Enhancing the Biocompatibility of Additively Manufactured Ti‐6al‐4 V Eli with Diamond‐Like Carbon Coating

Orthopedic implants provide patients with an opportunity to regain functionality lost from illness, disease, or injury. Recent advancements in additive manufacturing (AM) techniques have allowed for the increased customization of Ti‐6Al‐4V ELI (extra low interstitials) implants to complement natural variations in the human anatomy. Yet, the low bioactivity of Ti‐6Al‐4 V ELI and possible adverse effects from the leeching of aluminum and vanadium complicate the post‐operation recovery process. In this work, Ti‐6Al‐4 V ELI samples are printed using the electron beam melt technique in two directions and coated with diamond‐like carbon (DLC) to examine whether their biological properties can be improved. By conducting in vitro studies with Saos‐2 osteosarcoma cells, the effects of morphology and surface chemistry are correlated to the bioactivities of the coated and uncoated samples. The outcome of the study suggested that DLC coating is a viable method for controlling the surface bioactivity of a material. It indicates that a carbon coating, along with an appropriate topography, has the potential to promote the proliferation and maturity of bone cells and hence enhance the performance of additively manufactured products in next‐generation biomedical applications.


Introduction
Orthopedic implants are often required to restore functionality to the human body damaged by disease and degeneration.Present arthroplasty surgery requires surgeons to fit commercially DOI: 10.1002/admi.202300225available metallic implants into each individual patient.These implants are massproduced, biologically inert, and often do not conform to the specific requirements and dimensions of the individual patient.The lack of bioactivity and specificity can result in prolonged recovery, joint instability, and increase the risk of revision surgery. [1]As such, there is a real clinical need to advance the field of orthopedic implants to enhance the recovery and quality of post-operative life.
Biomaterials used for orthopedic implants have gone through systematic advancement throughout the ages.Beginning with the early generation of materials which focused on being biologically inert and non-toxic, to current generations of materials which are biologically activated for promoting specific cellular responses and healing. [2]A key challenge in this field is to generate biomaterials that can successfully provide physical and local environments that promote cell adhesion, proliferation, and differentiation to facilitate tissue development and long-term acceptance of the implant. [3]Currently, Ti-6Al-4 V Extra Low Interstitials (referred to as Ti64 from this point on) alloy is the most used implantable material for orthopedics due to its excellent strength and oxidation resistance. [4]xtensive studies by Geetha et al. and Shah et al. have verified the biocompatibility of commercially pure Ti64 for both in vitro and in vivo tests. [4,5]However, titanium alloys have innate drawbacks which limit their effectiveness for orthopedic applications.Firstly, titanium is a poor thermal conductor, causing work hardening of the surface during machining.This makes Ti and Tibased alloy both difficult and costly to work into complex geometry using traditional subtractive manufacturing methods for customized components as often needed in prostheses. [6]Secondly, although Ti-based alloys are biocompatible, they have limited capability to form strong bonds with the host tissue, in this case, bone. [7]Slow osseointegration will prolong healing which can increase the risk of fibrous tissue development during healing and loosening of the implant in the long term, leading to revision surgery. [8]As such, the development of implantable materials which have better bioactivity for specific purposes is highly desirable.
Additive manufacturing is well suited to biomedical applications because it enables the fabrication of highly complex geometry with a large degree of customization and cost-effectiveness for the production of bespoke components. [1,10]This approach addresses some of the limitations of working with Ti-alloy using traditional manufacturing techniques.3b,11] Despite the promises of additive manufacturing, there remains a need to study and understand the biocompatibility and cytotoxicity of additive-manufactured materials.It has often been assumed that their bioactivity remains similar regardless of the means of fabrication, however, the difference in processing conditions may induce a difference in surface chemistry and morphology, both of which can impact the biological response.Despite the general lack of passive dissolution of bulk Ti-alloys, corrosion and mechanical wear can both release metal ions that are cytotoxic. [9]Studies have raised concerns about the long-term use of Ti64 due to the breakdown of wear debris and the slow release of aluminum and vanadium in the body, which can lead to inflammatory reactions. [12]In particular, the leaching of aluminum is of potential concern as the accumulation of aluminum ions in the brain has been linked with the increased possibility of Parkinson's and Alzheimer's disease. [13]A study by Shah et al. [14] has indicated that although Ti64 manufactured by electron beam melting (EBM) promoted cellular attachment and proliferation, the interfacial tissue between implant and bone was found to be less mature than the native bone after a period of prolonged healing.This is a significant concern for biomedical applications and suggests that further work in this area is required to confirm the biocompatibility of 3D-printed implants.
The deposition of a thin layer of material to change the surface morphology and chemistry is a proven strategy to enhance the performance of implants. [15]Thin film coatings of additivemanufactured implants should likewise offer a significant enhancement to their performance.Diamond-like carbon (DLC) is a US Food and Drug Administration (FDA) approved biocompatible material that is chemically inert and offers mechanical flexibility with tunable compositions, which allows for its use in many biomedical applications. [16]DLC is a thin film comprising an amorphous form of carbon with a high proportion of carbon atoms in hybridized sp 3 bonds. [17]This bonding structure allows the DLC material to exhibit high mechanical hardness, high modulus of elasticity, good friction-reducing ability, and high wear resistance. [17,18]Furthermore, the material can be doped to change the surface energy so that it be used in antifouling and antibacterial applications. [19]They are superior to other diamond-based materials for orthopedic implants since their non-crystalline nature overcomes the limitations of delamination, extreme hardness, inhomogeneous surface chemistry, and particle leaching. [20]13c,21] In a study by Kim et al., DLC coatings were successfully deposited onto Ti-6Al-4 V materials and significantly showed enhancement to the wear performance of the material. [22]n our previous work, we investigated the performance of DLC thin films on bulk Ti-6Al-4 V. [23] Our results indicated that the DLC-coated specimen exhibited superior corrosion resistance in simulated body fluids and in vitro hemocompatibility.Though limited improvement in cellular differentiation and mineralization as compared to their bare counterpart when cultured with Saos-2 human osteosarcoma cells.
The combination of thin film coatings with additive manufacturing processes opens the possibility of tailoring both the surface chemistry and macroscopic structures.It is expected that such a combination would have an advantageous effect in eliciting specific cellular responses favorable to the healing and regeneration of the surrounding body tissue. [24]12b] Thereby mitigating potential However, published studies on the biological effects of DLC coatings deposited onto 3D-printed material substrates remain limited.The scope of this research is two-fold.It is to first investigate whether the surface chemistry of the printed material can be altered while leaving the topological features intact through the application of a conformal DLC coating.Secondly, it is to determine whether a DLC coating can enhance the bioactivity of the additively manufactured Ti64 for orthopedic applications.In addition to changing the surface chemistry, the effect of surface topography will also be studied by building the material in two different orientations.This work will garner insight as to how cellular response can be stimulated from both the physical and chemical environment of the implant.

Results and Discussion
The surface morphologies and compositions were controlled by the printing direction and DLC coating, respectively.The various types of samples used for this study are, Ti64(V) (Ti-6Al-4V-ELI printed vertically), Ti64(H) (Ti-6Al-4V-ELI printed horizontally), DLC-Ti64(V) (DLC coated Ti-6Al-4V-ELI printed vertically), and DLC-Ti64(H) (DLC coated Ti-6Al-4V-ELI printed horizontally).They can be visualized in the photos depicted in Figure S1, Supporting Information.The different build directions provided an effective way to vary the surface roughness, as the horizontal build exhibited a smoother prominent surface while the vertical build exhibited a rougher surface.This is because, with the horizontal build, the electron beam raster across the prominent surface melts the particles more effectively.Conversely, the electron beam does not raster across the prominent surface in the vertical build, resulting in a higher surface roughness as the particles are fused but only partially melted.The DLC coatings were deposited using a previously published technique, yielding a surface coating of ≈1 μm. [25]After the coating process, the surfaces turned from metallic grey to black.X-ray Diffraction (XRD) was performed to confirm that the coating process did not affect the crystalline structure of the printed materials (Figure S2, Supporting Information).The XRD patterns of the printed samples exhibit reflections consistent with those observed for Ti-6Al-4 V in the literature. [26]DLC-coating led to a slight increase in the lattice parameter when comparing Ti64(H) with DLC-Ti64(H), however, the lack of new peaks or discernible changes to the relative peak position indicates that the coating process did not affect the crystallographic ordering of the sample.

Surface Morphology Characterization
The topography of the samples was examined by scanning electron microscopy (SEM) and white light interferometry (WLI), as can be seen in Figure 1.These images agree with the visual inspection, with more protrusions observed in the vertically built sample as compared to those that were printed horizontally.From the micrographs in Figure 1a,b, partially melted particles embedded on the surface can be seen in Ti64(V) and DLC-Ti64(V), which led to a high surface roughness.Conversely, in Figure 1c,d, the horizontally built Ti64(H) and DLC-Ti64(H) contain smooth directional features where the powdered particles have been completely melted by the electron beam.These features are corroborated by the WLI measurements, which covered a larger area.Using the data from WLI, the arithmetic average (R a ) surface roughness of each sample over a representative sample of 233 × 211 μm is presented in Table 1.The partially melted particles 2.9 ± 0.1 μm seen in the vertically built sample have a significant impact on the surface roughness of the materials, resulting in R a values an order of magnitude higher than their horizontally built counterpart.Interestingly, the coating does not significantly alter the R a value.This indicates that the DLC coating is conformal, and the surface features were not altered by the coating process.These results confirmed the visual observations that built direction contributes significantly to the surface roughness while the coating process does not significantly alter the surface features.

Surface Composition Characterization
Figure 2 presents the spectra obtained from X-ray photoelectron spectroscopy (XPS) and low energy ion scattering (LEIS) of the four test materials.Both these techniques are surface sensitive, with XPS analysis probing to a depth of ≈5 nm, while LEIS interacts exclusively with the surface atomic layer.XPS scans of the four samples over a broad energy range are shown in Figure 2a.
From assessing the spectra, the built direction does not appear to influence the chemical composition.Peaks corresponding to Ti, Al, and V can be seen in both Ti64(H) and Ti64(V).There is also a strong oxygen peak and a trace amount of carbon.Coating the sample with DLC drastically changed the chemical composition of the surface.In both DLC-(H) and DLC-(V), the carbon peak dominates the spectra, with the metallic peaks no longer visible.There is a trace amount of oxygen detected, indicating that some functionalization might have happened.Figure 2b shows elemental information determined by the LEIS measurement of the outermost surface layer, which terminates the sample.The expected positions of the various elements are labeled.The presences of calcium and nitrogen are likely due to surface contamination from the fabrication of the Ti64 and DLC coating, respectively.The LEIS analysis of the surface composition gives a similar trend to that of XPS.In the uncoated samples, the surface is dominated by the Ti and O atoms.Noticeably, vanadium is largely absent from the surface.The lack of vanadium on the surface of Ti64 materials has been reported in the literature. [28]In the coated sample, the metallic features disappear while the carbon and oxygen are visible.The lack of metallic content observed on the surface agrees with the fact that DLC has complete coverage of the surface.
To further investigate the surfaces of the respective materials, the elemental compositions of the surfaces are quantified by integrating the area under the peak and adjusted according to the sensitivity factor of the different elements as preloaded from the CASA XPS software package.This can provide a better interpretation than just visually comparing peak heights, and the results are presented in Table 2.
For further analysis, narrow scans with both enhanced energy resolution and sensitivity were conducted around the peak regions of Ti2p and C1s.The spectra were deconvoluted to reveal the various chemical environment present (Figure S3, Supporting Information).In Ti64(H) and Ti64(V), the Ti2p peak is mainly comprised of titanium atoms in oxygen bonding (458 eV), with both titanium atoms in metallic and partially oxidized environment observed in the shoulder (453 and 455 eV).The narrow scans of DLC-Ti64(H) and DLC-Ti64(V) show that the Ti2p fea-ture was completely suppressed by the DLC coating.This further supports that the coating process is conformal to the surface.The C1s spectra for both the uncoated and coated sample are shown in the bottom row of the supplementary figure.The carbon feature seen in the uncoated Ti64 samples corresponds to adventitious carbon contamination with a trace of carbide formation in the lower energy shoulder (282 eV).The DLC-coated samples showed a carbon peak comprised predominantly of carbon atoms in sp 3 bonding configuration with a trace amount of oxygen-containing moiety visible in the higher energy shoulder.The chemical quantification confirms and highlights a few key observations.Firstly, the print direction does not significantly influence the chemical composition of the surface.This further support the view that the print direction is an effective way to control morphology without affecting the chemical composition.Secondly, despite the results from bulk measurements, the surfaces of the uncoated samples are dominated by metal oxides rather than the anticipated alloy composition.Lastly, the surface of the coated sample is predominantly carbon, with very low oxygen content.This reinforces the perspective that the DLC coating is predominantly composed of saturated carbon.To better illustrate the difference between the elemental composition between the surface and sub-surface layers, composition determined by energy-dispersive X-ray spectroscopy (EDS) is also presented in Table S1, Supporting Information.In comparison, EDS has a much longer penetration depth, as evident with the detection of the SiC adhesion layer.The most striking difference is the lack of oxygen detected and the increased vanadium content in the quantification of EDS as compared to XPS.The lack of oxygen supports the interpretation that the detected oxygen is predominantly from a surface oxide which supports both the XRD results and the fact that oxygenation was not part of the built process.The absence of vanadium on the surface is a reminder that surface composition can be quite different from the bulk.

Wettability
From the surface analysis, the coated and uncoated samples are made of different chemical species, which should influence the surface interactions.In the Ti64 samples, the surface oxide of ti- tanium dominates the surface, whereas this is replaced by carbon in the DLC-coated sample.This is reflected by the contact angle measurements as shown in Figure S4, Supporting Information and summarized in Figure 3. From our results, both uncoated Ti64 samples exhibited a similar wettability to water with contact angles of 63°and 65°for the vertically and horizontally printed, respectively.However, the DLC-coated samples show wettability highly dependent on the printing directions, DLC-Ti64(V) exhibited a contact angle of 86°while DLC-Ti64(H) exhibited a contact angle of 38°.This difference in the wettability trends between the uncoated and DLC-coated samples demonstrates the relationship between surface chemistry and topography.From the elemental analysis, the surface of the uncoated Ti64 is dominated by a surface metal oxide (Figure S3, Supporting Information).It is commonly accepted that titanium oxide is wettable due to the preferential adsorption of hydroxyl groups at their oxygen vacancy sites. [29]This explains the observed wettability despite the difference in topology.The coating process replaced the Ti64 surface chemistry while leaving the roughness of the respective printing directions unchanged.As the coating process does not influence the morphology, vertically printed samples have a much higher roughness as compared to horizontally printed samples.Evidently, the surface wettability of the DLC-coated sample is dominated by the surface roughness, where a smoother horizontally printed surface was easily wetted as compared to being pinned by the vertically printed surface, which exhibits higher surface roughness.

Biological Testing of the Material
The additive-manufactured samples exhibit a variety of surface roughness and surface chemistry based on the printing direction and DLC coating process.To ascertain the biological influences of the surface chemistry and morphology of the printed samples, in vitro testing with Saos-2 human osteosarcoma cells was conducted.
Indirect cytotoxicity was conducted to screen for the presence of any toxic leachable in the media collected from both the uncoated 3D-Ti64(V) and the DLC-Ti64(V) samples.The vertically printed samples were chosen as their higher roughness should facilitate greater contact between the surface and the medium.Quadruplicates of each sample were individually soaked in serum-free cell culture medium containing antibiotics for 4 days, after which the extract was pooled and created into a dilution series which was used to culture Saos-2 cells over a 24-h period, and cell viability was determined.According to international standards, cell viability at or above the 70% level is considered to be acceptable, whereas below that is deemed to be cytotoxic. [30]ata from the MTT assay conducted on the extract from both the uncoated 3D-Ti64(V) and the DLC-Ti64(V) samples are shown in Figure S5, Supporting Information.Exposure to extracts from 3D-Ti64(V) led to a continual decrease in cell viability with increasing concentration as compared to the TCPS (tissue culture polystyrene) control.Though, they remained acceptable to cells when dilutions were at and below 0.5.However, when exposed to a neat extract from the uncoated substrate, cell viability fell below 60%.This implies that the neat extract is considered cytotoxic.In constrast, exposure to extracts from DLC-Ti64(V) resulted in comparable cell viability to the TCPS control up to a dilution of 0.5.Despite a drop in cell viability to 70% when exposed to the neat extract, it is still considered to be acceptable.This shows DLC coating encapsulates the material and can minimize the release of cytotoxic leachable materials.
Figure 4 shows a visualization of cell morphology with fluorescent microscopy after 24 (Figure 4a-d) and 48 h (Figure 4e-h); the actin cytoskeletons of the cells are tagged using a phalloidin stain (which appear in red), and cell nuclei are tagged using a DAPI (C 16 H 17 Cl 2 N 5 ) stain (which appear in blue).All samples showed cell attachment and spreading after 24 h and continued to spread after 48 h.Cells on the 3D-Ti64(H), DLC-Ti64(H), and DLC-Ti64(V) surfaces were more evenly spread than cells adherent to the 3D-Ti64(V) sample.This suggests that both the surface chemistry and morphology of the printed samples affected cell growth.
Figure 5a-d shows images of the cells stained with Cell Tracker Green (CTG).From the images, it was clear that live cells were present on all surfaces after 24 h.The cell morphologies after 24 h on the different surfaces were investigated through SEM (Figure 5e-h).All cells showed evidence of spreading out after 24 h in culture, which indicates that all the surfaces support cell proliferation.Saos-2 cells on the 3D-Ti64(H) and DLC-Ti64(H) samples (Figure 5e,h) showed similar morphology, whilst cells on the 3D-Ti64(V) and DLC-Ti64(V) formed bridge-like structures between the valleys and the unmelted particles on the surface of the material.These bridge-like structures can be attributed to a low energy output/input expenditure from the cell, which is favorable for cell attachment.A similar behavior was reported in a study by Ponader et al., [31] where hFOB 1.19 human fetal osteoblasts also formed in low-lying areas around unmelted particles.The DLC-coated samples also contained cells with extensive filopodia which could indicate that the cells are spreading more actively on these surfaces as compared to the uncoated surfaces.
Though the observation of live cells indicates that these surfaces are not cytotoxic, a 7-day proliferation assay was conducted to further study how the different surfaces support cell proliferation (Figure 6).The data presented represents the mean value of cells adherent to the surface of each sample calculated from triplicate wells of each sample at each time-point with the difference in area considered and normalized to the optical measured on the TCPS control well at day 7. TCPS is a planar polystyrene material that has been commercially modified by gas plasma to provide oxygen-rich surface chemistry creating a very wettable surface (about 60-65°) which is specially designed for cells to stick and grow. [32]This gives a perspective in comparing the test samples with a surface optimized for cell growth.The outcomes from the cell proliferation assay indicated that the DLC-Ti64(V) sample supported the superior proliferation of cells as compared with all other materials, including TCPS, which served as our positive control.The proliferation of cells on the DLC-Ti64(H) samples matches that of the TCPS control but was less than that of the DLC-Ti64(V).Regardless of the printing direction, DLCcoated samples exhibit much higher levels of cell proliferation as compared to their uncoated counterpart, which supported a lower level of cell proliferation even when compared to the TCPS control.These results indicated that the DLC coating has the potential to provide a better surface environment for cell proliferation.This agrees with the SEM observation, which indicates that the cells on the DLC-coated showed enhanced activities.
Cell proliferation is not the only desirable trait in potential biomaterials; the ability to facilitate differentiation and mineralization could shorten the recovery time after the insertion of an orthopedic implant.As bone-forming cells mature, they produce calcium phosphate crystals in a biological process known as mineralization.In our investigation, both alkaline phosphatase (ALP)  and calcium production of the Saos-2 cells are measured after 23 days to assess the differentiation and mineralization of cells adherent to the various substrate over a period of 23 days.A standard culture medium was used to generate an undifferentiated control population of cells, while an osteogenic-specific culture regime was used to induce the differentiation of bone cells assessed by the measurement of calcium.
ALP is regarded as an early marker of cell differentiation.ALP produced by cells adherent to each surface in our assay was corrected for the number of cells attached to those surfaces as evaluated by a concurrent MTT assay.The outcomes presented in Figure 7 indicated that cells on all surfaces produced ALP at levels that were equivalent to or higher than the TCPS control under both standard and osteogenic culture conditions.Ti64(H) showed significantly higher levels of ALP when cells were maintained under osteogenic conditions, which indicated that these cells were in an earlier stage of cell differentiation at day 23.Contrarily, both the DLC-coated samples showed lower levels of ALP than the uncoated samples and had similar ALP levels under both  standard and osteogenic culture conditions confirming that they were at a more advanced stage of osteogenic differentiation where less ALP is expressed and more calcium production would be expected as mineralization occurs.
The outcomes of calcium production by the cells on the test surfaces on day 23, shown in Figure 8, confirmed this proposition.Calcium production by the adherent cells is a marker of the later phase of cell differentiation involving the mineralization of the extracellular matrix produced by the cells as they start to form bone.The data presented in Figure 8 represents the mean value of triplicate samples for each material, normalized to the value for the TCPS control in osteogenic medium and adjusted for the number of cells attached to each surface as determined by MTT assay which was run in parallel with the differentiation assay.From the results, cells on all test surfaces, including the TCPS control, produced significantly more calcium when maintained under osteogenic culture conditions than at standard conditions.The calcium production from all samples exceeds those of the TCPS control, with DLC-Ti64(V) samples exhibiting significantly higher levels of calcium under osteogenic conditions as compared to the other substrates, which had similar levels of calcium production.
The outcomes of the mineralization assay indicate that the cellular responses differed between the different surfaces.In the ALP assays, the uncoated samples printed in both directions exhibited a higher level of ALP production over 23 days of the assay compared to their DLC-coated counterparts, most noticeably in the case of the Ti64(H) surface.All other surfaces exhibited a similar level of ALP production between both osteogenic and normal culture medium regimes.Reduced ALP levels showed that the cells were no longer in an early stage of bone cell differentiation and had moved into the mineralization phase involving calcium production by the cells.This was confirmed by measurements of calcium on day 23 that showed the DLC-Ti64(V) produced the highest calcium levels of all the different substrates tested.This confirmed that the cells on this surface were in the later stages of bone cell differentiation with correspondingly higher levels of bone mineralization.
From the results presented, both surface chemistry and surface morphology can influence cellular response.The additive manufacturing process generated surfaces with different topography according to the printing direction.The DLC coating process replaced the metal oxide surface composition of the additively manufactured Ti64 materials with one terminated by sp 3 carbon without changing the surface roughness.Thereby separating the effects of surface topology, determined by the printing direction, from surface chemistry, determined by the surface coating.From the contact angle measurement, the bare surfaces terminated by metal oxide have a similar wettability regardless of the topography, with a considerable difference in wettability between the different surface textures after the application of a conformal DLC coating.The DLC-Ti64(V) surface showed a relatively high contact angle (86°), presenting a poorly wetting surface with low free energy compared with the more wettable DLC-Ti64(H) surface (38°).The consequence of these physicochemical measurements on the cellular response can be seen in the outcomes of a series of in vitro assays using bone cells.From the 23-day proliferation assay, the DLC-Ti64(V) surface supported more rapid cell proliferation than the uncoated Ti64(V) surface or either of the Ti64(H) surfaces or the TCPS control.It is reasonable to expect that an organic surface, such as that provided with a DLC coating, might better mimic an organic environment as compared to the uncoated metal oxide due to a higher rate of protein adsorption on the surface alongside more favorable surface energies.However, taking the current data together, we found that the DLC Ti64(V) surface was rougher (by SEM and R a measurements) and showed a higher contact angle than the DLC Ti64(H) surface.It is known that cells are known to be very sensitive to surface roughness (topography) and wettability (surface free energy) of the surface. [33]Data arising from cell assays conducted as part of this study on the two DLC-coated substrates showed that the DLC-Ti64(V) supported a higher level of cell growth, surpassing the planar TCPS control as well indicating that the slightly rough and relatively hydrophobic vertical (V) surface characteristics may be preferable to the growth of bone cells cultured over 7 days.From the cell differentiation assay indicative of cell maturity, the uncoated Ti64(H) surface had higher ALP production in the osteogenic culture medium than both the DLC-coated samples and the TCPS control with adherent cells to be in an early stage of bone cell differentiation despite the pro-osteogenic conditions.This contrasted with the DLC-Ti64(V) surface that showed lower levels of cell differentiation in both osteogenic and normal culture medium similar to the TCPS control, indicative of cells having moved into the later stage of differentiation involving mineralization by the day 23 endpoint.This was confirmed by DLC-Ti64(V) surfaces supporting higher calcium production as compared to other samples.Our assays showed that although the two DLC-coated surfaces started out in a similar way with regards to cell adhesion and spread, the longer-term proliferation/growth and differentiation of cells on the DLC-Ti64(H) surface was slower than on the DLC-Ti64(V) surface version as evidenced by higher ALP (i.e., less mature) and lower Ca in osteogenic medium (less mineralization by cells).Even though the DLC Ti64(H) has inferior mineralization, it was still slightly better than the TCPS standard.Taken together, data showed that both the surface morphology and the chemistry of the DLC-Ti64(V) samples provided an optimal environment for the proliferation and maturation of bone cells under culture conditions.

Conclusion
Additive manufacturing can create bespoke implants tailored to individual patients.From the work presented here, it can be seen that conformal surface coatings can further engineer the biomaterial for specific positive biological outcomes such as cell adhesion, proliferation, and maturity.As surface morphology and surface chemistry are often intimately linked, the ability to tune them independently opens new possibilities for material design.Though more work will need to be done to ascertain the exact mechanism which stimulates cellular interactions, we have shown that it is possible to enhance the biocompatibility of a material by tailoring the surface chemistry and morphology.

Experimental Section
Additive Manufacturing of Ti-6Al-4 V Samples by Electron Beam Melting: Samples were printed using a commercially available Arcam A1 electron beam melting (EBM) system from Ti-6Al-4V ELI powder with a particle size of 50-70 μm supplied by Arcam AB (Mölndal, Sweden) as the feedstock material.The samples were manufactured in both a horizontal and vertical direction for comparison.The process was performed in an inert helium environment at a pressure of 2 × 10 -3 mbar.The build temperature was maintained at 700 °C, with a starting plate temperature of 730 °C and an electron beam diameter of 0.2 mm.The schematic of the EBM system, CAD model, and description of the build orientation can be seen in Figure S6, Supporting Information.
All samples were cleaned after being printed using a three-stage solvent ultrasonic cleaning process.The samples were first washed in an X55 solvent solution for 15 min, then washed in an acetone solution for 15 min, and finally washed in an ethanol solution for a further 30 min.The samples were then dried with nitrogen gas and placed in a sealed petri dish.
Deposition of DLC Coatings: A custom-built plasma-assisted chemical vapor deposition (PACVD) system was used to coat the samples (Figure S7, Supporting Information). [25]Argon gas was fed into the chamber at a flow rate of 10 sccm to reach a gas pressure of 3 × 10 −2 mbar to perform plasma etching.The plasma etching process was performed at an RF power of 200 W for 10 min.Directly following the etching process, an amorphous silicon carbide (a:SiCH) adhesion layer was deposited by introducing tetramethylsilane (TMS) at a flow rate of 40 sccm to reach a working gas pressure of 3 × 10 −2 mbar, and then the RF power supply was set at 200 W for 5 min, resulting in a TMS layer ≈100 nm thick.Finally, the DLC layer was deposited using acetylene (C 2 H 2 ) gas which was fed into the chamber at a flow rate of 100 sccm to reach a working pressure of 6 × 10 −2 mbar.The deposition time of the DLC layer was set at 15 min at 200 W RF power.The DLC coating thickness was measured to be ≈1 μm by surface profilometry (Dektak 3030 stylus profilometer).
Material Characterization: The samples were examined with a variety of techniques to elucidate variations caused by the build direction and coating process.X-ray diffraction (XRD) was carried out to analyze the crystal structure of the material using the PANanalytical X'Pert Pro Materials Research Diffractometer (MRD) system with a Cu K  source (1.5405980Å).The surface roughness was determined by white light interferometry (WLI) measurement with the Bruker-AXS NT-9800 optical surface profiler instrument using a 0.80 numerical aperture and 5× Michelson objective.The microstructure of the surface and the cell morphology were examined using field emission scanning electron microscopy (SEM).SEM was conducted using a Carl Zeiss Auriga SEM instrument.Images were captured from a 15 keV electron beam energy with an InLens secondary electron detector.Critical point drying (CPD) of the samples was performed to preserve the cell morphology using an Emitech K850 CPD system with CO 2 for 2 h.The samples were imaged as is afterward.
The elemental composition of the samples was probed by different techniques to determine how the composition changes with depth.Energy dispersive X-ray spectroscopy (EDX) was obtained using an Oxford Instruments X-Max solid-state SSD attachment for the Zeiss Auriga SEM instrument with a 15 keV electron beam and analyzed using the OEM software package AZtecEnergy.Surface characterization analysis using X-ray photoelectron spectroscopy (XPS) was conducted using a SPECS Sage 150 system, with the MgK line at 1253.6 eV used as the excitation source.Both survey scans (between 0-1000 eV with an energy step of 0.5 eV) and high-resolution scans (step size of 0.1 eV) for the C 1s and Ti 2p features were conducted.Lastly, low energy ion scattering (LEIS) spectroscopy was recorded with a SPECS PHOIBOS 100 hemispherical analyzer (HSA) using a 2 keV helium ion source with a 138°scattering angle.
The wettability of each sample was determined by contact angle measurements.Contact angle measurements were obtained by analyzing photographic images of a 5 μL sessile water droplet placed on each sample using a pipette (Figure S4, Supporting Information).A digital camera with a macro lens was used to take images of the drop profile, which were later analyzed in the open-source program ImageJ.
Biological Characterization: In vitro studies using Saos-2 human osteosarcoma cells were conducted to measure the biological activities of the various samples as compared to tissue culture polystyrene (TCPS).
Indirect cytotoxicity using MTT assay with Saos-2 cells was done to screen for the presence of any toxic leachables in the media collected from each Ti64 sample.The Ti64 samples were first soaked in 70% ethanol for 2 h, after which they were dipped in sterile Baxter's water before being placed individually in wells of a 12-well TCPS plate.To each well, 1.5 ml of serumfree cell culture medium containing antibiotics was added to each well containing an individual wafer and the plates were sealed and held at 4 °C for 4 days.Saos-2 cells were seeded to wells at a density of 5 × 10 3 cm −2 in standard cell culture medium, and after 24 h, when cells were attached and spread, the culture medium was removed and discarded, then replaced with serial dilutions of pooled extracts from the Ti64 samples.A TCPS control in a serum-free medium was also included.After 20 h, phase contrast cell images were collected to determine the effect of the extracts on cell viability.
Cell adhesion and proliferation assays were used to determine the initial rate of cell adhesion to the test substrates and the rate of growth of those cells over a 7-day period.At time points of days 1, 4, and 7, cell numbers were measured by microculture tetrazolium (MTT) cell viability assay.A total of 16 samples of each test surface were used for this assay.For the cell adhesion study, two samples were used at each of the two time-points of 24 and 48 h.For the cell proliferation assay, three samples were used for MTT assay, and one sample was used to check cell morphology at each of the three time-points of days 1, 4. and 7.
ALP and calcium production of cells on each surface were measured after 23 days as a quantitative measure of cell differentiation.ALP is used as an early indicator of differentiation when cells are still proliferative and calcium as an indicator of the later stage of differentiation in which mineralization occurs.An MTT-based cell proliferation assay was also run concurrently to allow data outputs to be corrected for the number of cells adherent to each substrate.For both the ALP and calcium assays, the resulting values were adjusted to take into account the differences in sample areas and the control TCPS wells, with final values expressed as either p-NPP mM cm −2 or calcium mg cm −2 .Data were then normalized to the differentiated TCPS control value to give a percentage value and then corrected to take into account the number of cells present on surfaces as determined by the MTT assay run in parallel over the 23-day period.Details of the methodology employed for the biological characterization can be found in the Supporting Information.

Figure 1 .
Figure 1.SEM micrograph of the a,b) test samples with the corresponding 3D topological map as obtained from e-h) WLI.The depiction from the two methods corresponds to the same sample but at different magnifications; the SEM images are obtained at 200× magnification, whereas the WLI is obtained at 5×.

Figure 2 .
Figure 2. XPS and LEIS spectra of the samples, the peak positions of the various elements are indicated with the dashed lines.

Figure 3 .
Figure 3. Contact angle measurement obtained from the respective samples.

Figure 4 .
Figure 4. Fluorescent images of phalloidin (red) and DAPI (blue) stained cells on the different samples after a-d) 24 and e-h) 48 h.

Figure 5 .
Figure 5. a-d) CTG staining images of cell coverage on the additively manufactured samples after 24 h and e-h) SEM images of the cell morphology on the samples after 24 h.

Figure 6 . 7 -
Figure 6.7-Day proliferation assay with the optical density normalized to that of TCPS at day 7.

Figure 7 .
Figure 7. Alkaline phosphatase (ALP) production by cells after 23 days, the quantity produced is normalized to TCPS.

Figure 8 .
Figure 8. Calcium production by cells on the additively manufactured samples after 23 days, the quantity produced is normalized to TCPS on day 23.

Table 1 .
Surface roughness of the different samples as determined by WLI.