC_x was related to the vapour pressure of Si, i.e., the atmosphere where the Ti₃SiC₂ exits. They believed that the partial pressure of Si plays an important role in maintaining the stability of Ti₃SiC₂ whereby it has a high propensity to decompose in N₂, O₂ or CO atmospheres at temperatures above 1400ºC. This process of surface-initiated phase decomposition was even observed to commence as low as 1000 - 1200ºC in Ti₃SiC₂ thin films during vacuum annealing¹⁶. The large difference in observed decomposition temperatures between bulk and thin-film Ti₃SiC₂ has been attributed to the difference in diffusion length scales involved and measurement sensitivity employed in the respective studies. In addition, Ti₃SiC₂ has also been observed to react readily with molten Al, Cu, Ni and cryolite (Na₃AlF₆) at high temperatures.

In contrast, Barsoum and co-workers¹⁷ have shown that Ti₃SiC₂ was thermodynamically stable up to at least 1600ºC in vacuum for 24h and in argon atmosphere for 4h. They further argued that the reduced temperature at which Ti₃SiC₂ decomposed as observed by others was due to the presence of impurity phases (e.g. Fe or V) in the starting powders which interfered with the reaction synthesis of Ti₃SiC₂, and thus destabilized it following prolonged annealing in an inert environment¹⁸. However, mixed results have been reported by Radhakrishnan et al.¹⁹. In their investigation, Ti₃SiC₂ was shown to be stable in a tungsten-heated furnace for 10h at 1600ºC and 1800ºC in an argon atmosphere, but dissociated to TiC x under the same conditions when using a graphite heater.

These conflicting results suggest that the thermochemical stability of MAX phases is still poorly understood although its susceptibility to thermal decomposition is strongly influenced by factors such as:

• Purity of powders and sintered materials
• Temperature
• Vapour pressure
• Atmosphere, and
• The type of heating elements used.

In addition, the nature of the microstructure of the decomposed surface layer formed during annealing remains controversial, especially in relation to the role of pore sizes in the decomposition kinetics at the near surface.

In this article, the use of in-situ neutron diffraction to investigate the effect of vacuum annealing on the kinetics of thermal stability of several MAX phases in the temperature range 1000-1800°C is described. The role of pore size on the kinetics of phase decomposition is discussed.

Thermal Stability and Phase Transitions of MAX Phases

The phase transitions in several MAX phases and their relative phase abundances at various temperatures as revealed by in-situ neutron diffraction is shown in Figure 1. A weight loss of ~ 4% was observed for decomposed Ti₃SiC₂ which may be attributed to the release of gaseous Ti and Si by sublimation during the decomposition process. For Ti₃AlC₂, its decomposition into TiC and Ti₂AlC as lower order or intermediate phase was observed at ≥1400°C. However, at higher temperatures, when compared to TiC, a smaller growth rate for Ti₂AlC may indicate that Ti₂AlC experienced further decomposition into TiC via the sublimation of Al, similar to decomposition of Ti₃SiC₂. In contrast to Ti₃AlC₂, no intermediate or lower order phase was observed for the decomposition of Ti₃SiC₂. This difference can be attributed to the fact that Ti₃SiC₂ is the only stable ternary phase in Ti-Si-C diagram.

Research conducted in our laboratories showed that a weight loss of up to 20% was observed as a result of decomposition for all the MAX phases can be attributed to the release of gaseous Al by sublimation during the decomposition process because the vapour pressures of the A elements exceed the ambient pressure of the furnace (i.e. ≤5x10^-5 torr) at ≥ 1500°C. Since the vapor pressure of a substance increases non-linearly with temperature according to the Clausius-Clapeyron relation, the volatility of A elements will increase with any incremental increase in temperature .

(a)

(b)

(c)

(d)

Figure 1. Phase abundance as a function of temperature for the decomposition of (a) Ti₃SiC₂, (b) Ti₃AlC₂, (c) Ti₂AlC, and (d) Ti₂AlN in vacuum.

It is well known that A elements such as Si and Al have high vapour pressures and become volatile at elevated temperatures. Thus, at temperatures of well over 1500°C used in this study, both Al and Si should become volatile and sublime readily and continuously in a dynamic environment of high vacuum. When the vapor pressure becomes sufficient to overcome the ambient pressure in the vacuum furnace, bubbles will form inside the bulk of the substance which eventually appear as voids on the surface of decomposed MAX phase. The evidence of surface voids formation can be clearly discerned from the scanning electron micrographs of decomposed MAX phases shown in Figure 2 . Since Si has a lower vapour pressure than Al, it helps to explain why Ti₃SiC₂ is more resistant to decomposition than Ti₃AlC₂ or Ti₂AlN. In all cases, the kinetics of decomposition process are driven mainly by a highly restricted out-diffusion and sublimation of high vapour pressure A element (e.g. Al, Si) from the bulk to the surface of the sample and into the vacuum, i.e.,

M_n+1AX_n ---> M_n+1X_n + A

M_n+1X_n ---> (n+1)MX_n/(n+1)

Figure 2. Scanning electron micrographs of the surface microstructures of vacuum-decomposed MAX phases; (a) Ti₂AlN, (b) Ti₄AlN₃, (c) Ti₃SiC₂ and (d) Ti₃AlC₂.

Role of Pore Size on Decomposition Kinetics

Research conducted in our laboratories show that all the calculated activation energies are positive except for bulk Ti₃AlC₂. However, when Ti₃AlC₂ powder was used a positive activation energy was obtained which implies the importance of pore microstructures in the decomposition kinetics. Table 1 shows that the activation energies calculated from the Arrhenius equation for several MAX phases and the proposed reactions. A negative activation energy indicates that the rate of decomposition in Ti₃AlC₂ decreased with increasing temperature due to the formation of a dense TiC surface layer with very fine pores (<1.0 µm) which exert an increasing resistance to the sublimation process as the temperature increases (Fig. 2d). In contrast, a more porous decomposed layer with coarser pores (>2.0 µm) formed in other MAX phases and in powdered Ti₃AlC₂ which enabled the sublimation of Al or Si to progress with minimum resistance, thus resulting in an increasing rate of decomposition with temperature. In short, the pore sizes play a critical role in determining the value of activation energy and the rate of decomposition. Thus, the ability to manipulate the pore microstructure either through densification to reduce pore-size or engineering of pore-free microstructures will allow the process of decomposition in MAX phases to be minimized or arrested.

Table 1. Comparison of the kinetics of thermal decomposition in MAX-phase samples.

All the calculated activation energies are positive except for Ti₃AlC₂. A negative activation energy indicates that the rate of decomposition in Ti₃AlC₂ decreased with increasing temperature due to the formation of a TiC surface layer with very fine pores (<1.0 µm) which exert an increasing resistance to the sublimation process as the temperature increases. In contrast, coarser pores (>5.0µm) formed in other MAX phases which enable the sublimation of Al or Si to progress with minimum resistance and thus an increasing rate of decomposition with temperature. In short, the pore sizes play a critical role in determining the value of activation energy and the rate of decomposition.

The kinetics of isothermal phase decomposition as modelled using the Avrami equation and the Avrami exponents (n) of isothermal decomposition of the MAX phases are shown in Table 2. The low values (i.e. <1) of the exponent indicate that in all cases the decomposition is a highly restricted diffusion process, presumably of Al or Si from the bulk of the sample to its surface.

Table 2. Comparison of the Avrami decomposition kinetics in MAX phases.

Outlook

Before MAX phases can be widely used in extreme environments, the issues pertaining to their susceptibility to thermal decomposition need to be fully addressed. In particular, there remain several unresolved issues relating to the phase and thermal stability that require further study:

• The vapour pressure of element A is critical to the phase stability of MAX phases. The higher the vapour pressure of element A, the more susceptible the MAX phase is to phase dissociation at elevated temperature.
• The Avrami kinetics of phase dissociation is dependent on the rate of removal of vapourised element A. A dynamic atmosphere with a flowing gas or in high vacuum will facilitate the continual removal of the vapourised A and thus the continuous dissociation of the MAX phase. In contrast, a static atmosphere is expected to be most conducive for a MAX phase to resist phase dissociation.
• The role of microstructural modification due to phase dissociation on the mechanical performance of MAX phases. It remains unknown how microstructural changes will affect the mechanical properties. New stabilizers will be formulated to arrest the susceptibility of MAX phases to thermal dissociation at elevated temperature. TiSi₂ is an effective stabilizer for Ti₃SiC₂.
• Development of improved models to describe the chemical processes and kinetics of phase dissociation. No such models exist currently that can adequately describe and predict the property modification, especially for the ternary nitrides.

Acknowledgements

Fundings from the Australian Research Council ( DP0664586, LX0774743, LE0882725), ISIS and AINSE (P329, P606, P1431 & MI1488 ) are acknowledged. Also thanked are contributions from Drs W.K. Pang, R. Smith, V. Peterson, S. Kennedy and E/Prof. B. O’Connor.

1. M.W. Barsoum and T. El-Raghy, Am. Sci.89 , 334-343 ( 2001).
2. I.M. Low, et al., J. Am. Ceram. Soc.81, 225 (1998).
3. I.M. Low, J. Eur. Ceram. Soc. 18 , 709 (1998).
4. I.M. Low, Mater. Lett . 58 , 927-930 (2004).
5. I.M. Low and W.K. Pang, Mater. Aust. Mag.6 , 33-35 (2011) .
6. I.M. Low, et al., J. Am.Ceram.Soc. 90 , 2610 (2007).
7. I.M. Low et al., Physica B, 385-386, 499-501 (2006).
8. W.K. Pang, I.M. Low, et al., J. Alloys Compds.509, 172-176 (2010).
9. W.K. Pang, I.M. Low, et al., J. Physics: Conference Series, 251 , 012025 (2010) .
10. W.K. Pang and I.M. Low, J. Aust. Ceram. Soc. 45 , 39-43 (2009).
11. W.K. Pang, I.M. Low and Z.M. Sun, J Am. Ceram. Soc.93, 2871-2876 (2010).
12. I.M. Low, et al., J. Eur. Ceram. Soc.31, 159-166 (2011).
13. H. Zhang, et al., J. Am. Ceram. Soc .91, 494 (2008).
14. A. Feng, et al., J. Mater. Res. 14, 925 (1999).
15. N.F. Gao, et al., Mater. Lett. 55 , 61 (2002).
16. J. Emmerlich, et al., Acta Mater. 55, 1479 (2007).
17. M.W. Barsoum and T. El-Raghy, J.Am. Ceram. Soc.79 , 1953 (1996).
18. N. Tzenov, et al., J. Eur. Ceram. Soc. 20, 801 (2000).
19. R. Radakrishnan, et al., J. Alloys Compd. 285, 85 (1999).