Analysis of local composition gradients in the hard-phase grains of cermets using a combination of X-ray diffraction and electron microscopy

https://doi.org/10.1016/j.ijrmhm.2007.03.004Get rights and content

Abstract

Combination of X-ray diffraction (XRD) and scanning electron microscopy (SEM) was used for investigation of local composition gradients in the hard-phases of cermets. XRD revealed distribution of lattice parameters in hard-phase grains, from which the composition gradients in the hard-phases were estimated using an appropriate microstructure model. This microstructure model was build with the aid of SEM micrographs, which were taken with back-scattered electrons (BSE) and completed by the registration of the electron back-scatter diffraction (EBSD) patterns and characteristic X-ray spectra. SEM/BSE yielded the first information about the spatial distribution of elements in individual hard-phase grains, SEM/EBSD about the morphology, the size and the size distribution of these grains. The final interpretation of the distribution of lattice parameters, which was obtained from the X-ray line profile analysis, was done with the aid of the local elemental analysis that was performed using SEM with the energy dispersive analysis of the characteristic X-ray spectra (EDX) and the known dependence of the lattice parameters on concentration.

Introduction

Development of new materials with specific properties requires a detailed knowledge of their microstructure, as the materials properties are always related to their microstructural features. A comprehensive review about the correlation between the microstructure and properties of WC/Co hardmetals was recently published by Roebuck [1], who has shown how the size of WC grains and the Co contents influence various materials properties, such as, for example, hardness. The average grain size and the grain size distribution are the most frequently investigated microstructure parameters in hardmetals, see, e.g. [2], [3], [4]. In ceramic–metal composites (cermets), the phase composition, the size and morphology of grains, the local chemical composition and/or the local composition gradients are typically the subject of microstructural studies, see, e.g. [5], [6], [7], [8], [9].

The advantages and disadvantages of the experimental methods that are typically used for microstructural analysis on hardmetals and cermets, i.e. the optical microscopy, X-ray diffraction (XRD), scanning electron microscopy (SEM) with back-scattered electrons (BSE), scanning transmission electron microscopy (STEM) with the electron energy loss spectroscopy (EELS), secondary ion mass spectrometry (SIMS), Auger electron spectroscopy (AES) and the atom probe field ion microscopy (APFIM), were discussed in a review paper by Roebuck and Gee [10], who have shown that particularly the determination of composition gradients in materials with grain sizes below 1 μm is a difficult task. For determination of the composition or concentration gradients with a high lateral resolution, the use of the energy-filtered transmission electron microscopy (EFTEM) was proposed in [11]. This experimental method yields the composition gradients with a very high lateral resolution, but the preparation of the samples for TEM is cumbersome and the results are not always statistically reliable, which is due to the high magnification of the method and a very small volume of the sample.

Another difficulty of the analytical methods using the interaction of electrons with matter is a high absorption of electrons in solids, which strongly depends on the atomic number, on the energy of electrons and on the kind of interaction between electrons and matter. For these reasons, the determination of the penetration depths of primary electrons in SEM or their absorption in TEM is no simple task. According to [12], [13], the penetration depth of primary electrons can be approximated byxe[μm]=0.033ρE01.7-EC1.7,where E0 is the energy of primary electrons, EC the energy of the absorption edge of the respective species (both in keV) and ρ the materials density in g/cm3. This approximation was verified by Monte Carlo simulations [14]. For E0 = 20 keV, the penetration depths calculated using Eq. (1) range between 0.06 and 1.0 μm for NbC and TiC, respectively. For E0 = 25 keV, the range of the penetration depths calculated for these materials is between 0.38 and 1.5 μm. The choice of the primary energy is a critical issue, particularly for elements having their absorption edge close to the energy of primary electrons, e.g. for Mo with EC =  19.9995 keV.

In the 1970s, Houska [15], [16] had shown that XRD is an appropriate method for analysis of composition gradients in solids. The analysis of composition gradients using XRD exploits a very high resolution of XRD in the reciprocal space, which allows small differences in the interplanar spacings to be distinguished with a high precision. For materials in which the dependence of the interplanar spacing on the chemical composition is known, the composition gradients can be determined from the distribution of the interplanar spacings with the aid of an appropriate microstructure model, which takes into account the size and shape of individual grains and the kind of chemical or physical processes running during the production of the materials. However, due to the poor lateral resolution of XRD in direct space, this technique cannot assign the observed distribution of the interplanar spacing to the individual microstructural features present in the material under study. Thus, the low lateral resolution of XRD makes the determination of the composition profiles solely from the XRD data impossible, and a microstructure model for evaluation of the XRD patterns is necessary.

In order to create and refine microstructure model of composite materials with the grain size in micrometer and sub-micrometer range, we used a combination of XRD, SEM with registration of back-scattered electrons (SEM/BSE), SEM with registration of the electron back-scatter diffraction patterns (SEM/EBSD) and SEM with energy-dispersive analysis of the X-ray spectra (SEM/EDX). The microstructure model of cermets describes lateral distribution of lattice parameters assuming a certain distribution of elements in hard-phase grains, from which the composition gradients on microscopic scale are concluded. The capability of this method is illustrated on cermets with different chemical composition and microstructure. For sure, the combination of XRD, SEM/BSE, EBSD and EDX cannot compete with the local methods for chemical analysis regarding the local resolution and the absolute precision of the results. However, it combines experimental methods that are widely available, do not need a special sample preparation, have a good statistics and are less sensitive to the absorption phenomena, which is true particularly for XRD. On the other hand, the knowledge of the distribution of the lattice parameters and the local changes in the interplanar spacing is needed for prediction of the formation of microstructure defects and for estimation of the fracture behavior of composite materials.

Section snippets

Information contents of the XRD patterns

In crystalline materials, XRD is primarily used for qualitative and quantitative phase analysis, i.e. for phase identification and for quantification of the phase composition, and for precise determination of lattice parameters. Besides, the XRD patterns contain relevant microstructural information about lattice strain and local composition gradients that can be obtained from distribution of the interplanar spacing, which controls the broadening of the XRD lines [24], [25]. Because of the high

Microstructure model

The microstructure model of cermets used here assumes that individual hard-phase grains contain composition gradients [6], [9] and crystal lattice defects. Both microstructure features, i.e. composition gradients and crystal lattice defects, are responsible for local variations of the interplanar spacing and thus for the physical broadening of XRD lines. For cermets consisting of sub-micrometer hard-phase grains, mainly dislocations and small crystallites can be regarded as relevant crystal

Experimental details

Two series of samples with different chemical compositions were prepared and investigated in this study. Samples of both series were based on the (Ti, W)(C, N) + (Co, Ni) cermets. For preparation of samples of the series A, powders of Ti(C, N), (Ta, Nb)C, WC, Cr3C2, Co and Ni were employed. Samples of the series B were made from the mixture of Ti(C, N), TiC, Mo2C, WC, Co and Ni. The amount of binder was nearly the same in both sample series. The particle size of the starting materials was between 1 and

Phase identification

In the sample A, XRD identified three distinct crystalline phases on the basis of their different lattice parameters and the positions of diffraction lines: face-centered cubic (fcc) Co–Ni solid solution with the lattice parameter (0.35730 ± 0.00005) nm, a second fcc phase with the average lattice parameter (0.43108 ± 0.00003) nm, and a third fcc phase with the average lattice parameter (0.42808 ± 0.00003) nm. The fcc Co–Ni solid solution builds the binder; the other two fcc phases are contained in the

Summary

Comparison of results obtained for samples with different chemical compositions confirmed a high sensitivity of this combination of experimental methods to the microstructural changes. Sample A, which is based on the (Ti, W)(C, N) + (Co, Ni) cermet and contains additionally Ta, Nb and Cr, consisted of Ti-rich regular grains and Ti-depleted inverse grains contained in the hard phase and (Co, Ni) solid solution contained in the binder. The lattice parameter in the regular grains ranged between 0.4279

Conclusions

Combination of X-ray diffraction and scanning electron microscopy with analysis of back-scattered electrons, back-scattered electron diffraction patterns and characteristic X-rays was shown to be a suitable technique for determination of the microstructure features in the hard-phase grains of composite materials like distribution of lattice parameters, average grain size, grain size distribution and composition gradients in individual grains. This combination of complementary experimental

Acknowledgement

The authors acknowledge the financial support of this work through the German Research Council (DFG) under the project # KL 1274/4-1.

References (40)

  • B. Wollein et al.

    X-ray mapping of microstructures in hardmetals and cermets

    Surf Interface Anal

    (2002)
  • W. Lengauer

    Diffusional control of the near-surface microstructure in functional gradient hardmetals

    Mat-wiss Werkstofftech

    (2005)
  • G. Östberg et al.

    Microstructure changes during wear by plastic deformation of cemented carbide and cermet cutting inserts

    Metal Mater Trans A

    (2006)
  • B. Roebuck et al.

    TiC and Ti(C, N) cermet microstructures

  • H.A. Bethe et al.

    Proc Am Phil Soc

    (1938)
  • P. Duncumb
  • R. Shimizu et al.

    The Monte Carlo technique as applied to the fundamentals of EPMA and SEM

    J Appl Phys

    (1972)
  • C.R. Houska

    X-ray diffraction from a binary diffusion zone

    J Appl Phys

    (1970)
  • D.R. Tenney et al.

    X-ray diffraction technique for investigation of small diffusion zones

    J Appl Phys

    (1970)
  • A.R. Stokes

    A numerical Fourier-analysis method for the correction of widths and shapes of lines on X-ray powder photographs

    Proc Phys Soc London

    (1948)
  • Cited by (12)

    • The role of graphite addition on spark plasma sintered titanium nitride

      2020, Journal of Materials Research and Technology
    • Affordable real-time diffraction compensation in the spatial domain

      2017, Optik
      Citation Excerpt :

      The effect of diffraction is well-known and in the paraxial approximation can be reduced to a simple convolution with a two-dimensional kernel function [4,5]. Therefore, in cases where such compensating devices are impractical (such as x-rays lenses) it is advantageous to use deconvolution to reconstruct the diffracted image [6,7]. However, since diffraction occurs in the field domain, the deconvolution has to occur in the field domain as well.

    • Characterisation of Ti(C, N)-based cermets with various nitrogen contents studied by EBSD/SEM and TEM

      2017, Journal of Alloys and Compounds
      Citation Excerpt :

      The XRD results in Fig. 2 show that the cermets consists of dual phases, Ti(C, N) and Ni; both dual phases have the face centered cubic (FCC) NaCl-type crystal structure, which are consistent with P. Duwez and G. Levi's suggestion [30,31]. Due to D. Rafaja's suggestion [21], assignment of the lattice parameters to the chemical composition in the system TiWMoCN was done using the idealized Vegard-like dependence in a linear approximation, the difference in the lattice parameters between the rim and core can be decreased by increasing W, Mo and C contents. The decreasing of the lattice parameter of Ti(C, N) hard phases with the increasing of the N content can be identified form XRD profiles, which is intensified by the remove of Mo and W. Due to that the change of the lattice parameter with the W and Mo contents is much smaller, only the C and N contents are considered.

    • Effect of submicron Ti(C,N) on the microstructure and the mechanical properties of Ti(C,N)-based cermets

      2011, International Journal of Refractory Metals and Hard Materials
      Citation Excerpt :

      Their results showed even a coreless microstructure and cermets with a higher Palmqvist fracture toughness but with lower hardness. Three distinguishable cubic phases—two hard phases and the binder—were detected by means of XRD (Table 3), confirming the results of Rafaja et al. [13]. The binder has a larger cell parameter than a pure Co/Ni alloy (which for grade A would be 3.497 Å and for grade B 3.513 Å), because various elements are dissolved in it.

    View all citing articles on Scopus
    View full text