CeO 2 induced dispersive distribution of TiB 2 particles in in situ TiB 2 /Al composite

Abstract

The roles of CeO₂ additive during preparation of in situ TiB₂/Al composite, alleviating particle settlement in composite melt and significantly improving particle dispersion in final microstructure, are studied in this paper. It is evidenced that the CeO₂ additive reacts with Al melts to release Ce solute into the melts, and the released surface active Ce is absorbed in the Al/TiB₂ interfaces without any other reaction products. First principles calculations show that the interfacial energy of Al/TiB₂ interfaces is reduced owing to the presence of Ce in Al/TiB₂ interfacial area. Therefore, the wettability of molten Al on TiB₂ surface is increased and the dispersion of TiB₂ particles in Al matrix is eventually improved.

Keywords

Metal matrix composites Dispersion Interface First principles interfacial energy calculations

Introduction

The Al–Ti–B master alloy and in situ TiB₂/Al composite prepared by an exothermic reaction among K₂TiF₆, KBF₄ salts and Al melts have been extensively investigated due to good grain refining performance of master alloy^1–3 and higher specific strength and good wear resistance of composite.^4–6 A large amount of information^7–12 on microstructures indicates that TiB₂ particles distribute along grain boundaries in a form of particle agglomerations. It is easier for the larger particle agglomerations to deposit in the composite melt, which results in the formation of depleted and aggregative zones of TiB₂ particles in the solidified casting. It was reported that when the composite is produced with high volume fraction TiB₂, the particle agglomerations will weaken the reinforcing effects of TiB₂ particles, which accordingly reduces its final properties.^13,14 Thus, the distribution control of TiB₂ particles in aluminium matrix is important for producing TiB₂/Al composite with high specific modulus and strength and Al–Ti–B master alloys with good grain refining performance via casting technique.^10–15 Han et al.¹⁰ indicated that the morphology of TiB₂ agglomerations in the Al–Ti–B alloys could be changed into loose coral-like by applying high intensity ultrasound in the melts. Jha and Dometakis¹⁵ reported that Mg and Zr as surface active additives could improve the dispersion of TiB₂ particles in aluminium matrix, and suggested that the decrease in interfacial energy of Al/TiB₂ interfaces by adding surface active elements is responsible for the dispersion improvement. However, their inferences need experimental or theoretical evidence. Meanwhile, our report¹⁴ has shown that the addition of CeO₂ during preparation of TiB₂/Al composite is helpful for alleviating the deposition of TiB₂ particles in the melts and accordingly improves the particle distribution in the final microstructure. However, the role of CeO₂ in the composite has not been analysed in detail. Thus, although the importance of interfacial energy of particle/matrix on the particle dispersion in in situ and ex situ composites has been proposed,^14–17 the addition of surface active element on interfacial energy is necessary to be experimentally or theoretically investigated further.

The exploration of interfacial energy of Al–Ce/TiB₂ interfaces is crucial to exactly understand the action mechanism of CeO₂ in TiB₂/Al composite. To date, no applicable experimental methods are available to quantitatively investigate the metal/reinforced particle interfacial energy in composite melt. Recently, the first principles method has been successfully applied in the study on metal/ceramic interfacial adhesion, such as Al/TiB₂, Al/AlB₂, Al/Al₃Ti, Al/Mo₂B, Nb/Al₂O₃, Ni/ZrO₂, Ni/Al₂O₃, Fe/WC, Al/α–Al₂O₃, Al/WC and Al/TiN,^18–27 which can not only accurately provide the atomic structure of interface but also quantitatively predict the interfacial energy. It is the primary purpose of this paper to examine the effects of CeO₂ additive on the dispersion of TiB₂ particles in in situ TiB₂/Al composite by analysing the existent forms and distribution of Ce element and its influence on the interfacial energy of Al/TiB₂ interfaces using first principles methods.

Experimental and computational details

Experimental procedure

The in situ 5 vol.- TiB₂/Al composite was prepared by an exothermic reaction process among K₂TiF₆, KBF₄ and Al melts. The composite was remelted, and 0·5 wt- CeO₂ powder was added to the melt at 1103 K. The melt was held for 15 min and then was cast into graphite mould at 1023 K. To explore the roles of CeO₂ additive in TiB₂/Al composite, the composites with and without CeO₂ were remelted in Al₂O₃ crucible with φ30×100 mm and held for 10 min at 1053 K in a 5 kW electrical furnace. Then, the crucible was removed from the furnace and rapidly quenched in cold water. The distribution of TiB₂ particles in composite melt was analysed by observing the microstructure of samples taken at the same height of two remelting composite rods.

A Sirion 200 field emission scanning electron microscope (FE-SEM) was employed to study particle distribution in Al matrix, and a Nova Nano 230 FE-SEM equipped with energy dispersive X-ray spectroscopy (EDS) attachment was employed to study distribution of Ce element in the composite. Microstructural analysis of composite was carried out on a JEM-2010F field emission transmission electron microscope (FE-TEM); thin foils were prepared by the ion milling technique.

Computational details and interface model geometry

Calculations of the interfacial energy of Al–Ce/TiB₂ interfaces were based on our earlier report on Al/TiB₂ interfaces.¹⁸ The interface orientation relationship of was also adopted in this paper.^28–30 A (1×1) interfacial supercell with a five-layer Al (111) slab with Ce atom being placed on a symmetric seven-layer TiB₂ (0001) slab was used, which is thick enough to ensure the bulk-like interior interface.^18,31 We have proved that the stacking sequence of Al/TiB₂ interfaces in the steady state Al–5Ti–1B alloy was Ti terminated HCP, wherein the interfacial Al atoms reside atop the second layer sites of TiB₂.¹⁸ Meanwhile, the surface active element, Ce, prefers to localise in the interfacial area. Thus, in the present calculation, the model of Al–Ce/TiB₂ interfaces was established by replacing an interfacial Al atom with a Ce atom in the Al(111)/TiB₂(0001) interface with the stacking sequence of Ti terminated HCP, as shown in Fig. 1. The free surfaces of the Al and TiB₂ slabs with the in plane periodicity were separated by a 10 Å vacuum to prohibit their interactions. To compensate the mismatch between TiB₂ and Al surfaces and satisfy the periodic boundary conditions in a supercell calculation, the coherent interface approximation was adopted,^25,26 where the Al slab was modestly stretched by 5·75 to be commensurate with the TiB₂ slab. All atoms in the supercells were allowed to freely relax in three directions during geometry optimisation.

Model of Al–Ce/TiB₂ interface atomic structure

First principles calculations were performed using the total energy plane wave pseudopotential method based on the density functional theory with the generalised gradient approximation of Perdew et al.^32,33 The minimum total energy was achieved by automatically relaxing the internal coordinates in all directions using Broyden–Fletcher–Goldfarb–Shanno algorithm.³⁴ The ultrasoft pseudopotential was used for Ce atoms, which generated that the atomic configuration was 4f¹5d¹6s². A plane wave cutoff energy of 380 eV was employed in the calculations, which assured a total energy convergence of 10⁻⁶ eV atom⁻¹. Brillouin zone sampling was performed using 4×4×4 and 1×1×1 Monkhorst–Pack k point meshes for Ce bulk and Al–Ce/TiB₂ interface calculations respectively.

Results and discussion

Effect of CeO₂ additive on microstructure of TiB₂/Al composite

Figure 2 shows the SEM microstructures of the TiB₂/Al composites with and without CeO₂ after remelting and rapidly quenching in cold water. TiB₂ particles in the composite melt without CeO₂ are in a form of agglomerations, and the agglomeration size is ∼5–15 μm, as shown in Fig. 2a and c. With CeO₂ being added into composite melt, the distribution of TiB₂ particles is evidently improved, as shown in Fig. 2b and d. Most of TiB₂ particles uniformly distribute in the matrix, and the size of single particle is <2 μm. It is indicated that the role of CeO₂ additive is to improve the dispersion of TiB₂ particles in composite melt.

Microstructures (SEM) of composite a, c without and b, d with CeO₂ after remelting and rapidly quenching in cold water

In the case of metal matrix composites, the dispersion of reinforced particles within the matrix depends simultaneously on the moving velocity of particles in melt and the solidification rate of matrix alloy during solidification process. When the solidification rate of matrix alloy is larger than the moving velocity of particles, the particles are captured by the advancing solid/liquid interface and uniformly distribute within the final microstructure. On the contrary, the particles are pushed by the advancing solid/liquid interface and eventually segregate in the last solidification regions, such as grain boundaries and interdendritic regions. According to the stokes equation on the moving velocity of particle in melt:³⁵ (1) where η is the viscosity of molten metal (Pa s), d is the diameter of reinforcing particle (m), g is the gravity acceleration (9·8 m s⁻²) and ρ_m and ρ_p are the densities of molten metal and particle (kg m⁻³) respectively. It can be calculated that the moving velocity of particle agglomeration with a size of 10 μm in composite melt is 25 times faster than that of 2 μm isolated particle. Thus, compared with the particle agglomerations, the isolated TiB₂ particles in the composite melt with CeO₂ additive are easier to be captured by the advancing solid/liquid interface during solidification, which results in the uniform distribution of TiB₂ particles in the solidified casting.

With CeO₂ being added into the melts, the following reaction may happen: (2)

The Gibbs free energy for the above reaction at 1100–1200 K can be calculated by the following equation: (3)

When the CeO₂ additive is added, the melt temperature is 1103 K in the experiment. The thermodynamic analysis shows that CeO₂ will react with Al to release Ce solute in the composite melt.³⁶ The SEM microstructure and element planar distribution for a polished surface of the composite with CeO₂ additive are shown in Fig. 3. It can be confirmed by the distributions of Ti and B elements that the polygonal particles in Fig. 3a are TiB₂. Meanwhile, the released Ce mainly distributes on the surface of TiB₂ particles, that is, the Al/TiB₂ interfaces, as shown in Fig. 3e and f. Although the Ce solute is absorbed in the Al/TiB₂ interfaces, the interfaces are clean and directly bonded without any other reaction products as observed by TEM of interface in Fig. 4.

a SEM microstructure and b–f element planar distribution of polished surface of composite with CeO₂

Bright field images (TEM) of a composite with CeO₂, b corresponding selected area diffraction pattern and c Al/TiB₂ interface

Interfacial energy calculation of Al–Ce/TiB₂ interfaces

To further analyse the effect of surface active Ce element on the interfacial energy of Al/TiB₂ interfaces in a thermodynamic sense, the interfacial energy of Al–Ce/TiB₂ interfaces was calculated. The calculation methodology in the literature^25,31,37 was employed in this paper, wherein the chemical potentials of Ti, B and Ce atoms were taken into account. The interfacial energy, γ, can be given by the following equation: (4) where A is the interface area; E_total is the total energy of a fully relaxed interface supercell; N_B, N_Ti, N_Al and N_Ce are the numbers of the B, Ti, Al and Ce atoms respectively; , and are the chemical potentials of the bulk TiB₂, Al and Ce respectively; is the chemical potentials of Ti in the slab; and and σ_Al are the surface energies of TiB₂ and Al respectively. Parameters used for the calculation of interfacial energy are given in Table 1. The interfacial energy of Al–Ce/TiB₂ interfaces can be obtained by equation (4) and is compared with that of the Al/TiB₂ interfaces reported in our earlier study,¹⁸ as shown in Fig. 5. It is seen that the interfacial energy of Al–Ce/TiB₂ interfaces is lower than that of Al/TiB₂ interfaces over the entire range of the Ti chemical potential.

Interfacial energies of Ti terminated HCP Al/TiB₂ interfaces and Al–Ce/TiB₂ interfaces as function of , i.e. difference between Ti chemical potential in slab and its value for pure Ti bulk phase; vertical dashed lines indicate ultimate Ti chemical potential values where bulk TiB₂ could form

Table 1

Parameters used for calculation of interfacial energy of Al–Ce/TiB₂ interfaces

Calculated values in present investigation			Calculated values in our earlier study^18,31
A/Å²	E_total/eV	/eV	/eV	/eV	/eV	σ_Al/J m⁻²	/J m⁻²
31·76	−29669·61	−1059·50	−2817·228	−56·495	−1603·122	0·82	See Ref. 31

According to Young's equation^38,39 (5) where σ_S/V, σ_S/L and σ_L/V are the interfacial energies of solid/vapour, solid/liquid and liquid/vapour interfaces respectively. It is obvious that the smaller the interfacial energy σ_S/L is, the smaller the contact angle θ is. As indicated in the interfacial energies versus the Ti chemical potential (Fig. 5), the interfacial energy of Al/TiB₂ is reduced with the presence of Ce. Therefore, the contact angle between molten Al and TiB₂ particle surface is decreased when Ce element exists, which leads to the improvements in the wettability of molten Al on TiB₂ and the dispersion of TiB₂ particles in composite melt.^38–40 Thus, the calculation of interfacial energy of Al–Ce/TiB₂ interfaces confirms theoretically that improvement in the dispersion of TiB₂ particles in composite with CeO₂ is mainly attributed to the decrease in interfacial energy of Al/TiB₂ interfaces due to adsorption of the surface active Ce element in interfaces.

Conclusions

The roles of CeO₂ additive in TiB₂/Al composite have been investigated by experiment and first principles methods. It is found that TiB₂ particles in the composite melt without CeO₂ additive are in a form of agglomerations, while the particle dispersion is evidently improved and the isolated TiB₂ particles exist in composite melt with CeO₂ being added. The CeO₂ additive reacts with Al melts to release Ce solute in the composite melt. The absorption of the released surface active Ce element in the Al/TiB₂ interfaces without any other reaction products is responsible for the reduction of the interfacial energy of Al/TiB₂ interfaces. Thereby, the wettability of molten Al on TiB₂ surface is improved and the dispersion of TiB₂ particles in composite melt is improved. It is proved that the interfacial energy of particles/matrix interfaces plays an important role in governing the dispersion and agglomeration of reinforced particles in the metal matrix.

Footnotes

Acknowledgments

The authors wish to acknowledge the support of the National Natural Science Foundation of China (grant nos. 51001074 and 51274141) and National High Technology Research and Development Program of China (863 project) (grant no. 2008AA03A239). At the same time, the authors are grateful to the Instrumental Analysis Center of Shanghai Jiao Tong University for help in microstructure analysis.

References

Han

, Shu

, Wang

and Sun

: Mater. Sci. Eng. A , 2006, A430, 326–331.

Murty

, Kori

, Venkateswarlu

, Bhat

and Chakraborty

: J. Mater. Process. Technol. , 1999, 89–90, 152–158.

and Liu

: J. Alloys Compd , 2006, 425, 245–250.

Han

, Liu

and Bian

: Composites A , 2002, 33A, 439–444.

Mandal

, Chakraborty

and Murty

: Wear , 2007, 262, (1–2), 160–166.

Wood

, Davies

and Kellie

JLF

: Mater. Sci. Technol. , 1993, 9, 833–840.

Feng

and Froyen

: J. Mater. Sci. , 2000, 35, 837–850.

Mohanty

and Gruzleski

: Acta Metall. Mater. , 1995, 43, 2001–2012.

Lakshmi

, Lu

and Gupta

: J. Mater. Process. Technol. , 1998, 73, 160–166.

10.

Han

, Li

and Wang

: Mater. Sci. Eng. A , 2005, A405, 306–312.

11.

Youssef

, Dashwood

and Lee

: Composites A , 2005, 36A, 747–763.

12.

Watson

, Forster

, Lee

, Dashwood

, Hamilton

and Chirazi

: Composites A , 2005, 36A, 1177–1187.

13.

Tszeng

: Composites B , 1998, 29B, 299–308.

14.

Xue

, Wang

and Han

: J. Alloys Compd , 2011, 509, 1573–1578.

15.

Jha

and Dometakis

: Mater. Des. , 1997, 18, 297–301.

16.

Schaffer

, Miller

and Dahle

: Scr. Mater. , 2007, 57, 1129–1132.

17.

Luo

: Acta Mater. , 2006, 54, 47–58.

18.

Han

, Dai

, Shu

, Wang

and Sun

: Appl. Phys. Lett. , 2006, 89, 1–3.

19.

Han

, Dai

, Wang

, Shu

and Sun

: Appl. Surf. Sci. , 2011, 257, 7831–7836.

20.

, Yang

and Luo

: Comput. Mater. Sci. , 2012, 62, 136–141.

21.

Si Abdelkader

and Faraoun

: Comput. Mater. Sci. , 2011, 50, 880–885.

22.

Duffy

, Harding

and Stoneham

: Acta Mater. , 1996, 44, 3293–3298.

23.

Han

, Zhang

and Xu

: Chem. Phys. Lett. , 2003, 378, 269–272.

24.

, Gao

, Xiao

, Min

, Ma

and Yi

: Appl. Surf. Sci. , 2011, 257, 5671–5678.

25.

Siegel

, Hector

Jr and Adams

: Phys. Rev. B , 2003, 67B, 1–4.

26.

Siegel

, Hector

Jr and Adams

: Surf. Sci. , 2002, 489, 321–336.

27.

Liu

, Wang

and Ye

: Acta Mater. , 2004, 52, 3681–3688.

28.

Quested

: Mater. Sci. Technol. , 2004, 20, 1357–1369.

29.

Easton

and Stjohn

: Metall. Mater. Trans. A , 1999, 30A, 1613–1623.

30.

Zhang

, Kelly

, Easton

and Taylor

: Acta Mater. , 2005, 53, 1427–1438.

31.

Han

, Dai

, Shu

, Wang

and Sun

: J. Phys. Condens. Matter , 2006, 18, 4197–4205.

32.

Segall

, Lindan

PJD

, Probert

, Pickard

, Hasnip

, Clark

and Payne

: J. Phys. Condens. Matter , 2002, 14, 2717–2744.

33.

Perdew

, Burke

and Ernzerhof

: Phys. Rev. Lett. , 1996, 77, 3865–3868.

34.

Fischer

and Almlof

: J. Phys. Chem. , 1992, 96, 9768–9744.

35.

Sohrabi Baba Heidary

and Akhlaghi

: Acta Mater. , 2011, 59, 4556–4568.

36.

Liu

and Xu

: Mater. Lett. , 2004, 58, 432–436.

37.

Zhang

and Smith JR : Phys. Rev. B , 2000, 61B, 16883–16889.

38.

Sangghaleh

and Halali

: Appl. Surf. Sci. , 2009, 255, 8202–8206.

39.

Liu

, Muolo

, Valenza

and Passerone

: Ceram. Int. , 2010, 36, 1177–1188.

40.

Kennedy

, Karantzalis

and Wyatt

: J. Mater. Sci. , 1999, 34, 933–940.

CeO 2 induced dispersive distribution of TiB 2 particles in in situ TiB 2 /Al composite

Abstract

Keywords

Introduction

Experimental and computational details

Experimental procedure

Computational details and interface model geometry

Results and discussion

Effect of CeO2 additive on microstructure of TiB2/Al composite

Interfacial energy calculation of Al–Ce/TiB2 interfaces

Conclusions

Footnotes

Acknowledgments

References

Effect of CeO₂ additive on microstructure of TiB₂/Al composite

Interfacial energy calculation of Al–Ce/TiB₂ interfaces