Influence of Different Acid Etchings on the Superficial Characteristics of Ti Sandblasted with Al2O3

Some implant manufactures use Al 2 O 3 instead TiO 2 powder to sandblast the machined dental implant, because Al 2 O 3 powder is commercially more easily available and is cheaper than TiO 2 powder. However, Al 2 O 3 powder usually leaves aluminum oxide contamination on the surface, which is potentially toxic. In this work, we subjected Ti discs previously sandblasted with Al 2 O 3 powder to 5 different acid etchings in order to verify which treatment is able to remove incorporated particles of Al 2 O 3 from the surface. One group of samples were only sandblasted and served as control. The samples were analyzed by electron microscopy (SEM, EDS), scanning probe microscopy, and grazing incidence XRD. The control group showed presence of Al 2 O 3 on the surface. Three acid etchings were efficient in removing the alumina from the tested samples. Almost all the tested samples showed higher roughness parameters values than the control samples. Titanium hydride was found in almost all test groups. Moreover, the results suggest that there is no incorporation of the whole Al 2 O 3 particle into the titanium surface after the collision, conversely a particle fragmentation occurs and what remains on the titanium surface are Al 2 O 3 residues.


Introduction
The design of endosseous dental implants have been revisited extensively to decrease treatment time frames by reducing the healing period for osseointegration 1 .Alterations in surface texture and chemistry are the commonly used modifications to increase the biological response to implants.Preparation of a roughened titanium surfaces has long been held and demonstrated to be an effective way to promote the interfacial biomechanical properties of bone-anchored implants by means of increasing the interlocking capacity of surface and consequently enabling a favorable stress distribution of the functional loading of an implant at the interface 2 .Several earlier biomechanical studies in various animal models found that the surface texture of titanium implants has a significant influence on their anchorage in bone [3][4][5] .Some authors also demonstrated the biological advantage of a rougher surface compared to machined or polished surfaces.Fu et al. 6 showed that osteoblast cell adhesion, proliferation and alkaline phosphatase (ALP) enzyme activity on sandblasted surface and laser-scanned surface were higher than those on machine-tooled and polished surfaces.ALP activity is considered to play a major role in bone formation and mineralization 6 .
Bone interlocking or micromechanical anchorage at the interface is not a feature common to all surfaces; to achieve it, a certain level of roughness seems to be required.Some authors 7,8 showed that machined surfaces do not achieve micromechanical anchorage at the interface.
Bone was not found attached to the machined surface when removed, whereas anchorage has been documented for other surfaces with physical and chemical surface modifications 4,7,8 .The surface topography obtained by acid etching can be modulated according to prior treatment, for example, by sandblasting, using acid mixtures, using different temperatures, and using different etching times 9 .A combination of blasting and acid etching has been a commonly used surface modification technique during the last two decades.The reason for the combination of methods is that the blasting procedure hypothetically achieves an optimal roughness for mechanical fixation whereas the additional etching smoothes out some sharp peaks.The resulting surface has an improved potential for protein adhesion, considered to be important for the early bone-healing process 10 .Acid etching is a subtractive method, wherein pits are created on the titanium surface 5 .The etching process corrodes the titanium surface greatly, creating irregular pits of varying depth, and produces a microroughness in the range of 0.5 to 3 µm, depending on the etching conditions.When the implant surface is sandblasted prior to etching, a microroughness is superimposed on top of the macroroughness.Etched surfaces have been documented to lead to more bone apposition 11 and to enhance the interfacial strength as measured by removal torque 3 or push-out tests 12 when compared to machined surfaces.
Regarding the first step of surface modification, the sandblasting process on machined dental implant is usually made with aluminum oxide (Al 2 O 3 ) or titanium oxide (TiO 2 ) powder.Some dental implant manufactures use Al 2 O 3 to sandblast the machined dental implant before the acid etching process, because Al 2 O 3 powder is commercially more easily available and is cheaper than TiO 2 powder.However, as the sandblasting particles used during the roughening step may not be completely removed from the implant surface during the etching process, sharp-edged alumina particles are left and can potentially be released during implant insertion 9 or during the osseointegration process.
The effects of Al 2 O 3 are definitely a cause for concern.Although Piattelli et al. 13 showed that residual Al 2 O 3 particles on the implant surface could not affect the osseointegration of titanium dental implants, their in vivo study confirmed this only for a very short-term, since the rabbits used in the study were euthanized after only 4 weeks.More recently, Canabarro et al. 14 evaluated the response of osteoblasts derived from human alveolar bone on to different modified titanium surfaces.They found that the presence of Al 2 O 3 could possibly interfere with the nucleation of apatite crystals during the process of mineralization.Therefore the potential effects of slow and continuous release of trace metals cannot be ignored.In the short term, however, trace elements may be responsible for tissue toxicity and wound breakdown, leading to failure of the implant 15 .Although there are no clear reports of machined titanium implants failing due to contamination of the surrounding tissues by impurities, the mere presence of associated metallic elements in the adjacent tissues is a matter of concern 15 .It is therefore relevant to develop a technique of producing roughness on titanium surfaces with the use of Al 2 O 3 powder in the sandblasting process without the addition of surface impurities.
As the blasting procedure with Al 2 O 3 powder produces desired superficial characteristics to osseointegration, but leaves sharp-edged alumina particles on the surface (which may be toxic), it is the aim of the present study to prepare commercially pure titanium (cpTi) grade IV discs sandblasted with Al 2 O 3 and submit to 5 different acid etchings described in the literature, in order to verify which acid etching (if any) is able to remove the Al 2 O 3 from the previously sandblasted surface.

Material and Methods
Eighteen machined cpTi grade IV discs (12.7 × 2.0 mm) were used as the substrate material for the experiment.All discs were cut from a rod using a IsoMet  Low Speed Saw (Buehler  , Lake Bluff, USA) with a Diamond Wafering Blade No. 11-4244 (4" diameter, 0.012" thickness) from the same manufacturer.
The samples were embedded in polymethyl methacrylate, in order to be polished by a polishing machine.The samples were sandblasted at a pressure of 4 bar with Al 2 O 3 powder (average granulometry of 250 µm) in a KaVo Strahlstation EWL 5423 (KaVo Dental GmbH, Biberach/Riβ, Germany).The samples were kept manually at a distance of approximately 2 cm from the blast nozzle, and the sandblasting was oriented perpendicular to the disc surface.Then all samples were ultrasonically cleaned in acetone for 30 minutes, in alcohol for 30 minutes, and finally in deionized water for more 30 minutes, to remove loose particles of Al 2 O 3 .
The discs were separated into a control group with 3 discs and 5 test groups with 3 discs each.Each test group was acid-etched by five different acid etchings (defined as groups AT1 to AT5) previously described in the literature [16][17][18][19][20] .Table 1 presents a detailed description of these 5 acid etching treatments.The firing in vacuum was performed in a VITA Vacumat 40T vacuum furnace (VITA Zahnfabrik H. Rauter GmbH & Co.KG, Bad Säckingen, Germany).
The surface morphology of treated samples was examined by scanning electron microscopy (SEM -JEOL, model JSM-5310, Tokyo, Japan).The SE mode with an acceleration voltage of 25 kV was selected for SEM analysis and the pressure was maintained below 1×10 -5 Torr.The load current (LC) was approximately 85 μA.For a direct comparison of the surface morphology, the same magnification of 1000× was selected for all samples.
In order to obtain quantitative analysis of the surface roughness, atomic force microscopy (AFM -NTegra Aura, NT-MDT, Moscow, Russia) of the samples was performed.AFM images were acquired in air using semicontact mode with a NSG 01 sharpened gold-coated silicon tip (nominal spring constant of 2.5-10 N/m and nominal resonance frequency of 110-200 kHz, NT-MDT).The scanning area for the measurements was 50 × 50 µm 2 .The images obtained by AFM were characterized by 2 nd order extraction filter, using the software "Image Analysis 2.1.2"(NT -MDT, Moscow, Russia).The seven amplitude surface roughness parameters determined by the software were evaluated (S y , S z , S a , S q , S sk , and S ka ).The mean value and standard deviation of  20 these parameters were obtained from 15 satisfactory scans of each group (5 from each of the 3 samples), from random sites on the surface.The surface chemical composition was analyzed by energy dispersive x-ray spectroscopy (EDS -JEOL, model JXA-8900RJ, Tokyo, Japan).A significant area at the center of each sample was chosen, without incorporating chemical analysis of the sample holder.All the analyses were made with a magnification of 200×.The elemental chemical composition was determined by the mean value and standard deviation from the 3 samples of each group.
Moreover, grazing incidence X-ray diffraction (GIXRD) measurements were carried out in a Ultima IV X-ray diffractometer (Rigaku, Tokyo, Japan), using Cu-Kα 1 radiation at 30 kV and tube current of 20 mA, without any filter or monochromator, in the angle range of 10°-90° (2θ) with a grazing incidence of 3°, making the diffraction sensitive to the surface.The step of measurement was set to 0.05° with a scan rate of 0.5° per minute.The divergence slit was set to 1 mm, with a Div H.L. Slit of 2 mm.The results were analyzed in Search-Match software (Crystallographica, Oxford, United Kingdom).GIXRD experiments were carried out in order to distinguish chemical compounds at the sample surface, mainly aluminum and titanium compounds.

SEM analysis
Figures 1a to 1f revealed characteristic differences at the microscopic level according to the surface modification methods, as measured by SEM.The control group samples were mainly characterized with facets produced by blasting, with some smooth areas and other very rough areas (Figure 1a).Differences of the pits and facets were obvious between the implants due to differences of the etching processes.Incorporated particles were clearly observed in SEM analyses in the control samples and in the AT5 samples (white little clusters throughout the surface -Figures 1a and  1f).Samples submitted to acid etching classified as number 1 (AT1) showed the presence of shallow pits (Figure 1b).AT2 samples showed a surface with many pits, with large variation in the diameter of the pits (Figure 1c).AT3 samples showed a surface with a great number of pits (Figure 1d).AT4 samples showed a smoothening of the facets on the surface (Figure 1e) when compared with the control group samples, which suggests that the acid etching was not effective to change surface morphology.AT5 showed the presence of similar irregularities, as the facets present in control group, but now with the creation of small pits throughout the surface (Figure 1f).

AFM analysis
The qualitative and quantitative surface topography demonstrated different degrees of roughness.Table 2 presents the mean values of tridimensional roughness parameters for the control group and acid-etched samples (groups AT1 to AT5), as determined by AFM.It can be observed that the surface of the samples from group AT1 showed smaller values of the amplitude roughness parameters than the control group, whereas the samples from AT2, AT3, and AT4 groups showed higher values.AT5 showed closest values of S a , S q , and S z as compared to the control group.

EDS analysis
EDS analysis of the surfaces showed titanium and oxygen to be the most common elements in all groups, followed by aluminum (Table 3).The presence of oxygen showed a modest variation in atomic concentration (%at) between the groups (about 40%).There was a higher variation in concentration between the samples with the elements titanium and aluminum.Groups control and AT5 showed a smaller %at of titanium (Figure 2a).Aluminum was found in higher concentrations in groups control, AT3, and AT5, with values ranging 6.5-7.5% in these groups (Figure 2b).It was not found in group AT2, and showed low concentration in groups AT1 and AT4 (0.3% and 1.4%, respectively).Sulfur was only detected in groups AT3 and AT4, at low concentrations (about 1%).

XRD analysis
Figure 3 presents a typical GIXRD diffractogram for a sample from the control group, where diffraction peaks were labeled according to Miller indices, as described elsewhere 21 .The diffractogram analysis confirmed the presence of titanium (Ti) and aluminum oxide (Al 2 O 3 ).The GIXRD results obtained for the samples from groups submitted to acid etching showed the presence of Ti in group AT1 (Figure 4a), Ti and titanium hydride (TiH 2 ) in group AT2 (Figure 4b), Ti, Al 2 O 3 , and TiH 2 in group AT3 (Figure 4c), Ti and TiH 2 in group AT4 (Figure 4d), and Ti, Al 2 O 3 , TiH 2 , and rutile (TiO 2 ) in group AT5 (Figure 4e).

Discussion
Surface contamination of the alloys is a recognized problem 22 and must be avoided in order to retain the surface characteristics of the metal as well as to prevent additional metallic ion release into the host tissues.This is particularly essential in metals/alloys intended for surgical implantation, i.e. endosseous dental implants, where potentially both local and distant toxic effects may occur.
Populations of macrophages have been shown to undergo increased metabolic activity lasting up to 90 days in the case of rough-surface implants, by which time smoothsurface implants were quiescent and encapsulated by fibrous tissue.Therefore, whatever the nature of the contaminant, its effects on the implant surface characteristics, and the consequent reaction at the implant-tissue interface, need to be recognized as one possible factor in clinical failure 23 .
One possible contaminant of great concern is the residues of Al 2 O 3 from the sandblasting process.Al Jabbari et al. 24 has recently demonstrated that alumina fragments embedded on the alloy surface resist steam jet cleaning, suggesting that the fragments are very firmly retained.A similar conclusion was reached from previous findings that Al content after sandblasting with Al 2 O 3 did not decrease significantly after hot steam cleaning 25 .Acid etching seems to be the only method of removing or significantly reducing the presence of residues of Al 2 O 3 on the surface of a metal/alloy after sandblasting with Al 2 O 3 .The results of the present study suggests the same, since the samples were ultrasonically cleaned for 90 minutes in three different ways (acetone, ethyl alcohol, and deionized water), and even after this procedure, the samples of the control group showed a considerable presence of Al 2 O 3 in the surface, as observed in the GIXRD diffractogram (Figure 3).Thus, the particles are somehow incorporated into the titanium surface.However, it can also be suggested that the Al 2 O 3 particles were fragmented after collision with the titanium surface, since no whole Al 2 O 3 particle (250 μm diameter) were observed in any of the SEM images from the samples of the control group.
Nonetheless, the acid etching may affect the surface roughness and morphology, but not always in a negative way.As already mentioned, etched surfaces have been documented to improve bone apposition 11 .In the groups AT2 and AT3, SEM images showed that the etching has created micropits over the surface macrorugosities obtained by sandblasting, as compared to the control group, which is also reflected in higher values of the amplitude roughness parameters for these samples.However, specific to dental implants, studies have shown that histologic and biomechanical characteristics were improved due to increases in the as-machined surface texture by varied methods resulting in average implant surface roughness (S a ) ranging from 0.5 to 2 µm 26,27 .Thus, the values found in the present study for the roughness parameter S a in all studied groups except AT1 seem to be appropriate (0.56-0.82 µm, see Table 2).
Another observation concerning the surface roughness is that the observed values of the dimensional amplitude roughness parameters (S y , S z , average, S a , S q ) were higher than the control sample, agreeing with An et al. 23 .On the other hand, Bathomarco et al. 29 have observed the opposite, i.e. that the chemical etching lead to a decrease in the surface roughness.This divergence may result from differences in the conditions of blasting procedure.While the present study used Al 2 O 3 particles with 250 μm in diameter with a blast pressure of 4 bar, the other study 29 used titanium oxide (TiO 2 ) particles with 50 μm in diameter with a blast pressure of 3 bar.
Buser et al. 4 compared machined, dual-etched and titanium plasma-sprayed (TPS) implants with    sandblasted + etched surfaces (as performed in the present study) in a pig model.The R a for the machined, etched and TPS surfaces was 0.15, 1.3 and 3.1 µm, respectively, and 2 µm for the blasted + etched surface.The removal torque was significantly higher for the blasted + etched implants.The results can be interpreted in a way that optimal roughness also exists for blasted + etched surfaces or that the additional micropits added to the surface may be more important than a further increase of the height deviation 10 .Similar findings were found in a rat model by Abron et al. 30 and by Marinho et al. 31 .They have found the strongest bone response with blasted + etched implants in comparison with machined surfaces.
It has been reported that the etching process modifies the Ti surface composition of SLA-treated implants, and XRD and metallographic microscopy analysis indicated the presence of 20 to 40% of titanium hydride (TiH x , x ≤ 2) in addition to Ti 32 .Before attacking the metallic titanium, the acids must first dissolve the protective titanium oxide layer.During the course of the corrosion process of titanium, native hydrogen ions (H + ) are released 33 .These small ions diffuse rapidly into the metal because the latter is left without its dense protective oxide layer.Therefore, the sub-surface is enriched with hydrogen 34 .When hydrogen saturation is reached, titanium hydride is formed.Titanium hydride may be biologically important because a hydride layer is much better suited as a template for binding biomolecules chemically onto a titanium surface 35 .In the present study, as the surface of the samples were also modified by acid etching after sandblasting, it was not surprising that titanium hydride was found in almost all test groups (from AT2 to AT5), as demonstrated by GIXRD analysis.As the AT1 samples did now show the presence of titanium hydride on the surface, it can not be suggested that the acid-etching classified here as AT1 is weak.Two observations favor this opinion.First, AT1 was able to remove almost all content of Al from the surface.Second, according to the observations of Szmukler-Moncler et al. 36 , for cp Ti, the absorbed hydrogen could not be related to the vigor of the etching bath.
EDS analysis of the surfaces showed titanium and oxygen to be the most common elements in all groups, most probable due to the natural formation of a passivating titanium oxide layer (mainly 2 ) just after sample surface preparation.It was previously observed, by using XPS analysis, that dental implants surfaces treated only by sandblasting and acid etching consists of oxidized titanium 37 .TiO 2 can exist in three crystalline forms named anatase, rutile, and brookite, all with different physical properties.The quasi-amorphous oxides can be crystallized by heating to temperatures of 400-500 °C, which leads to transformation to more crystallized phases 38 .Thus, the presence of rutile phase in group AT5 can be explained by the thermal treatment by which the samples of this group were subjected (firing in vacuum at 600 o C for 10 minutes).
As the substrate, Ti was detected in all samples (control and test groups).The GIXRD diffractograms also showed the presence of Al 2 O 3 in groups AT3 and AT5, demonstrating a similarity with the results of EDS, which also showed a greater presence of Al in the same test groups.As the chemical analysis performed by EDS does not distinguish compounds, the results suggest that the aluminum signal comes from alumina.
Sulfur in the samples of groups AT3 and AT4 can be the originated from the H 2 SO 4 used in the acid etching process.As this acid was also used to etch samples from groups AT2 and AT5 and were not detected by the EDS analysis, it can be suggested that whether the EDS technique was not sufficiently sensitive as to detect low concentrations of sulfur in these groups or differences in the acid etching processes of each group may have influenced the absence of this element in the samples of groups AT2 and AT5, because acid mixture, bath temperature, and etching time may affect the chemistry of the surfaces 5 .
Considering that the observed roughness parameter values, morphology, and atomic concentration of Al for the samples from group AT5 are similar to the control group, it can be suggested that the AT5 was not effective in changing the roughness and the chemical composition of the titanium disc surfaces when compared to samples only sandblasted (control group), except that the heat treatment has led to a higher crystallinity of TiO 2 on the surface in the form of rutile phase.
It was stated that the sandblasting particles used during the roughening step are not completely removed from the implant surface during the etching process 9 .Therefore sharp-edged alumina particles are left and could potentially be released during implant insertion.This raises questions regarding the added value of sandblasting prior to etching on the anchorage of etched implants 5 .However, it was demonstrated in the present study that some acid etchings can completely (or almost completely) remove the aluminum from a titanium surface previously sandblasted with Al 2 O 3 .

Conclusions
The present study demonstrated that the acid etchings here named as AT1, AT2, and AT4 are efficient in removing residues of Al 2 O 3 from the previously sandblasted surface.This is suggested based on the surface sensitive analyses presented here (SEM, EDS, GIXRD, and AFM).Moreover, it can also be suggested that there is no incorporation of the whole Al 2 O 3 particle into the titanium surface after the collision (for the blasting parameters used, 4 bar and 250 µm average particle diameter), conversely a particle fragmentation occurs and what remains on the titanium surface are Al 2 O 3 residues.Last but not least, since the present study has used discs as the substrate material, the observed results can not be directly extrapolated to complex surfaces like those in cylindrical threaded titanium implants.

Figure 2 .
Figure 2. Graphic presentation of atomic concentration of titanium (a) and aluminium (b) in the sample groups, as determined by EDS analysis.
o C for 10min Iwaya et al.

Table 3 .
Atomic concentration (%at) of elements according to the sample groups, as determined by EDS analysis (mean values from 3 samples).