Exploring How Dopants Strengthen Metal-Ni/Ceramic-Al₂O₃ Interface Structures at the Atomic and Electronic Levels

Abstract

The metal-based/ceramic interface structure is a key research focus in science, and addressing the stability of the interface has significant scientific importance as well as economic value. In this project, the work of adhesion, heat of segregation, electronic structure, charge density, and density of states for doped-M (M = Ti, Mg, Cu, Zn, Si, Mn, or Al) Ni (111)/Al₂O₃ (0001) interface structures are studied using first-principle calculation methods. The calculation results demonstrate that doping Ti and Mg can increase the bonding strength of the Ni–Al₂O₃ interface by factors of 3.4 and 1.5, respectively. However, other dopants, such as Si, Mn, and Al, have a negative effect on the bonding of the Ni–Al₂O₃ interface. As a result, the alloying elements may be beneficial to the bonding of the Ni–Al₂O₃ interface, but they may also play an opposite role. Moreover, the Ti and Mg dopants segregate from the matrix and move to the middle position of the Ni–Al₂O₃ interface during relaxation, while other dopants exhibit a slight segregation and solid solution in the matrix. Most remarkably, the segregation behavior of Ti and Mg induced electron transfer to the middle of the interface, thereby increasing the charge density of the Ni–Al₂O₃ interface. For the optimal doped-Ti Ni–Al₂O₃ interface, bonds of Ti–O and Ti–Ni are found, which indicates that the dopant Ti generates stable compounds in the interface region, acting as a stabilizer for the interface. Consequently, selecting Ti as an additive in the fabrication of metal-based ceramic Ni–Al₂O₃ composites will contribute to prolonging the service lifetime of the composite.

1. Introduction

Metal-based ceramic composites are advanced, interesting materials whose performances are based on a combination of metal and ceramics. They are widely used in industrial fields, such as aviation, energy, sensors, and automotive [1,2,3,4]. Metal-based ceramic composites mainly include a metal–ceramic coating and metal-based-reinforced ceramic particles. It is worth noting that as a novel Ni–Al₂O₃ composite, the interface between Ni and Al₂O₃ determines the stability and lifespan of the composite. Therefore, it is of great scientific significance to explore the properties of the Ni–Al₂O₃ interface.

Regarding experiments, most studies primarily deal with the production process of metal-based ceramic composites and assess their chemical stability and mechanical characteristics. For instance, research conducted by Martínez-Franco et al. [5] found that a Ni–Al₂O₃ composite prepared through the high-kinetic processing of ball milling and spark plasma sintering possessed excellent mechanical properties, making it a potential candidate for various high-temperature applications. Just recently, Gao et al. [6] showed that ceramic nanoparticle-enhanced multilayer Ni–Ni–Al₂O₃ coating structures exhibited superior toughness, wear resistance, and corrosion resistance. Moreover, it was discovered that this multilayer coating structure was less prone to cracking. Shen et al. [7] suggested that an Ni–Al₂O₃ coating pretreatment is expected to develop as a low-cost, novel, and effective method for promoting nitride. However, Fu et al. [8] suggested that the Ni–Al₂O₃ interface can crack and expand under hot conditions so that the material could fail early. Our early experiments [9] have also confirmed this. Currently, the biggest challenges faced by Ni–Al₂O₃ composites during manufacturing and use are the following: (1) poor wettability between the matrix and ceramic; and (2) weak bonding strength at the matrix–ceramic interface. Indeed, effective dopants would be an effective measure to strengthen the Ni–Al₂O₃ interface binding, as shown in Figure 1. Therefore, it is necessary to conduct a more in-depth exploration of the modification of the Ni–Al₂O₃ interface, especially at the atomic and electronic scales.

First-principle calculations based on quantum mechanics can study the properties of metal–ceramic interfaces, shedding light on the microscopic features at minimal material dimensions. Hocker et al. [10] applied first-principles calculations to predict the tensile properties of metal–ceramic interfaces at the electronic dimension. Our team [11,12] found that active alloying elements and rare earth elements can act as binders, reducing interfacial spacing and enhancing the bonding strength of the Fe–Al₂O₃ interface. Research by Chen et al. [13] has shown that doping Mg and Cu can make the charge distribution on the Al–Al₂O₃ interface uniform, thereby enhancing its mechanical properties. Chen et al. [14] successfully predicted the fracture process of the NiTi–Al₂O₃ interface, thereby obtaining the failure mechanism of the structures. Guo et al. [15] found that dopant-Ti can form strong Ti–O ionic bonds at the Ag–Al₂O₃ interface, enhancing the bonding strength of the interface. Consequently, it can be seen that employing the first-principles calculation method can obtain the properties of the interface and reveal the bonding mechanism of heterogeneous materials from the atomic and electronic scales, a task that is unachievable through experimental methods.

Only very recently, our group has conducted research on the properties of Ni–Al₂O₃ interfaces and discovered that there are three types of Ni–Al₂O₃ interfaces (single Al-terminated, double Al-terminated, and O-terminated), which have significant differences in their bonding strengths [16]. Among them, the single Al-terminated Ni–Al₂O₃ interface is the easiest to form in the actual environment, serving as the basic model for this research. This work primarily involves doping alloying elements into the single Al-terminated Ni–Al₂O₃ interface, aiming to optimize the interfacial environment and strengthen the interfacial bonding. The research details are presented below. Firstly, we utilized the work of adhesion to screen for dopants-M (M = Ti, Mg, Cu, Zn, Si, Mn, or Al), which are advantageous for the bonding of the Ni–Al₂O₃ interface. Following this, we undertook a detailed atomic and electronic structural analysis of the Ni–Al₂O₃ interfaces showing positive effects, thereby uncovering their micro-level characteristics. In the final stage, we conducted comprehensive testing on the best-performing doped-Ti Ni–Al₂O₃ interface conditions with the aim of elucidating the specific mechanisms behind the strengthening effects at this interface.

2. Results and Discussion

2.1. The Binding Strength of the Interface

Generally, the work of adhesion (abbreviated here as the W_ad) can be used to evaluate the binding strength of the interface semi-quantitatively, as one of the criteria for judging the bonding properties of heterogeneous materials. In this work, the W_ad of the doped-M (M = Ti, Mg, Cu, Zn, Si, Mn, or Al) and non-doped Ni–Al₂O₃ interface structures are calculated to reveal the effect of doping on the bonding of the interface, as shown in the following Equation (1) [17,18]:

$$W_{ad}={\displaystyle \frac{E_{Ni+M}+E_{A{l}_2{O}_3}-E_{Ni+M/A{l}_2{O}_3}}{A}}$$

(1)

Here, E_Ni+M and $E_{{Al}_2{O}_3}$ denote the total energies of the Ni-side with dopant-M and Al₂O₃-side, respectively. $E_{Ni+M/{Al}_2{O}_3}$ is the total energy of the doped-M Ni–Al₂O₃ interface structure, and A stands for the area of the Ni–Al₂O₃ interface.

Figure 2 shows the W_ad of the doped-M and non-doped Ni–Al₂O₃ interface structures. In contrast to the non-doped Ni–Al₂O₃ interface, the W_ad of the doped-M Ni–Al₂O₃ interfaces, in descending order, is the doped-Ti (4.23 J/m²) > doped-Mg (1.9 J/m²) > doped-Cu (1.29 J/m²) > non-doped (1.25 J/m²) > doped-Zn (0.85 J/m²), showing that Ti and Mg has a positive effect on strengthening the bonding of the Ni–Al₂O₃ interface. Yet, the Zn dopant weakens the bonding strength between the Ni and Al₂O₃. More noteworthy, the W_ad of the Ni–Al₂O₃ interfaces doped with Si, Mn, and Al are negative, suggesting that these dopants would even prevent the formation of the interface. The main reasons for this phenomenon include the compatibility between elements in materials and the activity of the atoms (energy and size, etc.). In a previous study by our group [19], it was also found that titanium can improve the tensile properties of the Fe–Al₂O₃ interface structure. In this way, Ti seems to be effective in improving the interface properties of metals and Al₂O₃, which requires further verification with similar metals–Al₂O₃ structures. Consequently, doping alloy elements in the Ni–Al₂O₃ interface can have an uncertain effect on the bonding of that interface, which may be enhanced or weakened.

The process of doping will cause changes in the diameter and energy of the atomic position, which will induce a relaxation behavior in the atoms to form a new interface structure. The heat of segregation (ΔG_seg) is the driving force that forms the stable compound, which can be used to evaluate the stability of dopants in the interface structure. The calculation of ΔG_seg is performed utilizing the following Equation (2) [20]:

$$\mathrm{\Delta }{G}_{seg}={\displaystyle \frac{1}{n}}\left(E_{Ni/A{l}_2{O}_3}-E_{Ni/A{l}_2{O}_3.nx}+n{\mu }_x-n{\mu }_{Ni}\right)$$

(2)

where $E_{Ni/{Al}_2{O}_3}$ and $E_{Ni/{Al}_2{O}_3.nx}$ represent the total energies of the non-doped and doped-M Ni–Al₂O₃ interface structures, respectively. Here, n denotes the total number of atoms doped with x (where x can be Ti, Mg, Cu, Zn, Si, Mn, or Al), and μ signifies the chemical potential of the dopant element. If ΔG_seg is positive, it means that dopant atoms are easy to segregate from the matrix; if ΔG_seg is negative, it indicates that the doped atoms are a solid solution in the matrix and not prone to segregation.

Figure 3 shows the ΔG_seg of the doped-M and non-doped Ni–Al₂O₃ interface structures. It is obvious that the ΔG_seg of the doped-Ti, Mg, and Cu Ni–Al₂O₃ interface structures are larger than 0, which are 2.34, 1.42, and 0.86 eV, respectively, indicating that the dopants are in an unstable state in the matrix. However, the ΔG_seg of the doped-Zn, Si, Mn, and Al Ni–Al₂O₃ interfaces are negative and very close to 0, which means that the dopants are more prone to solid solutions in the matrix. From the perspective of the atomic structure, the Ti and Mg atoms have a large relaxation behavior, and the remaining dopants change very little, which matches the result of the ΔG_seg. Indeed, a higher heat of segregation corresponds to a lower total energy in the atomic structure’s system, resulting in a more stable structure. Taken together, Ti, Mg, and Cu dopants are beneficial to the stability of the Ni–Al₂O₃ interface, and hence they are selected for further studies.

2.2. The Electronic Properties of the Interface

The formation of heterogeneous materials is complex, especially with chemical and physical changes at the interface region, which will induce a transfer and redistribution of electrons. To better understand the strengthening mechanism of the Ni–Al₂O₃ interface by doping elements, it is necessary to study the electronic properties of the interfacial region. Next, we will discuss the spatial distribution of the electrons, the charge density of the plane, the electron overlap population, and the density of states for the Ni–Al₂O₃ interface structures.

Figure 4 shows the spatial electron distribution of the non-doped and doped-M (M = Ti, Mg, and Cu) Ni–Al₂O₃ interface structures. In general, the electron distribution positions mean the interactions between the atoms. Compared to the non-doped Ni–Al₂O₃ interface, Mg and Ti dopants relax towards the center of the interface, transferring their surrounding electrons onto the interface. Interestingly, it is the segregation of the dopant that allows the electrons to be continuously distributed to connect the metal and the ceramic (see Figure 4c,d), which is a microscopic manifestation of the strengthening effect. It is worth noting that the number of electrons at the doped-Ti interface is more than that of other interfaces, indicating that Ti has the most significant effect on the Ni–Al₂O₃ interface. This phenomenon matches the results of the W_ad above. Moreover, we found that Cu had no obvious effects on the electronic environment of the Ni–Al₂O₃ interface. Consequently, effective dopants can modify the interfacial electronic environment, thereby achieving the purpose of the strengthening effect. To put it differently, the properties of the interface are dictated by the electronic environment.

In order to better understand the characteristics of the Ni–Al₂O₃ interface, the charge density of the plane was analyzed, as shown in Figure 5. The plane with the greatest number of atoms is presented in this work, where red represents the charge aggregation region and blue represents the charge depletion region.

The distribution of charge density on the Ni-side and Al₂O₃-side are regular, indicating that the doping behavior of the interfacial region did not affect the non-interfacial region. The introduction of dopants (Ti, Mg, or Cu) at the Ni–Al₂O₃ interface widens the charge distribution region relative to the non-doped interface (illustrated by the yellow arrow in the Figure 5), marking a stronger bonding interaction from the interface. In particular, the Mg and Ti atoms had relaxed to the middle of the interface, effectively increasing the charge density of the interface. Herein, the segregation behavior of the dopants induces a lattice distortion at the interface, which may be beneficial for the bonding of the Ni–Al₂O₃ interface. Among all the doped Ni–Al₂O₃ interfaces, the reinforcement effect of Ti on interfacial bonding was the most significant, making it the best choice for practical applications. Therefore, further research on the strengthening mechanism of doped-Ti Ni–Al₂O₃ interface is necessary.

The overlap population of bonds provides a direct means to assess whether a valid chemical bond forms between atoms and unveils insights into the chemical changes occurring in the interfacial region. According to the traditional Mulliken formula (Equation (3)) [21], the overlap population of electrons for chemical bonds (Z_bond) can be derived by using Equation (4). Both of these equations are expressible as the following:

$$X_{Mulliken}={\displaystyle \frac{A+I}{2}}$$

(3)

$$Z_{bond}=a{e}^{-b(\mathrm{\Delta }{X}_{Mulliken}{)}^2}+c$$

(4)

Here, A represents the affinity energy of the atom, I represents the ionization energy of the atom, and a, b, and c are constants. The Z_bond represents the electron overlap population for the bonds, with a greater value indicating a greater strength of bond. A Z_bond value greater than 0 indicates that the two atoms are in a bonded state, while a Z_bond value less than 0 indicates that the two atoms are in a non-bonded state.

Figure 6 is the overlap population of bonds at the doped-Ti Ni–Al₂O₃ interface. The values of overlap populations between Ti and O, between Ti and Ni-1, and between Ti and Ni-2 at the interface were 0.46, 0.2, and 0.17 e, respectively. This demonstrates that Ti segregates to the interface and undergoes chemical reactions, forming new chemical bonds with the atoms at both sides of the interface. In fact, the formation of new chemical bonds progressively integrates the two interfacial sides into a unified entity. Therefore, it can be inferred that the Ti undergoes complex chemical reactions in the Ni–Al₂O₃ interface region and produces new substances that contribute to the interface’s stability, which is the strengthening mechanism of the Ni–Al₂O₃ interface. As a result, the dopant acts as an adhesive at the interface, which can be understand as a “glue effect”.

In order to reveal the bonding mechanism at the doped-Ti Ni–Al₂O₃ interface, we calculated the partial density of states (PDOSs) of the atoms. Figure 7 shows the PDOS of the doped-Ti Ni–Al₂O₃ interface (the positions of the atoms are provided in Figure 6). Relative to the Ni-interior atom, the electron p-orbits of the Ni-1 and Ni-2 at the interface exhibit new peaks ranging from −34.57 to −33.58 eV, matching the shape of the p-orbit of the Ti atom. It can be suggested that the orbital hybridization of electrons has occurred between the Ti and Ni-1, as well as between the Ti and Ni-2, which are a characteristic feature of covalent bonds. Furthermore, the peak of the s-orbit of O-1 atom at the interface becomes stronger within the range of −21.34 to −18.25 eV. The s, p, and d-orbits of the Ti atoms exhibited peaks that match the shape of the s-orbit of the O-1 atom. It can be concluded that electrons are transferred from Ti to O-1, and orbital hybridization of electrons occurs between the Ti and O-1, resulting in the formation of a Ti–O ionic compound. Consequently, the Ti–O1, Ti–Ni1, and Ti–Ni2 bonds exhibit typical bonding peaks (see the green and yellow boxes in Figure 7). This bonding behavior promotes and accelerates the formation of a stable Ni–Al₂O₃ interface.

Here, our conclusions will provide theoretical guidance for practical applications, advancing the innovation of Ni–Al₂O₃ composites. In future work, our group will consider the effect of the defects of the materials on the Ni–Al₂O₃ interface and the strengthening effect of the dopants on the interface under the defects.

3. Calculation Method and Details

3.1. Calculation Parameter

For this research, every calculation, from bulks to surfaces and interfaces, are executed with the Cambridge Sequential Total Energy Package (CASTEP) code [22], which is based on density functional theory (DFT). Initially, the geometry optimization of unit cells, surfaces, and interfaces are carried out using the Broyden–Fletcher–Goldfarb–Shanno (BFGS) minimization algorithm [23]. Convergence tests confirmed that a plane-wave cutoff energy of 380 eV is sufficient. The Brillouin zone was sampled using an 8 × 8 × 1 k-point grid for both interface and surface calculations. To ensure the accuracy of the calculations, the optimization convergence criteria for the original crystal cell structure was set as the following: an energy deviation per atom of 0.01 meV, a stress deviation of 0.05 GPa, and a maximum displacement of 0.001 Å. Subsequently, the ultrasoft pseudopotential (USPP) [24] was used to describe the interaction between the ionic nucleus and valence electrons. The valence electrons of the three atoms are set as O-2s²2p⁴, Al-3s²3p¹, and Ni-3d⁸4s², respectively. Ultimately, the generalized gradient approximation (GGA) with the Perdew–Burke–Enzerh (PBE) [25] method served as the electron exchange and correlation function in this work.

3.2. Model Building

According to the actual situation of material, the γ-Ni and α-Al₂O₃ cells serve as the basic units of the structure. The lattice constants for the Ni unit cell are a = b = c = 3.528 Å, and Al₂O₃ has a = b = 4.815 Å and c = 13.135 Å. Typically, stable heterogeneous interface systems are composed of crystal planes with the least energy of surface. It has been shown that γ-Ni (111) [17] and α-Al₂O₃ (0001) [26] belong to the surface structures with the lowest surface energy in the low-index crystal surfaces. Therefore, the α-Al₂O₃ (0001)/γ-Ni (111) interface system is established, which consists of 5 layers of Ni and 12 layers of Al₂O₃, as depicted in Figure 8.

Furthermore, the outermost atomic layer on the surface of the material has the greatest influence on the bonding properties of the interface, so the central position of the interface layer was chosen as the doping site to achieve effective contrast, as shown in Figure 9.

4. Conclusions

In this work, we conducted a systematic investigation into the bonding strength and strengthening mechanism of Ni–Al₂O₃ interfaces with doped-M (M = Ti, Mg, Cu, Zn, Si, Mn, or Al) at the atomic and electronic scales, yielding the following conclusions:

(1) The doping of Ti and Mg can increase the work of adhesion of the Ni–Al₂O₃ interface by up to factors of 3.38 and 1.52, respectively, while the dopant Cu almost does not change the work of adhesion of the interface. Conversely, the dopants of Zn, Si, Mn, or Al can not only fail to improve the work of adhesion, but also weaken the bonding strength of the interface. Accordingly, different dopants may exert either positive or negative effects on interfacial bonding.

(2) The dopant Ti and Mg atoms will precipitate from the matrix Ni and diffuse into the middle of the Ni–Al₂O₃ interface in the relaxation process. More importantly, the diffusion behavior of the Ti and Mg can induce the transfer and redistribution of electrons in the Ni–Al₂O₃ interface region. Compared to the non-doped Ni–Al₂O₃ interface, the doping of Ti and Mg increase the number of electrons and broaden the area of charge distribution in the interface region. As a result, effective dopants can optimize the electronic structure and enhance interatomic bonding strength at the interface.

(3) Of all the doped Ni–Al₂O₃ interfaces, Ti is the best candidate. It is found that the stronger bonds of Ti–Ni and Ti–O are formed at the interface, which significantly improved the stability of the Ni–Al₂O₃ interface. In other words, the dopant Ti appears to act as an adhesive at the Ni–Al₂O₃ interface.