**1. Introduction**

One of the many issues faced with regards to foreign objects being inserted into the body is the reduction or elimination of adverse e ffects caused by a stimulated immune response. Many such e ffects can be reduced with a careful material selection; however the current trend is to use the response of the body to improve the initial success of implant materials which could lead to a longer device lifetime, reducing the need for revision surgery. In the case of load bearing implants, such as in total hip replacement (THR) or total knee replacement (TKR), further complications such as the mechanical requirements of the material to bear load can reduce the number of materials available. It is often the case that materials that possess the mechanical properties are not always the most compatible in these sites. A possible solution is to coat mechanically sound devices with materials which will be

more biologically favourable [1]. A considerable amount of work has already gone into this area [2,3] but many problems such as selection of the most appropriate correct coating technique is required to control properties such as coating adhesion [4], composition [5] and topography [6], which may ultimately influence the biological response of the surface.

While many material systems offer attractive properties for biomedical applications, HA has attracted much attention due to its excellent bioactive ability for the application of bone repair and bone replacement [7]. Unfortunately, the nature of ceramic materials due to their reduced mechanical performance in comparison to metals, means that they, including HA, are restricted to the application of non-load bearing sites [8]. This problem can be overcome by applying an HA coating onto metallic materials with superior mechanical strength and stiffness [9]. Currently, plasma spraying is the only suitable commercial and FDA approved coating process employed [10,11], but coatings are compromised by the inclusion of undesired calcium phosphate phases leading to bioresorption and variable tissue response [12]. In addition, these coatings are often poorly bonded to the metallic substrates, leading to failure in the form of cracking at the coating/substrate interface [13]. Physical vapour deposition (PVD), as an attractive alternative, offers dense, defect free, well adhered single phase coatings with controllable deposition parameters [14,15].

Numerous authors have showed RF magnetron sputtering could be a potentially beneficial technique for applying coatings to future implantable devices [16–20]. Current work has shown that HA coatings are bioactive and exhibit similar properties to bulk materials. Many different compositions have been achieved for HA thin films due to differing sputtering parameters [16–18,21]. However, it is still unclear what suitable deposition parameters are and how factors such as power density and sputtering environment affect resultant films. One important compositional parameter is the Ca/P ratio, which can give rise to changes in the thermal stability, biological response, mechanical properties and dissolution potential [17].

SiHA thin films have been shown to elicit an enhanced cellular response compared to HA, which may provide a way to increase bone response in vitro and in vivo, thus improving the longevity of implants for future generations [22,23]. Work concerning RF magnetron sputtered SiHA thin films have only recently been investigated in the last decade [19,23–26]. Thin film systems often differ from bulk material systems due to the manufacturing process. Therefore, it is assumed that silicate groups substitute for phosphate groups and it is not fully understood what may happen at higher silicon concentrations above 5 wt.% [23]. Furthermore, optimal silicon values (ca. 2.2 wt.%) have been suggested but these are not in agreemen<sup>t</sup> [27,28]. Therefore, higher silicon additions [29] up to 13.4 wt.% and altered crystallinities have been investigated in this study.

### **2. Materials and Methods**

### *2.1. Substrate Preparation*

A commercially pure grade 1 titanium sheet (Timet, Swansea, UK) (CPTi) was wire eroded into 10 mm diameter, 1 mm thick discs, which were then ground to a mirror finish (Ra ~ 16 ± 4 nm) using a series of silicon carbide paper (P240-P1200; Struers, Rotherham, UK), polished with a mixture of colloidal silica (Buehler, Germany) and 10% hydrogen peroxide (Fisher Scientific, Loughborough, UK) and finally water. The discs were cleaned for 30 min in acetone, IMS and distilled water and dried under flowing nitrogen (BOC, London, UK).

Silicon single crystals (100) orientated Czochralski wafer (Compart Technology Ltd., Peterborough, UK) were cut into squares approximately 10 mm × 10 mm in size with a diamond tipped scribe, and were used as substrates for XRD to reduce peak contributions from the substrate.

### *2.2. Target Preparation*

The HA targets were manufactured by plasma spraying HA powder (Plasma Biotal, Buxton, UK) onto a circular copper backing plate 75 mm in diameter and 5 mm in thickness, with a target material

thickness of ca. 500 μm. For the Si doping, a single 99.999% pure silicon target (Kurt J. Lesker, Hastings, UK) was used, with dimensions 300 mm × 100 mm × 0.635 mm. Figure A2 demonstrates the target composition via XRD analysis to be mostly HA.

### *2.3. RF*/*Pulsed DC Magnetron Sputtering*

The HA and SiHA thin film coatings were produced using a TEER UDP-650 type 2 unbalanced magnetron sputtering rig (Teer Coatings Ltd., Droitwich, UK). Before deposition, the chamber was pumped using successive rotary (Edwards E2M40 rotary pump; 10−<sup>2</sup> Torr) and di ffusion (Edwards 750 di ffusion pump) pumping steps to a minimum of 2 × 10−<sup>5</sup> Torr. The chamber was then backfilled with argon (36 sccm; 2 × 10−<sup>3</sup> Torr; 99.999% purity; Pureshield BOC ©, London, UK).

The HA target was powered by an Advanced Energies RF power supply unit whilst the silicon target was powered by an Advanced Energy Inc. (Bicester, UK) DC Pinnacle pulsed DC power supply. The frequency and pulse time were maintained constant at 150 kHz and 1500 ns, respectively. A −30 V bias was applied to the substrates. The samples were mounted on a stainless steel plate with double sided adhesive tape and rotated at 4 RPM. The samples were bias-cleaned for 2 min prior to sputtering. The total deposition time for each coating type was maintained at 2 h, with approximate sputtering rates of ca. 100 nm/h.

### *2.4. Post Deposition Heat Treatments*

The obtained thin film coatings were recrystallised via heat treatment at a range of temperatures, with a set time of 2 h using a Lenton Thermal Design tube furnace (Hope Valley, UK) in an argon atmosphere. Argon was flowed at 100 sccm for 30 min before ramping the temperature up to 20 ◦C min−1. Mass spectrometry indicated that all gases and water vapour were at untraceable levels before ramping. The samples were then left to cool under flowing argon.

### *2.5. Materials Characterisation*

#### 2.5.1. Scanning Electron Microscopy (SEM) and Energy Dispersive X-Ray Spectroscopy (EDX)

A Phillips XL-30 scanning electron microscope (LaB6, Surrey, UK) was used to obtain micrographs at accelerating voltages between 10–20 keV and a working distance of 10 mm. An Oxford Instruments energy dispersive X-ray microanalysis (EDX) system (High Wycombe, UK) was utilised with a collecting time of 50 s. Ten randomly selected spots were analysed. Approximately 90,000 total spectrum counts were taken per sample area.

#### 2.5.2. Focused Ion Beam SEM (FIB-SEM) and Transmission Electron Microscopy (TEM)

The HA and SiHA thin films were initially sectioned using a FEI Quanta 200 3D FIB-SEM (Cambridge, UK) fitted with a Quorum cryo-transfer unit, an Omniprobe micromanipulator and an INCA Oxford Instruments EDX analysis system. The FIB-SEM was operated at an ion beam accelerating voltage of 30 kV and electron beam accelerating voltages of 5–20 kV. A coating of tungsten was deposited in situ by chemical vapour deposition (CVD) to protect the HA coating. FEI-Runscript software was employed to mill inspection trenches, using reducing milling currents of 7 to 1 nA for rough sectioning, followed by milling currents of 0.5 nA to 30 pA in order to polish the lamella surfaces and to fashion U-shaped cuts into the lamella to facilitate lift-out. A cross sectional microstructural and chemical analysis was performed using a JEOL 2000-FX-II TEM (Welwyn Garden City, UK) operating at 200 keV.

### 2.5.3. X-Ray Di ffraction (XRD)

A Bruker AXS D8 Advance X-ray di ffractometer (Coventry, UK) was used in glancing angle X-ray di ffraction mode. Cu k α X-rays (λ = 1.5406 Å) incidented the samples at a 2θ range of 20◦–55◦, a step size of 0.02◦ and a dwell time of 11 s. The samples were mounted and rotated at 30 RPM. A 2θ

range and dwell time of 25◦–35◦ and 2 s, respectively, was used for recrystallisation measurements. Temperature scans were performed from room temperature to a maximum of 800 ◦C to investigate the recrystallisation of thin films.

PC-APD version 3.5B DOS-based software was used to determine crystallite size of thin films using the Scherrer equation:

$$B = \frac{0.9\lambda}{t\cos\theta} \tag{1}$$

where *B* is the peak broadening (full width half maximum (FWHM) in radians), λ is the wavelength of the XRD source material and *t* is the crystallite diameter. B factors in instrumental broadening (*B* = *B0bs* − *Binst*) where *Bobs* is the observed line broadening, which includes instrumental factors such as detector slit width, area of the specimen irradiated and possible Kα<sup>2</sup> X-rays.

### 2.5.4. Reflective High-Energy Electron Diffraction

A RHEED unit coupled with a JEOL 2000fx TEM operated at 200 keV was used to assess the surface crystallinity of films. The samples were held perpendicular to normal on a stage positioned immediately below the objective lens. The samples were tilted so that the shadow edge was positioned close to the primary beam to access near surface diffraction information. A GaN/GaAs single crystal standard reference sample with known *d*-spacings and a camera constant of 33.8 ± 0.5 cm was used to calculate thin films' *d*-spacings.

### 2.5.5. X-Ray Photoelectron Spectroscopy (XPS)

A Kratos Instruments Axis Ultra (Manchester, UK) with a monochromated Al Kα X-ray source was run at 10 keV and 15 mA. The instrument was operated in constant analysis energy (CAE) mode with a pass energy of 20 eV for high-resolution scans. Chemical and compositional information was obtained between 0–1400 eV. All samples were charge corrected to the C 1s adventitious carbon peak, which was set to a value of 284.8 eV. The region areas were selected manually and peak deconvolution was carried out using Gaussian-Lorentzian (GL30) line shapes. The data analysis and compositional quantification were carried out using CasaXPS software (version 2.3.22).

### 2.5.6. Fourier Transform Infrared Spectroscopy (FTIR)

Fourier transform infrared spectroscopy (FTIR) was performed on the thin films using a Bruker Optics Tensor 27 spectrometer (Coventry, UK) in glancing angle mode at 80◦ with a liquid nitrogen cooled MCT detector. The chamber was flushed with compressed air continually at 500 sccm to reduce background signal from water vapour. Samples scans were taken in triplicate for each coating with a total of 90 scans per specimen. Resultant spectra were analysed using an OPUS software (version 7.0).

### 2.5.7. Surface Roughness – Optical Profilometry

A Mitutoyo Surftest SV-600 profilometer (Coventry, UK) equipped with Surfpak SV software was used to measure surface roughness of the coatings. The stylus had a 5 μm radius tip. A scan length of 2 mm with a scan speed of 0.2 mm s<sup>−</sup><sup>1</sup> and a range of 80 μm were used for all samples. Calibration was carried out prior to every session using a Mitotoyo Precision reference specimen with an Ra value of 2.95 μm.

### 2.5.8. Surface Wettability—Sessile Drop/Contact Angle

Wetting angles of water droplets on the sample surfaces were tested using a sessile drop experiment. H2O was pumped from a syringe at a rate of 1.0 <sup>μ</sup>L·s<sup>−</sup>1, from a height 4.0 ± 0.5 mm above the sample's surface. A drop settling time of 10 s was implemented prior to data collection.

### *2.6. In Vitro Biocompatibility*
