1 Introduction

Hydroxylapatite (HAp) is a widely accepted material that has been used for preparing coatings on orthopedic implants since 1980 [1–4], due to its excellent biocompatibility and bioactivity, properties that make it well suited for biological fixation. Titanium alloy (Ti-alloy) has become one of the most frequently used biomaterials due to its excellent corrosion resistance, good mechanical properties and low toxicity [5, 6]. Combining these two materials to produce a HAp-coated titanium alloy thus becomes a promising way to obtain implants for orthopedic applications, because they are assumed to increase bone repair and remodeling compared with non-coated titanium alloy [7].

Methods such as plasma-spraying [3, 8], sintering and laser melting [9] have been developed to coat HAp on commercially pure titanium (cp-Ti) or titanium alloy. These methods are successful and useful ways to obtain high quality HAp coated titanium alloy. However, the high preparation temperatures necessary for these methods can lead to some negative effects, and the resulting HAp coatings with low crystallinity [10, 11] and poor mechanical stability of the interface between HAp coating and metal substrate in vivo [12]. Many methods have been developed to improve properties, sush as heat treatment, chemical treatment [13, 14] or spark plasma sintering [15]. Despite all these, there are still some problems.

Therefore, it is necessary to develop a new technique that can fabricate a bioactive coating on a Ti-alloy substrate with relatively high crystallinity and excellent mechanical strength at lower process temperature. As a bonding method, the eutectic melting bonding (EMB) method may be one probable solution. As far as we know, it has not been reported using the EMB method preparing coatings on the surfaces of metal substrate. The EMB method has now become a basic process in some cutting-edge techniques (such as active brazing or transient liquid phase diffusion bonding) used for welding different metals or metals and ceramics. An important characteristic of the EMB method is contact melting at a temperature higher than the eutectic point in the phase diagram i.e., if the two metals have contact with each other at a temperature higher than the eutectic point, inter-diffusion between the two metals will occur and result in eutectic melting. The eutectic liquid phase will form and grow around the interface until one of the metals is completely consumed and enters the liquid phase or until the temperature is lowered to that below the eutectic point. Obviously, this melting process is not driven by increasing temperature, but by constitutional superheating. Thus, the temperature of the EMB method is usually significantly lower than either of the melting points of the bonded metals. Therefore, if a eutectic reaction can occur between two different metals, the bonding temperature will be far lower than the melting points of the bonded metals, and a low bonding temperature can be successfully achieved and a high bonding strength can also be expected because the interface is metallurgically bonded.

Ti is low cytotoxicity metals [16], and Ti has a eutectic point with Ni at 942 °C with 24 at.% Ni and 76 at.% Ti (as seen in the phase diagram in Fig. 1 [17]). These characteristics make them suitable for use as components for preparing the HAp coatings. We hypothesize that if a HAp powder is placed above a Ni-foil, pressed on a Ti-alloy substrate and heated to a temperature above the eutectic temperature, the eutectic liquid phase could infiltrate through the HAp powder and create a coating that contains Ti, Ni and HAp after cooling to room temperature. The components in the coating would be distributed in a gradient: from Ti-alloy to the eutectic phase of Ti and Ni and finally to HAp. Therefore, the coefficients of thermal expansion of the transition layer would gradually vary between those of the HAp and Ti-alloy. Hence, a firm bonding between the substrate and the HAp could be achieved.

Fig. 1
figure 1

Ti–Ni phase diagram [17]. Points 1 and 4 indicate the solid phase composition of 49 and 9 at.% Ni at 1,050 °C, respectively. Points 2 and 3 indicate the liquid phase composition of 34 and 22 at.% Ni at 1,050 °C, respectively. Point A is the eutectic point (24 at.% Ni, 76 at.% Ti). Point B is a middle phase whose composition corresponds to the eutectic point A at 1,050 °C

Full size image

Based on the above theoretical speculation, we conducted a series of experiments to prepare HAp coatings on a Ti-alloy using the EMB method firstly. This is the first time to use EMB method to prepare HAp coatings on the surface of Ti-alloy, even the first report to prepare coatings using EMB method as far as we know. The phase components, microstructure, bonding strength and cytotoxicity of these coatings were characterized, and the potential applications were also anticipated.

2 Materials and methods

2.1 Materials

Ti-alloy (Ti–6Al–4V, Northwest Institute for non-ferrous metals Research) samples with dimensions of 3 × 54 × 54 mm3 with a shallow indent (1 × 50 × 50 mm3) at one end were used as the substrates. The surface of the indent was polished using emery paper (Hubei Fengpu Abrasive Co., LTD, Tongcheng, Hubei, China) with a #600 grit and then a #1,500 grit to remove the oxide. After polishing, the samples were ultrasonically cleaned in acetone (Tianjin Kemiou Chemical Reagent Co., Ltd., Tianjin, China) for 20 min and dried in an oven (Shanghai Jing Hong Laboratory Instrument Co., Ltd., Shanghai, China) at 50 °C for 20 min. The samples were then immersed in an oxidative acid solution made by mixing equal volumes of 18 % HCl (Shanghai Jing Hong Laboratory Instrument Co., Ltd., Shanghai, China) and 48 % H2SO4 (Shanghai Jing Hong Laboratory Instrument Co., Ltd., Shanghai, China) at 50 °C for 30 min to obtain roughened surfaces.

Ni-foils (Northwest Institute for non-ferrous metals Research, Purity: 99.5 %) with dimensions of 0.1 × 50 × 50 mm3 were used as one of the precursor components for the Ti–Ni eutectic reaction. Both sides of the foils were polished gently using emery paper with a #1,500 grit to remove the oxide, and the foils were then ultrasonically cleaned in ethanol for 10 min and finally kept in analytical ethanol (Shanghai Jing Hong Laboratory Instrument Co., Ltd., Shanghai, China) for further use.

The starting HAp powder was made using a wet chemical method [18]. The production of the powder was based on the following chemical reaction:

$$ 6 {\text{H}}_{ 3} {\text{PO}}_{ 4} + { 1}0{\text{Ca}}\left( {\text{OH}} \right)_{ 2} \to {\text{Ca}}_{ 10} \left( {{\text{PO}}_{ 4} } \right)_{ 6} \left( {\text{OH}} \right)_{ 2} + {\text{ 18H}}_{ 2} {\text{O}} $$
(1)

Heat-treatment of the powder in a muffle furnace (Tianjin Taisite Instrument Co., LTD, Tianjin, China) at 700 °C for 1 h was performed to achieve a complete crystalline structure. The average size of the powder generated using the previously described wet chemical method was 37.23 nm, according to laser particle analyzer (Malvern, Zetasizer Nano ZS).

2.2 Preparation of HAp/Ti2Ni coatings

A vacuum hot press furnace (FJK-2) equipped with a radiation heating system was utilized for preparation of the HAp/Ti2Ni coatings in this study. The FJK-2 was developed [19] by Northwestern Polytechnical University, and its highest temperature is 1,350 °C (control precision ±3 °C). The working vacuum during heating can reach up to 4.3 × 10−4 Pa. The furnace can simultaneously press the sample during the heating and vacuuming processes. The loading capacity is 100 N–100 kN.

The preparation of HAp/Ti2Ni coatings comprises two stages (shown in Fig. 2a): solid diffusion bonding of the Ni-foil and the Ti–6Al–4V substrate at 900 °C and 4 MPa for 60 min (Fig. 2a, stage I) and EMB of the HAp powder and the Ti–6Al–4V substrate bonded with Ni-foil at 1,050 °C for 40 min (Fig. 2a, stage II).

Fig. 2
figure 2

Schematic illustrations of the preparation of HAp/Ti2Ni coatings and the tensile test specimen. a Illustration of the preparation of HAp/Ti2Ni coatings. Ni-foil was put in the indent of the Ti-6Al-4V substrate, and then a pressure of 4 MPa was applied to the Ni-foil and the substrate at 900 °C to bond them together (Stage I). Then, HAp powder was evenly spread onto the surface of the bonded Ni-foil and heated up to 1,050 °C and the temperature was maintained for 30 min to obtain the final coating (Stage II). b Illustration of the geometrical arrangement for the bonding of the Ni-foil and the Ti-alloy substrate by solid diffusion bonding. c Illustration of the preparation of the sandwich tensile test specimens for the ASTM C633 adhesion test

Full size image

2.2.1 Solid diffusion bonding of Ni-foil and Ti-alloy substrate

Figure 2a, stage I and Fig. 2b schematically show the process of bonding the Ni-foil with the Ti-alloy substrate. First, a Ni-foil was placed into the indent of the Ti-alloy substrate (Fig. 2a), and then 7 g of well-dispersed nano-HAp powder was spread evenly on the Ni-foil in the indent of the Ti–6Al–4V substrate (Fig. 2a, stage II). Second, the sample (Ti-alloy substrate with the Ni-foil and the HAp powder) was placed between two Al2O3 welding-barrier plates, which are used to isolate the sample to avoid welding with the pressing head or the pedestal. The thickness of the plates was 0.5 mm. The sample in the welding-barrier plates was placed on the pedestal in the vacuum hot press furnace (FJK-2) (Fig. 2b). Finally, the pressing head located right above the sample applied an axial pressure of 4 MPa to the sample at 900 °C for 60 min in a vacuum of 1.4 × 10−3 Pa. Because the temperature is lower than the eutectic point, eutectic melting would not occur, but solid diffusion bonding would occur and a Ti-alloy substrate with a joined Ni-foil was obtained.

2.2.2 Eutectic melting bonding of the coatings and the Ti–6Al–4V substrate

After the solid diffusion bonding in stage I, the load (4 MPa) was relieved and the temperature was increased from 900 to 1,050 °C and kept constant for 40 min under a vacuum (1.4 × 10−3 Pa). The Ti-Ni phase diagram (Fig. 1) shows that the region around the Ti/Ni interface will melt at 1,050 °C due to the Ti/Ni eutectic reaction. Thus, during this EMB process, the Ti/Ni eutectic liquid would form until the Ni-foil is completely consumed. The liquid phase infiltrates into the HAp powder layer and the layer would be joined to the Ti-alloy substrate after cooling down to room temperature.

2.3 Preparation of tensile specimens

The bonding strength between a coating and a substrate is generally estimated by detaching the coating from the substrate in a tensile test. The tensile specimen is typically made of two as-coated samples that are bonded together using an organic adhesive. However, because the bonding strength between the coating and the substrate in the current case was higher than that of the organic adhesive, it was not possible to detach the coating from the substrate using the organic adhesive. To obtain the bonding strength, we made special sandwich tensile specimens by uniting two Ti-alloy rods (Φ25 mm × 40 mm) between which the HAp/Ti2Ni coating was prepared by the EMB method using processes similar to those described above. The processes are as follows: first, diffusion bonding of the Ni-foil with the Ti–6Al–4V rod. Two rods (one with an indent and the other with a convex head illustrated in Fig. 2c) were aligned head to head coaxially. Ni-foils were attached to the rod head and a layer of 0.8 g HAp powder was evenly spread between the Ni-foils. The parameters for bonding the Ni-foils and the Ti-alloy rods were the same as those described in Sect. 2.2.1. Second, the two rods were bonded using the EMB process. After the first step, the 4 MPa load was relieved and the sample was heated to 1,050 °C to carry out the EMB process. The temperature was maintained at 1,050 °C for 40 min to obtain the specimens. The size of the tensile specimens were made to meet the requirements of the standard tensile adhesion test (ASTM C633 [20]) that was specially designed for plasma-sprayed coatings. In the subsequent test, these specimens were subjected to tensile force at a constant crosshead speed of 0.05 mm/min until failure using a tensile test machine (Instron 3382, America).

2.4 Cell culture

A cell culture test was performed on the coatings. Prior to the cell culture test, samples were cut into 8 × 8 mm2 square shapes, and ultrasonically cleaned in ethanol and rinsed with distilled water. All samples were air dried and sterilized in sealed plastic bags in an autoclave (ZHN-3, China) at 180 °C for 2 h.

The cell culture test was carried out using the MG-63 cell line, which was established at the School of Life Sciences, Northwestern Polytechnical University. The MG-63 cells were incubated on the autoclaved samples in a 24-well plate with culture medium (Dulbecco’s modified eagle medium, DMEM) containing 10 % fetal bovine serum (FBS) and 0.5 % antibiotics. The cultures were maintained in an atmosphere of 100 % humidity, 5 % CO2, and 37 °C in the 24-well plate. After incubation for 1, 4, or 6 days, the cells were fixed using 0.3 % glutaraldehyde, and then dehydrated in acetonitrile and critical-point dried.

2.5 Characterization

Due to the vital importance of understanding the correlation between the microstructure of the HAp coating and its properties such as morphology and bioactivity, the microstructure and the element distribution of the coating and the morphology of the cells adhered to the coating were characterized using a scanning electron microscope (SEM: Tescan VEGA, Czech). The phases of the coating were detected by X-ray Diffractometer (XRD) (Rigaku D/MAX-2400, Japan) operated at 45 kV and 40 mA with CuKα radiation (λ = 1.5406 Å). The coating was scraped with a knife carefully and finely milled in an agate mortar.

3 Results

3.1 Phase composition and crystallinity of the coatings

The XRD analysis of the coating is shown in Fig. 3. The main peaks matching those of the crystalline HAp showed that the HAp in the coating was of relative high crystallinity. From the peaks, it was determined that the main phase was hydroxyapatite, and the amount of this phase is 90 %. A small amount (totally 10 %) of Ca4O(PO4)2 and Ti2/Ni can also be detected. There were no other phases detected beyond the above components. However, the coating is not totally crystalline HAp, as it can be seen in Fig. 3 that an obvious background and a slight halo in the XRD pattern which shows a partial non-crystalline structure.

Fig. 3
figure 3

XRD pattern of the HAp/Ti2Ni coating. The major composition in the coatings was HAp. There was also a small amount of Ca4O(PO4)2 and Ti2Ni

Full size image

3.2 Element distribution and Ca/P ratio of the coating

By examining the cross-section of the coating using SEM and energy-dispersive X-ray spectroscopy (EDS), we obtained an SEM image and the distribution of the elements in the coatings. Figure 4 shows the cross-section of the coating (Fig. 4a) and its element distribution obtained by EDS (Fig. 4b). The microstructure shows that the coating consisted of scattered darker-colored particles embedded in a lighter-colored matrix. The size and number of the particles gradually decreased from the surface to the substrate. The EDS spectrum shown in Fig. 4b indicates that the major elements in the coating were Ca (atomic ratio: 21.32 %), P (atomic ratio: 10.75 %), and O (atomic ratio: 67.93 %), while those in the matrix were Ti and some Ni. From the surface to the inner part of the coating, the amounts of the elements Ca, P, and O decreased sharply, whereas the amount of Ti increased. The thickness of the coating was estimated to be approximately 300–400 μm.

Fig. 4
figure 4

SEM image of the coatings cross-section. a Single arrows show the coating/substrate interface and the double-headed arrow indicates the thickness (approximately 305 μm) of the coating. b Element distribution on the coating cross-section

Full size image

It should be noted that the Ca/P ratios of the coatings produced in this study were approximately 1.98, which was higher than that of pure HAp (1.67). However, this result was not surprising because some phases other than pure HAp appeared in the coatings. For instance, the XRD patterns illustrated in Fig. 3 showed the presence of non-HA phase Ca4O(PO4)2, the Ca/P ratio of which is 2. Based on these results demonstrated in Figs. 3 and 4, we can conclude that the coatings were mainly composed of HAp particles and a Ti2Ni matrix.

3.3 Surface morphology of the coating

The surface morphology of a coating is shown in Fig. 5. It was observed that the surface of the coating was coarse and uneven with both concave and convex shapes, as illustrated in Fig. 5a. Many irregular shaped particles can also be observed on the surface, as shown in Fig. 5b. The size of the particles was typically in the range of 0.5–2 μm. The observed faceted morphology is a characteristic of ceramic phases (such as HAp) rather than alloy phases. Thus, it is reasonable to deduce that there is more HAp presented on the surface of the coating. This deduction is consistent with the results shown in Figs. 3 and 4b. As for biocompatible implants, the coarse and uneven morphology on the coating surface would be very helpful for the in-growth of osseous tissue.

Fig. 5
figure 5

SEM images of the coating surface. a A low magnification image showed a typical solidification metal morphology. b A high magnification image showed a rough and porous surface morphology

Full size image

3.4 Bonding strength

The bonding strength between the coating and the substrate was evidently high because the usual method failed to determine the bonding strength. In the usual method, two coated samples are usually bonded together using an organic adhesive to obtain the tensile test specimen for determination of the bonding strength. In a standard tensile test, when the bonding strength between the coating and the organic adhesive is higher than that between the coating and the substrate, the tensile strength can be obtained by detaching the coating from the substrate. However, in our experiment, the specimens were always broken at the interface between the coating and adhesive. Because the usual method could not be applied in the current situation, the new tensile test specimens as described in Sect. 2.2 were prepared and the tensile test was carried out using the standard tensile adhesion test method (ASTM C633). The results show that the bonding strength between the coating and the substrate was 69.68 ± 5.15 MPa (n = 3), which was indeed very high.

3.5 Cell morphology

Figure 6 shows typical SEM images of the MG-63 cells cultured on the HAp/Ti2Ni coatings for 1, 4, and 6 days. As illustrated in Fig. 6a–f, the cells were typically flat and attached tightly to the surface of the coatings. The cell also underwent shape changes, producing finger-like or sheet-like protrusions (filopodia or lamellipodia, respectively), suggesting that the coatings had good cell viability. A few differences were observed in the morphology of cells cultured on the surface of coatings for 1, 4, or 6 days.

Fig. 6
figure 6

SEM images of MG-63 cells on the surface of the coatings at 1 day (a, b). Cells attached to the surface of the sample and stretched out many filopodia. MG-63 morphology at 4 days (c, d). The number of cells was increased, cells extended lamellipodia and cell–cell contacts occurred. MG-63 morphology at 6 days (e, f). The cells spread out and covered most of the area of the sample surface, and cell–cell contacts also covered a large area

Full size image

After incubation for 1 day, the spindle-shaped cells attached, spread, and grew on the surface of the HAp/Ti2Ni coatings, with filopodia extending from the cell edges (Fig. 6a). Cell–cell contact via the filopodia was also observed (Fig. 6b).

After incubation for 4 days, the number of cells increased (Fig. 6c) compared with a 1-day incubation, though the morphology and size did not change remarkably. In addition, part of the cell filopodia was developing into lamellipodia, and some cells were joined together as shown in Fig. 6d. The cells had a greater degree of contact, which could enhance gap junctional communications between cells.

After incubation for 6 days, cells covered almost 80 % of the coating surface (Fig. 6e) and were observed adhering and spreading on top of each other via developed lamellipodia, forming a multilayer of cells without any specific orientation (Fig. 6e, f). Cells grew into and spread well on the concave of the coatings (Fig. 6e). The cell morphology changed from spindle-shaped to slice-shaped. Cell–cell contact was increased and few areas with no cells were observed (Fig. 6f).

4 Discussion

The EMB method proposed in this paper is a new effort to firmly bind a HAp coating to a Ti-alloy substrate. To better understand the preparation processes and the characteristics of the coatings, the processes of the preparation method and the coatings’ phase composition were analyzed from a thermodynamic viewpoint. The benefits of the preparation method will also be discussed in this section.

4.1 Thermodynamic analysis of the coating preparation

The preparation process used for either the HAp coating samples (Sect. 2.2) or the tensile specimens (Sect. 2.3) consists of two stages: the solid diffusion bonding at 900 °C and the EMB method at 1,050 °C. During the first stage, although Ni and Ti-alloy were kept in solid states, it has been confirmed by many studies that at approximately 900 °C three intermetallic layers of Ti2Ni, TiNi, and TiNi3 would be produced and arranged from the Ti-alloy side to Ni-foil side [21–24], as shown in Fig. 7. Moreover, the thickness of the layers can also be estimated by the following equation [25]:

$$ x = \left( {k_{0} \exp \left( {{{ - Q} \mathord{\left/ {\vphantom {{ - Q} {{\text{R}}T}}} \right. \kern-0pt} {{\text{R}}T}}} \right)t} \right)^{{{1 \mathord{\left/ {\vphantom {1 2}} \right. \kern-0pt} 2}}} $$
(2)

where x is the layer thickness (m), k 0 is the growth constant (m2/s), Q is the activation energy for layer growth (J/mol), R is the universal gas constant (8.314 J/mol/K), and T and t are the temperature (K) and time (s) of diffusion bonding, respectively. It has been previously reported [21] that the thicknesses of the Ti2Ni, TiNi, and TiNi3 layers are 3.27, 5.2 and 5.9 μm, respectively when heated at 900 °C for 45 min. Thus, using Eq. 2, it can be easily calculated that the thicknesses of the Ti2Ni, TiNi, and TiNi3 layers at the end of the first stage in this study (900 °C for 60 min) were 3.8, 6 and 6.8 μm, respectively.

Fig. 7
figure 7

Schematic of the preparation process for the Ti2Ni/HAp coatings, which consisted of three stages: first, a, b, solid diffusion bonding at 900 °C for 60 min; second, c–e, the EMB process at 1,050 °C for 30 min; third, f, the cooling down period

Full size image

When the samples were heated to 1,050 °C in the second stage (the EMB method), the 3.8 μm thick Ti2Ni layer would melt first due to its lower melting point (984 °C, see Fig. 1). Then, the liquid layer would grow thicker by the eutectic melting of βTi and TiNi if their Ni contents were in the range of 9–49 at.% (from point 4–1 in Fig. 1). The Ni content of the produced TiNi eutectic liquid was from 22 to 34 at.% (from point 3–2 in Fig. 1). The growth of the eutectic liquid would continue until the Ni-foil was completely exhausted. The growth rate of the liquid layer (or the melting rate of the Ni-foil) would be very high because it was dominated by the fast inter-diffusion of Ti–Ni in liquid. This deduction was confirmed by our results. The 420 μm thick liquid layers were observed at the Ti–6Al–4V/Ni interface that had been heated at 950 °C for 45 min [22]. Based on Eq. 2, we calculated that the 100 μm thick Ni-foil layer used in this study would be exhausted in 2.6 min even at 950 °C, so the eutectic melting of the Ni-foil would be quicker during the 1,050 °C EMB process in this study.

When the Ni-foil completely melted, the TiNi eutectic liquid would come into contact with and infiltrate into the HAp layer. Hydrogen and oxygen were produced by the decomposition of HAp at the EMB temperature (1,050 °C) as shown in Eq. 3:

$$ {\text{Ca}}_{ 10} \left( {{\text{PO}}_{ 4} } \right)_{ 6} \left( {\text{OH}} \right)_{ 2} \to {\text{Ca}}_{ 4} {\text{O}}\left( {{\text{PO}}_{ 4} } \right)_{ 2} + 2 {\text{Ca}}_{ 3} \left( {{\text{PO}}_{ 4} } \right)_{ 2} + {\text{H}}_{ 2} {\text{O}} $$
(3)

As a result, the strong chemical affinities between Ti/H [26] and Ti/O [27] would result in a good wetting between the Ti/Ni liquid and HAp. Consequently, a complete infiltration could be observed (Fig. 4).

The as-prepared coating thickness basically arises from extended HAp powder and Ti2Ni eutectic mixture, the latter derived from infiltration and solidification of the Ti/Ni eutectic liquid. The eutectic liquid infiltrated into and wrapped HAp powder groups, and then solidified with cooling down. The thickness of spread HAp powder is about 100 μm in a loose station of as-spread. If the composition of the Ti/Ni liquid is the same as that of the eutectic point A (24 at.% Ni, Fig. 1) and the few volume change of Ti/Ni mixture from liquid to solid is ignored, the thickness of Ti/Ni liquid phase can be seen as the one of Ti/Ni solid phase. The thickness of the liquid at the end of the Ni-foil melting, W, mainly depends on the thickness of Ni-foil, can be expressed as follows:

$$ W = \frac{L}{{0.24v_{Ni} }}\left( {0.24v_{Ni} + 0.76v_{Ti} } \right) $$
(4)

where L is the initial thickness of the Ni-foil (100 μm) and v Ni and v Ti are the molar volumes of Ti (10.54 × 10−6 m3/mol) and Ni (6.59 × 10−6 m3/mol), respectively.

According to Eq. 4, it can be estimated that W should be approximately 665 μm, then the total thickness of coating should be approximately 765 μm. However, the coating thickness (approximately 300–400 μm) was reasonably thinner than the estimated value because during the EMB process, the pressing head (see Fig. 2b) would squeeze out some Ti/Ni liquid due to the self-weight of the head.

After the Ni-foil was exhausted, the concentration of Ti in the liquid would increase (from point B to point 3 in Fig. 1) due to the diffusion of Ti flux driven by the concentration gradient. Then, the liquid composition would reach point 3, and an isothermal solidification would begin [28, 29]. The time of the isothermal solidification, t s , would be as follows [30]:

$$ t_{s} = \frac{{2\pi W^{2} }}{{16DC_{\hbox{max} }^{2} }} $$
(5)

where C max is the maximum concentration of the solute in the liquid, i.e. Ti in the Ti/Ni liquid (wt%) and D is the diffusion coefficient of the solute in the residual solid, i.e. that of the Ti in Ti-alloy boundary layer near the liquid (m2/s), which can be described by an Arrhenius function as follows:

$$ D = D_{0} \exp \left( {\frac{ - G}{RT}} \right) $$
(6)

where D 0 is the pre-exponential factor and G is the diffusion activation energy.

In the case of the Ti–Ni solid phase, D 0 and G are 1.8 × 10−8 (m2/s) and 155 × 103 (J/mol), respectively [31]. When T = 1,050 °C, D is 1.36 × 10−14 (m2/s). Using the terms C max  = 74 wt% (as seen as the top X-axis in Fig. 1: 74 wt% of Ti) and W = 6.65 × 10−4 m, the isothermal solidification time can be estimated as 6,474 h. Even if the actual coating was thin (300–400 μm), and the liquid thickness was assumed to be half of the value W, the isothermal solidification time t s will be 329–586 h. Therefore, we determined that, in this study, the solidification of the Ti/Ni liquid was driven by the temperature decrease in the cooling process rather than the Ti diffusion in the residual Ti-alloy.

4.2 Phase composition analysis

As previously reported [21–24, 32] and can be deduced from the phase map (Fig. 1), three intermetallic layers (Ti2Ni, TiNi, and TiNi3) should have been produced at the beginning of the EMB process (see Fig. 7b). Furthermore, the thermodynamic conditions are in favor of the residual of TiNi because the Gibbs formation free energies for Ti2Ni, TiNi, and TiNi3 are −26.3, −34.5, and −30.6 kJ/mol, respectively. However, in our experiment, only Ti2Ni was observed in the final coating (see Fig. 3). The reason for this inconsistency between the actual intermetallic product and the product predicted by thermodynamics is that the amount of Ni used in this study was limited. If the amount of Ni is limited, the type of residual intermetallics should be determined by the effective Gibbs formation free energy [33, 34], which can be evaluated by the following equation:

$$ \varDelta G_{e}^{i} = \varDelta G\frac{{c_{e} }}{{q^{i} }} $$
(7)

where ΔG i e is the effective Gibbs formation free energy of phase i, c e is the concentration of the limited element in the reactant (c e  = 1 in this study), and q i is the molar ratio of the limited element in phase i (q i is 33, 50, and 75 % for Ti2Ni, TiNi, and TiNi3, respectively). From Eq. 7, ΔG i e for Ti2Ni, TiNi, and TiNi3 can be estimated to be −79.7, −69, and −40.8 kJ/mol, respectively. Thus, it can be seen that a residue of Ti2Ni in the final coating is a reasonable thermodynamic result.

4.3 Advantages of the EMB preparation method

A relative high degree of crystallinity, chemical purity, and phase stability [35, 36] is necessary for increasing the bone adaptation of HAp coatings. The EMB method proposed in this study kept the coating preparation temperature at 1,050°C, which is much lower than the melting point of either Ti (1,670 °C) or Ni (1,455 °C). This temperature is not only near the lower end of the temperature range (1,000–1,360 °C) in which HAp releases its OH− ions and transforms into OHAp but also is far below the threshold temperature (1,360 °C) above which the OHAp decomposes into TCP and TTCP [37]. Thus, the surface layer of the coatings in this study contained 90 % crystalline HAp (Fig. 3) and a small amount of Ti2Ni and Ca4O(PO4)2. An appropriate degree of crystallinity of the HAp coatings is necessary to influent the coating degradation rate [37]. Amorphous HAp coatings with tend to dissolve rapidly in the physiological environment, so that the coatings with low crystallinity would become weak quickly and may cause inflammatory responses, whereas the coatings with relative high crystallinity can maintain their stability in the complex physiological environment for a longer time. Such coatings exhibited good cell adaptation in vitro (see Fig. 6). However, the nucleation of bone-like apatite may not be preferential to the higher crystalline HAp surface in the physiological environment [38]. As it can be seen in Fig. 3, there is a mild halo in the XRD pattern which shows a partial amorphous structure. As a result, as-prepared HAp/Ti2Ni coating has an appropriate crystallinity which leads no quick degradation and keeps ability to forming bone-like apatite in vivo.

The surface morphology of the coatings prepared by the EMB method was also good for improving the biocompatibility. Cells are sensitive to topographical features ranging from mesoscale to nanoscale. Rough and porous coating surfaces are always required because living cells need a proper topography to adhere, proliferate, and grow once the surface is implanted in the body. Moreover, the rough and porous surface structures of HAp/Ti2Ni coatings can enlarge the micro-surface area of the coating and greatly increase the interface area between the coating and the fresh osseous tissue, which can provide good growth conditions for tissue and can even help to improve the osteal combination ability. The micropores (Fig. 5b) in the coatings provided attachment points for cell filopodia in the period of early adhesion. The early adhered cells would send signals via gap junctions [39] that the location is suitable to attach, which induced other cells to attach and adhere to the coatings. The quality of cell adhesion will influence their morphology and their capacity for proliferation and differentiation [39]. Moreover, the stretching out of the filopodia means the osteoconduction of the coating substrates was high. These results also indicated that HAp/Ti2Ni coatings served as good substrates for the attachment and adhesion of MG-63 cells, which is required for their subsequent differentiation.

Another important point that is necessary to address is the inherently high bonding strength between the substrate and the coating prepared by the EMB method. The high bonding strength of the coatings prepared by the EMB method originated mainly from the excellent bonding by the strong metallurgical interface. The low preparation temperature, another benefit for the high bonding strength, decreases the thermal mismatch between the coating and substrate and the residual thermal stress around the coating/substrate interface, which favors a high bonding strength. Therefore, the bonding strength between the substrate and the HAp coatings prepared by the EMB method could easily exceed 60 MPa, while the strength of the HAp coatings prepared by traditional methods (such as plasma spraying, pulsed-laser deposition, sol–gel, or electron-deposition) is usually not higher than 40 MPa [40].

5 Concluding remarks

In this paper, we introduced using the EMB method for preparing HAp/Ti2Ni coatings. The most prominent features of these coatings are that they have a very good bonding strength to the substrate (a bonding strength higher than 60 MPa can be easily achieved) and a relatively high degree of crystallinity, both of which are desired features for orthopedic implants. Furthermore, the surface morphology of the coatings was also suitable for cell attachment and growth, thus, the coatings may be a good candidate material in practical application. The EMB method of preparation may be potentially useful in the field of preparing biocompatible materials, not only in binding HAp coatings with a Ti-alloy but also in the preparation of other similar types of coatings on metal or alloy substrates.