- Email: [email protected]
Protective layer formation on magnesium in cell culture medium
Materials Science and Engineering C 63 (2016) 341–351
Contents lists available at ScienceDirect
Materials Science and Engineering C journal homepage: www.elsevier.com/locate/msec
Protective layer formation on magnesium in cell culture medium V. Wagener, S. Virtanen ⁎ Chair for Surface Science and Corrosion, Department Materials Science, University of Erlangen-Nuremberg, Martensstr. 7, 91058 Erlangen, Germany
a r t i c l e
i n f o
Article history: Received 27 October 2015 Received in revised form 18 February 2016 Accepted 1 March 2016 Available online 3 March 2016 Keywords: Magnesium Calcium phosphate coating Protein influence Corrosion protection electrochemical impedance spectroscopy
a b s t r a c t In the past, different studies showed that hydroxyapatite (HA) or similar calcium phosphates can be precipitated on Mg during immersion in simulated body fluids. However, at the same time, in most cases a dark grey or black layer is built under the white HA crystals. This layer seems to consist as well of calcium phosphates. Until now, neither the morphology nor its influence on Mg corrosion have been investigated in detail. In this work commercially pure magnesium (cp) was immersed in cell culture medium for one, three and five days at room temperature and in the incubator (37 °C, 5% CO2). In addition, the influence of proteins on the formation of a corrosion layer was investigated by adding 20% of fetal calf serum (FCS) to the cell culture medium in the incubator. In order to analyze the formed layers, SEM images of cross sections, X-ray Photoelectron Spectroscopy (XPS), Xray diffraction (XRD), Energy Dispersive X-ray Spectroscopy (EDX) and Fourier Transformed Infrared Spectroscopy (FTIR) measurements were carried out. Characterization of the corrosion behavior was achieved by electrochemical impedance spectroscopy (EIS) and by potentio-dynamic polarization in Dulbecco's Modified Eagle's Medium (DMEM) at 37 °C. Surface analysis showed that all formed layers consist mainly of amorphous calcium phosphate compounds. For the immersion at room temperature the Ca/P ratio indicates the formation of HA, while in the incubator probably pre-stages to HA are formed. The different immersion conditions lead to a variation in layer thicknesses. However, electrochemical characterization shows that the layer thickness does not influence the corrosion resistance of magnesium. The main influencing factor for the corrosion behavior is the layer morphology. Thus, immersion at room temperature leads to the highest corrosion protection due to the formation of a compact outer layer. Layers formed in the incubator show much worse performances due to completely porous structures. The existence of proteins in DMEM seems to hinder the formation of a corrosion layer. However, protein adsorption leads to similar results as concerns corrosion protection as the formed calcium phosphate layer. © 2016 Elsevier B.V. All rights reserved.
1. Introduction The scientific focus on magnesium as bio-degradable implant material in the orthopedic field is not only based on the non-toxicity of magnesium and its mechanical properties, but also on the positive influence of magnesium ions on bone formation [1–3]. Magnesium is a trace element in the human body and thus Mg ions exist abundantly in blood and several tissues. Indeed, half of the Mg ions in the body are stored in bone tissue [1]. Several studies showed that bone formation can be stimulated by the presence of Mg implants [4]. It is assumed that released Mg ions trigger the formation of calcium phosphate compounds on the implant surface that may be pre-stages to hydroxyapatite (HA), which is a main component of human bone [5,6]. The pH increase due to Mg dissolution favors the formation of calcium phosphate as well [7]. ⁎ Corresponding author. E-mail address: [email protected] (S. Virtanen).
http://dx.doi.org/10.1016/j.msec.2016.03.003 0928-4931/© 2016 Elsevier B.V. All rights reserved.
It is as well reported that calcium phosphates and especially HA act as good corrosion protective coatings on magnesium [7–10]. Thus, in several studies HA was deposited on Mg or on Mg alloys in order to increase corrosion resistance for biomedical application [9, 11–13]. For the application as bone implant not only the formation of calcium phosphate in vivo but also the pre-coating with HA is of interest. Blind et al. [14] reported improved osseo-integration due to HA coatings, for example. Different studies showed that precipitation of calcium phosphate compounds on magnesium in simulated body fluids or cell culture media can occur [1,15–20]. At the same time, however, a dark grey or black layer is formed under the white calcium phosphate precipitations in most cases [17,21,22]. Degner et al. propose a corrosion protective effect of this black layer [21]. Willumeit et al. [17] found out that those layers as well mainly consist of calcium phosphate compounds. However, the morphology of the layers was not investigated in detail. In this study, magnesium was immersed in cell culture medium under different conditions. Immersion time, temperature and CO2 concentration
342
V. Wagener, S. Virtanen / Materials Science and Engineering C 63 (2016) 341–351
Fig. 1. SEM images after immersion of Mg in DMEM at room temperature; a)–c) immersion for one day; d)–f) immersion for three days; g)–i) immersion for five days.
were varied. The formed layers were analyzed concerning their thickness, morphology and composition and corrosion resistance. In addition, the influence of proteins in cell culture medium on the formation of a corrosion protective layer was investigated. 2. Materials and methods 2.1. Sample preparation Samples were cut from a cp magnesium rod (25.4 mm diameter, 99.9% purity, Chempur Feinchemikalien GmbH & Co. KG), and afterwards deburred and ground with SiC abrasive paper up to a grid size of 1200. After sonication in ethanol and drying, samples were immersed in ca. 50 ml of cell culture medium for one, three and five days at room temperature and in the incubator at 37 °C (5% CO2, ~95% humidity). As cell culture medium, Dulbecco's Modified Eagle's Medium with 1 g glucose and 3.7 g NaHCO3 (DMEM, Biochrom AG) was used for the immersion at room temperature. For the immersion in the incubator 1% Penicillin-Streptomycin-Glutamine (PSG, Sigma) was added to the described DMEM, in order to prevent growth of microorganisms. For a third testing row, 20% of fetal calf serum (FCS, Gibco) was added to the cell culture medium in the incubator, in order to investigate the influence of serum proteins on the layer formation on magnesium in DMEM. After immersion in cell culture medium for the desired time, samples were rinsed with water and dried with nitrogen. 2.2. pH measurements The pH of the cell culture medium was measured before and after immersion for all parameters. As reference, pH was measured for DMEM at room temperature, DMEM (+ PSG) in the incubator and
DMEM (+PSG) + 20% FCS in the incubator without immersed samples after one, three and five days. 2.3. SEM Cross sections were prepared with help of an ion mill (IM4000, Hitachi), using an acceleration voltage of 6 V and a discharging voltage of 1.5 kV. Samples were cut in the “cross milling” mode with three repetitions per minute with an angle of ± 30°. Afterwards, the cross sections were characterized with a scanning electron microscope (FESEM S4800, Hitachi), using an acceleration voltage between 3 kV and 10 kV. The working distance was varied between 4 and 6 mm. 2.4. XPS XPS measurements were carried out with a high-resolution X-ray photoelectron spectrometer (PHI 5600 USA) using aluminum Kα radiation (1486.6 eV, 300 W) for excitation. The binding energy of the target elements was determined at a pass energy of 23.5 eV with a resolution of b0.4 eV (values measured every 0.2 eV for the high resolution spectra and 0.8 for survey) and a takeoff angle of 45° with respect to the surface normal. The binding energy of the C1s signal was used to correct the spectra for charging. The background was subtracted using the Shirley method in all spectra. To obtain the molar fractions of each species, the peak areas of the measured XPS spectra were corrected with the photoionization cross sections of Scofield [23] σ and the asymmetry parameter β (orbital geometry) [24], which are contained in the sensitivity factors of the acquisition software (MultiPakV6.1A, 99 June 16, copyright Physical Electronics Inc.,1994–1999). For each immersion parameter a survey spectrum and high resolution spectra for C1s, O1s, N1s, Ca2p, P2p and Mg2p signals were measured. C1s narrow scans for one day immersion were fitted with Origin (Gaussian multi-peak fitting).
V. Wagener, S. Virtanen / Materials Science and Engineering C 63 (2016) 341–351
343
Fig. 2. SEM images after immersion of Mg in DMEM in the incubator (37 °C, 5% CO2); a)–c) immersion for one day; d)–f) immersion for three days; g)–i) immersion for five days.
2.5. XRD X-ray diffraction (XRD) spectra were recorded for the different immersion parameters for one day, in order to gain information concerning crystallinity of the formed layers. Therefore, an X-ray diffractometer (X'PERT PW 3040 MPD, Panalytical) with a monochromatic Cu-Kα radiation (λ = 1.54 Å) was used, the divergence slit was put to 0.19 mm. Diffraction patterns were calculated from 2θ (10°–80°) for every sample, with a step width of 0.03° and an incident angle of 1°.
Potential (OCP) was recorded for 30 min to guarantee a rather stable potential for the EIS measurements (not shown). Afterwards, the impedance was measured at the particular OCP of the samples in the range of 100 kHz to 10 mHz with an excitation amplitude of ±10 mV. In addition, polarization curves were measured after EIS. Therefore, voltage was varied within a range of −300 mV relative to the particular OCP and 0 V with a scan rate of 3 mV/s and with a 10 mA current limit in the anodic region. 3. Results and discussion
2.6. FTIR Fourier Transformed Infrared Spectroscopy (FTIR) was conducted with an infrared spectrometer (Thermo Scientific Nicolet 6700/Smart iTR) with a diamond reference crystal for the immersion parameters for one day. Measurements took place in a wave number range of 500–4000 cm−1 with a penetration depth of 2 μm. 2.7. Electrochemistry Electrochemical impedance spectroscopy (EIS) was used to characterize the different Mg samples concerning corrosion resistance. All measurements were conducted using an electrochemical workstation “IM6eX”, a potentiostat “XPot” and the corresponding “Thales”-Software (Zahner-Elektrik GmbH & Co. KG, Kronach, Germany). The experiments were carried out in an electrochemical cell with a three electrode set-up, in which the particular Mg samples acted as working electrodes. As counter electrode a platinum sheet and as reference an Ag/AgCl electrode with 3 M KCl were used. As electrolyte 100 ml of DMEM was chosen, measuring temperature was kept constant at 37 °C (±1 °C). The surface exposed to the electrolyte was 0.7854 cm2. The Open Circuit
3.1. Layer morphology Cross sections were prepared and analyzed by SEM in order to gain detailed information on morphology and thickness of the layers formed during immersion in DMEM under different conditions. Fig. 1 depicts the SEM images of the cross sections for the immersion of magnesium in DMEM at room temperature. All three immersion times show the formation of layers with similar morphology. Layers can be divided into two areas with different structures. While the inner layer is porous, the outer regions show compact structures (see Fig. 1b, e and h). In general, a decrease of porosity in the inner layer from substrate to the outer compact region can be observed. The high porosity near the substrate may be due to hydrogen evolution caused by initial strong magnesium dissolution (see Eq. (1)). Mg þ 2H2 O→Mg2þ þ 2OH− þ H2
ð1Þ
The formation of gas bubbles leads to the formation of a porous initial layer. With the growth of the layer during immersion, Mg corrosion slows down due to isolating effects of the layer and a pH increase (compare Fig. 5) [25], resulting in a decreased hydrogen evolution and thus in
344
V. Wagener, S. Virtanen / Materials Science and Engineering C 63 (2016) 341–351
Fig. 3. SEM images after immersion of Mg in DMEM + 20% FCS in the incubator (37 °C, 5% CO2); a)–c) immersion for one day; d)–f) immersion for three days; g)–i) immersion for five days.
a decrease of porosity. The overview pictures (Fig. 1a, d and g) show crack formation for all immersion times. Cracks extend to the whole layer thickness and may be caused by internal stresses during immersion in DMEM or during drying. For 5d immersion, the formation of structures beneath the layer is found (see Fig. 1i), indicating corrosion. These structures seem to begin directly at the substrate-layer interface and spread into the substrate. Fig. 2 depicts the SEM images for layer formation on magnesium in DMEM in the incubator. Compared to DMEM immersion at room
Fig. 4. Average thicknesses of layers built after immersion of Mg in DMEM at room temperature, in DMEM in the incubator (37 °C, 5% CO2) and in DMEM + 20% FCS in the incubator (37 °C, 5% CO2) with standard deviations (n = 4).
temperature, the morphology of the formed layers differs for the different immersion times. The immersion for one day shows a porous structure with no gradient in pore size or density over the layer thickness. No compact outer layer can be observed (see Fig. 2b and d). In comparison, the 3d and 5d immersion show again different regions with a variation in pore density (Fig. 2e and h). For 3d, pore density decreases from inside to the outer areas. Nevertheless, no completely compact outer layer is formed, as can be seen in Fig. 2e. 5d immersion shows as well areas with different pore density. However, again no completely compact outer layer is
Fig. 5. Development of pH during immersion of magnesium in DMEM at room temperature (RT) in the incubator at 37 °C (Inc) and in DMEM + 20% FCS in the incubator (20FCS). Reference measurements of medium without samples showed no significant differences in pH values.
V. Wagener, S. Virtanen / Materials Science and Engineering C 63 (2016) 341–351
345
Fig. 6. XPS results after immersion of Mg in DMEM for 1, 3 and 5 days; a) atomic concentrations of several signals after immersion in DMEM at room temperature; b) atomic concentrations of several signals after immersion in DMEM in the incubator (37 °C, 5% CO2); c) atomic concentrations of several signals after immersion in DMEM + 20% FCS in the incubator (37 °C, 5% CO2); d) Ca2p/P2p ratios for the DMEM immersion with the different parameters.
formed (Fig. 2i). The reason for the missing compact outer layer for the immersion in DMEM in the incubator may be the rather small increase in pH due to the better buffered system caused by the high CO2 concentration in the environment (see Fig. 5). Thus, magnesium dissolution is not slowed down due to a rise in pH as it may be the case for the immersion at room temperature. Crack formation seems to be more distinctive than for the immersion in DMEM at room temperature, especially after three and five days, where layer thicknesses increase (Fig. 2d, g and h). This may be explained by increasing internal stresses in the thicker layers. The SEM images of the cross sections for the immersion of Mg in DMEM + 20% FCS in the incubator are shown in Fig. 3. Compared to the immersion in DMEM without proteins, rather thin layers are formed and the border between substrate and layer seems to be rougher. While three days immersion shows a similar pore structure as the layers built in DMEM without proteins (see Fig. 3e, d and f), the formed structures after 1d and 5d passivation seem to be more compact. Regularly cracks were only detected after three days, where layers reached a thickness of about 1 μm. This confirms the assumption that cracks originate from internal stresses for sufficient thick layers. Layer thicknesses were taken from the particular cross section SEM images. Mean values were calculated from four different positions in the layer on each sample. The results can be seen in Fig. 4. For the immersion in DMEM at room temperature a slight decrease in layer thickness over time is observed. The maximum values are reached for 1d (3.5 μm). The low standard deviations indicate a homogeneous layer formation. In comparison, the layers formed in DMEM in the incubator show a strong increase in average thickness with immersion time. The thickest layers are achieved after five days (up to 24 μm). Nevertheless, deviations increase strongly for longer immersion times indicating
inhomogeneous layer growth, as it can be seen in Fig. 2g. The addition of 20% FCS to DMEM in the incubator causes a significant drop in layer thicknesses to values of 1 μm and lower. It is assumed that the proteins in the cell culture medium hinder the formation of the corrosion layer. Layer growth at room temperature is for all immersion times very similar and seems to be self-limited. This may be due to the formation of a compact outer layer. Further surface analysis indicates that the formed layers mainly consist of calcium phosphates (see Section 3.2). It is reported that the formation of calcium phosphate layers is triggered by the release of Mg ions and favored by an increase of pH values [7]. The reduced corrosion during layer growth and pH increase as well as the formation of a compact outer layer (compare Fig. 1) decreases the release of Mg ions strongly. As a consequence, calcium phosphate layers do not extend further after the compact layer is built. SEM images show that this stage is already reached after one day. Another indicator for layer growth limitation is the development of pH in the only weakly buffered electrolyte. As Fig. 5 shows, pH values increase within the first day of immersion for the passivation at room temperature. For three and five days, pH values stay more or less on the same level, pointing out that Mg corrosion reaches a plateau. For the immersion in DMEM in the incubator, a more or less steady increase of layer thickness over time occurs. SEM images show that porous layers without completely compact outer regions are formed in the incubator. Thus a limitation like it is observed for the immersion at room temperature does not exist. The pH values show only a small rise due to the better buffering in the incubator. It is assumed that corrosion and thus Mg release occur during the whole immersion time. The large variation in layer thickness may be caused by impurities in the Mg substrate or slightly different pH values in the cell culture medium. As layer formation seems not to be limited within the investigated period
346
V. Wagener, S. Virtanen / Materials Science and Engineering C 63 (2016) 341–351
Fig. 7. C1s narrow scans with fitted curves for 1d immersion in a) DMEM at room temperature, b) in DMEM in the incubator and c) in DMEM + 20% FCS in the incubator.
due to buffering, slightly different immersion conditions may have an influence on the growth. Layer formation in DMEM + 20% FCS seems again to underlie a limitation. As already mentioned, proteins seem to hinder the formation of a calcium phosphate layer.
3.2. Layer composition In order to analyze the composition of the formed layers, XPS spectra and narrow scans were measured. Atomic concentrations for C1s, N1s, O1s, Ca2p, P2p and Mg2p are shown in Fig. 6. In addition, C1s narrow
scans for 1d immersion under the three different conditions can be found in Fig. 7. The layers formed on Mg in DMEM at room temperature (Fig. 6a) show for all immersion times' similar values for the selected signals. The largest variations for the different times can be observed for C1sand O1s-signals. While C1s decreases significantly after three days (~ 20.2 at.%), O1s reaches maximum values (~ 55.5 at.%). N1s signals show values b1.5 at.% for all immersion times with peaks being overlaid by noise. This leads to the assumption that no nitrogen compounds are included in the outer layer (as detected by XPS). Ca2p signals lie around 12 at.%, P2p signals around 7 at.% for all immersion times. Mg2p signals are with values smaller than 3 at.% comparably low, showing that magnesium is not included as a main component in the formed (outer) layer. The high O1s, Ca2p and P2p signals indicate the formation of a calcium phosphate layer. C1s signals may be caused by atmospheric contamination. However, the relatively high values hypothesize the incorporation of carbonate ions in the calcium phosphate layer [26, 27]. The analysis of C1s narrow scan (Fig. 7a) shows that indeed most of the carbon signal seems to originate from atmospheric contamination (C–Hx) or carboxyl or carbonate groups [28]. As for the DMEM immersion at room temperature, the DMEM immersion in the incubator shows no significant variations of the selected signals for the different immersion times (Fig. 6b). Compared to the immersion at room temperature, C1s signals decrease slightly, while O1s signals show an increase. N1s signals show again values b1.5 at.% leading as well to the assumption that no nitrogen compounds are included in the (outer) layer. Ca2p and P2p signals lie in the same range as for the immersion in DMEM in the incubator. However, Mg2p signals show an increase (6 at.%). The higher concentration of Mg in the outer layer may be due to higher Mg dissolution because of the better buffering in the incubator, leading to an incorporation of Mg in the whole layer. The analysis of the C1s narrow scan (Fig. 7b) shows similar results as for the immersion in DMEM at room temperature. Again, the carbon signal is mainly caused by contamination and carboxyl and carbonate groups. However, relative to the peak at about 284.5 eV (C–Hx) the peak at about 289.4 eV (O = C–O, CO2– 3 ) is higher. This may indicate that more carbonate groups are incorporated in the layer due to the significantly higher CO2 concentration in the incubator. The addition of 20% FCS to the cell culture medium in the incubator leads to a significantly different chemical composition of the outer layer as observed by XPS: A strong increase of C1s and N1s signals for all immersion times is seen, while O1s, Ca2p and P2p signals show a decrease (Fig. 6c). The increase of C1s and N1s signals indicates adsorption of proteins on the surface, whereas the lower amounts of Ca and P detected verify the assumption that proteins hinder the formation of calcium phosphate layers on the surface. The C1s narrow scan shows as well that proteins are adsorbed on the surface. Compared to the immersion of Mg in DMEM without FCS, the shape of the overall C1s signal changes due to an additional distinctive peak caused by C–N bindings (Fig. 7c). Mg2p reaches the comparably highest values (7–8 at.%) for this system. The buffering of the system seems to be even better than for DMEM in the incubator because of the additional buffering effect of albumin and other proteins contained in FCS (see Fig. 5). Thus it is probable that Mg is strongly incorporated in the layers due to continuing Mg dissolution. The clear Ca2p and P2p signals for all parameters hypothesize the formation of calcium phosphate layers on the Mg surfaces after immersion in cell culture medium. This assumption is supported by several studies, describing the formation of calcium phosphates on magnesium in SBF or cell culture medium [1,15–17]. To gain more information about the formed calcium phosphate compounds, Ca2p/P2p ratios were calculated for each immersion parameter (Fig. 6d). For the layer formation in DMEM at room temperature, all Ca2p/P2p ratios show values around 1.67, indicating the formation of hydroxyapatite (HA). The ratio of 1.67 can be deduced from the molecular formula of HA, Ca5(PO4)3OH [14,29]. Small deviations can be caused by substitution
V. Wagener, S. Virtanen / Materials Science and Engineering C 63 (2016) 341–351
347
Fig. 8. Results of EDX line scans after 1d immersion of Mg in DMEM and the corresponding SEM images showing the position of the lines; a) immersion of Mg in DMEM at room temperature; b) immersion of Mg in DMEM in the incubator (37 °C, 5% CO2); c) immersion of DMEM + 20% FCS in the incubator (37 °C, 5% CO2).
of Ca ions by Mg ions and/or phosphate groups by carbonate groups [7, 26]. For the immersion in DMEM in the incubator, a strong time dependence is observed. With increasing immersion time an almost linear drop of the Ca2p/P2p ratio from ca. 1.53 for 1d to 1.29 for 5d occurs. Such big deviations from the Ca2p/P2p ratio of HA cannot be exclusively explained by substitution of Ca and phosphate groups. It is assumed that the formed layers mainly consist of pre-stage calcium phosphates such as amorphous calcium phosphate (Ca/P: ~1.00) [30], tri-calcium phosphate (TCP, Ca/P: ~ 1.50) or octa-calcium phosphate (OCP, Ca/P: ~1.33) [29]. As well considerable is a mixture of different calcium phosphate compounds [29]. The decrease of the Ca/P ratio with immersion time may be due to progressing Ca substitution by Mg ions that are
available as well in the outer areas of the layer because of on-going corrosion in the buffered system. It has been described that in the beginning of immersion pre-stage calcium phosphate compounds are formed on a surface. With time and rising pH the transformation into HA can occur [7].Thus, especially the smaller increase in pH for the immersion in the incubator may explain why no HA is built. The addition of 20% FCS to DMEM in the incubator leads to a further decrease of the Ca/P ratio. For 1d and 3d, values are about 1.31, for 5d the Ca/P reaches minimal values (1.25) compared to all other parameters. In addition to the very surface sensitive XPS analysis, EDX line scans were conducted to gain information not only about the outer regions but also about the whole formed layers. Exemplary, 1d passivation in
348
V. Wagener, S. Virtanen / Materials Science and Engineering C 63 (2016) 341–351
Fig. 9. XRD spectra after one day immersion of Mg in DMEM at room temperature, in the incubator (37 °C, 5% CO2) and in DMEM + 20% FCS in the incubator (37 °C, 5% CO2).
DMEM at room temperature, in the incubator and in DMEM + 20% FCS in the incubator were analyzed. Fig. 8 depicts the results of the EDX line scans. Lines were placed in the way that information from substrate, interface and layer could be gained (see Fig. 8d, e and f). Analyzed were Mg, O, P and Ca signals. For none of the samples C and N could be detected. For all immersions the substrate shows high Mg signals, while signals for O and especially P and Ca approach zero. At the transition from substrate to layer a drop in Mg signals and an increase of oxygen occurs. For the immersion in DMEM at room temperature and in the incubator, as well a rise in P and Ca can be observed. While Mg is decreasing continuously from inner to outer layer regions, O, P and Ca increase more or less continuously over the layer thickness. After the addition of FCS to DMEM, neither Ca nor P can be detected in the substrate or in the layer. The results show that not only the outer regions of the formed layers consist mainly of Ca and phosphates. However, an increase of Ca and P content from inner to outer layer regions can be observed. At the same time the Mg concentration is decreasing. Due to the depth dependent variation in the layer composition, different calcium phosphate compounds may be found. However, caution has to be paid to EDX data due to limited resolution. In order to gain information about crystallinity of the formed layers, XRD spectra were recorded for the different DMEM immersion treatments for 1d (Fig. 9). All three curves show that no crystalline calcium phosphates were formed during immersion in DMEM. The peaks for RT, Inc and 20FCS are all caused by the crystalline Mg substrate. However, especially for the immersion in DMEM at room temperature an elevation in the
Fig. 10. FTIR spectra after one day immersion of Mg in DMEM at room temperature, in the incubator (37 °C, 5% CO2) and in DMEM + 20% FCS in the incubator (37 °C, 5% CO2).
range of 25° to 40° can be observed, indicating the existence of amorphous species. Most of the characteristic peaks for HA lie between 25° and 55° [31]. Accordingly, it is assumed that for the immersion in DMEM at room temperature amorphous HA is built. This can be due to incorporation of Mg in the layers. It has been reported that crystallization of HA and other calcium phosphates is hindered by the presence of Mg ions due to the substitution of Ca by Mg [7,32–36]. For the immersion at room temperature, a peak for Mg(OH)2 is visible which is not present for the other samples. It seems that magnesium hydroxide is only built in DMEM at room temperature, possibly due to the increase of pH values up to about 9 in this case. According to the Pourbaix diagram for Mg, Mg(OH)2 is only formed at pH values higher than 8.5 [37]. FTIR results for the three different 1d immersion treatments are shown in Fig. 10. For the immersion in DMEM at room temperature as well as for the immersion in cell culture medium with and without FCS in the incubator, phosphate groups and carbonate groups are detected, supporting the assumption that during immersion in cell culture medium calcium phosphate compounds with incorporated carbonates are deposited on the Mg surface. Positive ions such as Ca2+ and Mg2+ cannot be detected with FTIR. Influence of formed layers on the corrosion behavior of Mg. The influence of the layer formation in DMEM on corrosion resistance was investigated using EIS and I/E measurements. As a reference, electrochemical measurements for cp Mg without passivation in DMEM were conducted (Fig. 11). Fig. 12 depicts the results for the EIS measurements after immersion of Mg in DMEM. One representative Nyquist plot is shown for every sample. Figs. 11a and 12a–c show that the shapes of all curves are very similar, consisting of two semi-circles (compare as well Bode plots in the supporting information) whereat the first semi-circle is much smaller than the second one at lower frequencies. In general, two semi-circles in the Nyquist plot indicate the formation of a porous layer on the surface, with the second semi-circle representing the resistance and the capacity of the layer [38]. In addition, inductive loops occur for most of the curves, indicating dissolution of the Mg substrate under the formed layer [38]. For comparison of the results of the different immersion parameters, the total charge transfer resistance (Rct) was calculated by subtraction of Z′ values for the first intersection with the y-axis at high frequencies from Z′ values for the second intersection at low frequencies. Results are displayed in Fig. 12d. The orange hatched area marks the range of Rct values for cp Mg without DMEM treatment. The highest total charge transfer resistances and thus the best corrosion resistances are achieved by the immersion in DMEM at room temperature, whereat three days show maximum Rct values (~ 32 kΩ ∗ cm2). Compared to the immersion at room temperature, the layer formation in the incubator leads for all immersion times to a decrease in the corrosion resistance. With Rct values between 2 and 4 kΩ ∗cm2, the immersion in the incubator shows results that lie in the range of cp Mg without DMEM treatment. Noticeable are the rather good results for untreated Mg. This indicates the formation of a protective layer on the sample surface already during the electrochemical measurement in DMEM. The results for the polarization measurements are shown in Fig. 13. The layer formation at room temperature (Fig. 13a) shows for all immersion times similar curves. After the OCP a short increase in current density with a following plateau occurs, indicating passivity in this region. Between − 1.29 V and − 1.17 V, a breakdown of passivity (i.e., an abrupt increase of the anodic current density) can be observed for all samples. A difference for the immersion times occurs in the position of the OCP, in the level of the passive current density and the breakdown potential. Best results are again achieved by 3d immersion (most positive OCP, lowest passive current density, most positive breakdown potential). Immersion in DMEM in the incubator leads to similar polarization curves as the immersion at room temperature. However, curves are shifted to more negative potentials and to higher current densities, indicating a decrease in corrosion resistance. In addition, no real passive
V. Wagener, S. Virtanen / Materials Science and Engineering C 63 (2016) 341–351
349
Fig. 11. Representative Nyquist plot (a) and polarization curve (b) for cp Mg without passivation.
region can be observed for the DMEM passivation in the incubator. 1d and 3d immersion time show very similar curves, for 5d corrosion resistance is decreased. Addition of FCS during immersion leads to no significant changes in OCP. However, a passive region is again observed for three and five days of immersion. Nevertheless, passive current densities lie with values of about 40 μA/cm2 four to 20 times higher than passive current densities for DMEM immersion at room temperature. A comparison with the polarization measurement of cp Mg shows a very similar curve progression as after immersion at RT with a small plateau in the anodic region. This again indicates the formation of a protective layer during the electrochemical measurements in DMEM that take all in all about 1 h. The comparison of the corrosion current densities for the different treatments as determined from polarization curves is shown in
Fig. 13d. The orange hatched area marks the range of icorr for cp Mg. It is visible that the lowest icorr is achieved by 3d DMEM immersion at room temperature (b1 μA/cm2). In general, immersion in DMEM at room temperature leads to the best results. However, a significant increase of icorr occurs for five days. Here the deviations are comparably high. The layers formed in the incubator lie in the same region for 1d and 3d, considering standard deviations. While corrosion current densities increase to an overall maximum of about 9 μA/cm2 for DMEM, icorr decreases for DMEM + 20% FCS and reaches the level of 5d DMEM at room temperature. Nevertheless, DMEM immersion in the incubator does not lead to a decrease of icorr compared to cp Mg. Corrosion resistance even seems to be much worse for 5d in the incubator. In general, results of EIS and potentio-dynamic polarization show a good correlation. However, a positive effect of proteins during immersion
Fig. 12. EIS results of the DMEM immersion of Mg; a) representative Nyquist-plots after immersion for 1, 3 and 5 days in DMEM at room temperature; b) representative Nyquist-plots after immersion for 1, 3 and 5 days in DMEM in the incubator (37 °C, 5% CO2); c) representative Nyquist-plots after immersion for 1, 3 and 5 days in DMEM + 20% FCS in the incubator (37 °C, 5% CO2); d) mean total charge transfer resistances (Rct) of all immersion parameters with according standard deviations (n = 3). The orange hatched area marks the range of Rct for cp Mg.
350
V. Wagener, S. Virtanen / Materials Science and Engineering C 63 (2016) 341–351
Fig. 13. IE results of the immersion of Mg in DMEM; a) representative polarization curves after immersion for 1, 3 and 5 days in DMEM at room temperature; b) representative polarization curves after immersion for 1, 3 and 5 days in DMEM in the incubator (37 °C, 5% CO2); c) representative polarization curves after immersion for 1, 3 and 5 days in DMEM + 20% FCS in the incubator (37 °C, 5% CO2); d) mean corrosion current densities (icorr) of all immersion parameters with according standard deviations (n = 3). The orange hatched area marks the range of icorr for cp Mg.
is only visible for the polarization curves. For EIS no significant difference between the immersion in DMEM and in DMEM + 20% FCS can be detected. A correlation between layer thickness and corrosion resistance cannot be found. This may be due to crack formation that is increasing with increasing layer thickness and porosity. Main influencing factor on corrosion resistance seems to be the morphology of the formed layer. Best results concerning corrosion resistance are achieved by the immersion of Mg in DMEM at room temperature. The formed layers possess a compact outer region that may hinder ion transfer and thus reduce corrosion. The Ca/P ratio of the outer layer region that is very close to the value for HA indicates as well a strong reduction in Mg corrosion (i.e., less Mg2+ available for the layer formation). For the immersion in the incubator corrosion resistance decreases significantly, as already described. SEM images show that porous layers are formed and a completely compact outer layer does not exist. Thus, ion transfer is not reduced to the same degree as for the DMEM immersion at room temperature. The addition of 20% FCS to DMEM during immersion does not lead to an increase in charge transfer resistance. The formed layers are comparably thin and as well porous. However, a positive influence of proteins on the corrosion resistance can be observed for the polarization curves. Hence, proteins in FCS hinder the formation of a calcium phosphate layer on the Mg surface, but protein adsorption on the surface can at least partially replace the corrosion protective effect. 4. Conclusion It could be shown that immersion of magnesium in DMEM with varying parameters can lead to the formation of a corrosion layer consisting mainly of calcium phosphate compounds that can strongly enhance corrosion resistance of magnesium. Best results concerning corrosion protection are achieved by the immersion of Mg in DMEM
at room temperature; in this case, the corrosion rate is decreased by a factor of ca. 3–10 compared with bare Mg. Main influencing factor on the corrosion resistance is the layer morphology. The formation of a compact outer layer in DMEM at room temperature reduces corrosion significantly. Immersion of Mg in DMEM in the incubator does not improve corrosion resistance of Mg, in spite of a formation of thicker layers. However, due to stronger buffering in the incubator, Mg dissolution is not slowed down as strongly as for immersion at room temperature (with stronger pH increase), and completely porous layers are formed due to continuous hydrogen gas formation. The addition of FCS to DMEM shows similar results for electrochemical measurements as for passivation in the incubator without proteins. Surface analysis indicates that proteins are adsorbed on the surface but the formation of the calcium phosphate layer is hindered. Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.msec.2016.03.003. Acknowledgments The authors thank Martin Weiser, Sarah Höhn, Helga Hildebrand and Ulrike Marten for technical support with surface analysis and DFG (VI 350/7-1) for funding. References [1] M.P. Staiger, A.M. Pietak, J. Huadmai, G. Dias, Biomaterials 27 (2006) 1728–1734. [2] F. Witte, V. Kaese, H. Haferkamp, E. Switzer, A. Meyer-Lindenberg, C.J. Wirth, H. Windhagen, Biomaterials 26 (2005) 3557–3563. [3] R. Zeng, W. Dietzel, F. Witte, N. Hort, C. Blawert, Adv. Eng. Mater. 10 (2008) B3–B14. [4] F. Witte, J. Reifenrath, P.P. Muller, H.A. Crostack, J. Nellesen, F.W. Bach, D. Bormann, M. Rudert, Mater. Werkst. 37 (2006) 504–508. [5] L. Yang, N. Hort, D. Laipple, D. Höche, Y. Huang, K.U. Kainer, R. Willumeit, F. Feyerabend, Acta Biomater. 9 (2013) 8475–8487. [6] D. Williams, Med. Device Technol. 17 (2006) 9–10.
V. Wagener, S. Virtanen / Materials Science and Engineering C 63 (2016) 341–351 [7] S. Hiromoto, A. Yamamoto, Electrochim. Acta 54 (2009) 7085–7093. [8] D. MacConnell, Apatite: Its Crystal Chemistry, Mineralogy, Utilization, and Geologic and Biologic Occurrences, Springer-Verlag, New York, 1973. [9] J.X. Yang, Y.P. Jiao, F.Z. Cui, I.S. Lee, Q.S. Yin, Y. Zhang, Surf. Coat. Technol. 202 (2008) 5733–5736. [10] Y.F. Zheng, X.N. Gu, F. Witte, Mater. Sci. Eng. R 77 (2014) 1–34. [11] S. Hiromoto, M. Tomozawa, Surf. Coat. Technol. 205 (2011) 4711–4719. [12] Y.W. Song, D.Y. Shan, E.H. Han, Mater. Lett. 62 (2008) 3276–3279. [13] S. Shadanbaz, G.J. Dias, Acta Biomater. 8 (2012) 20–30. [14] O. Blind, L.H. Klein, B. Dailey, L. Jordan, Dent. Mater. 21 (2005) 1017–1024. [15] L.C. Li, J.C. Gao, Y. Wang, Surf. Coat. Technol. 185 (2004) 92–98. [16] H. Kuwahara, Y. Al-Abdullat, N. Mazaki, S. Tsutsumi, T. Aizawa, Mater. Trans. 42 (2001) 1317–1321. [17] R. Willumeit, J. Fischer, F. Feyerabend, N. Hort, U. Bismayer, S. Heidrich, B. Mihailova, Acta Biomater. 7 (2011) 2704–2715. [18] S. Virtanen, Mater. Sci. Eng. B Adv. 176 (2011) 1600–1608. [19] Y. Xin, T. Hu, P.K. Chu, Acta Biomater. 7 (2011) 1452–1459. [20] X.N. Gu, Y.F. Zheng, L.J. Chen, Biomed. Mater. 4 (2009) 065011. [21] J. Degner, F. Singer, L. Cordero, A.R. Boccaccini, S. Virtanen, Appl. Surf. Sci. 282 (2013) 264–270. [22] F. Singer, T. Distler, S. Virtanen, Int. J. Electrochem. Sci. 9 (2014) 7965–7976. [23] J.H. Scofield, J. Electron Spectrosc. Relat. Phenom. 8 (1976) 129–137. [24] R.F. Reilman, A. Msezane, S.T. Manson, J. Electron Spectrosc. Relat. Phenom. 8 (1976) 389–394.
[25] [26] [27] [28]
[29] [30] [31] [32] [33] [34] [35] [36] [37] [38]
351
G.L. Makar, J. Kruger, Int. Mater. Rev. 38 (1993) 138–153. R. Astala, M.J. Stott, Chem. Mater. 17 (2005) 4125–4133. H. El Feki, J.M. Savariault, A. Ben Salah, J. Alloys Compd. 287 (1999) 114–120. J.F. Moulder, J. Chastain, Handbook of X-ray Photoelectron Spectroscopy: A Reference Book of Standard Spectra for Identification and Interpretation of XPS Data, Physical Electronics Division, Perkin-Elmer Corporation, 1992. O. Suzuki, Acta Biomater. 6 (2010) 3379–3387. R. Rettig, S. Virtanen, J. Biomed. Mater. Res. A 88A (2009) 359–369. F. Witte, F. Feyerabend, P. Maier, J. Fischer, M. Stormer, C. Blawert, W. Dietzel, N. Hort, Biomaterials 28 (2007) 2163–2174. I.V. Fadeev, L.I. Shvorneva, S.M. Barinov, V.P. Orlovskii, Inorg. Mater. 39 (2003) 947–950. K.S. TenHuisen, P.W. Brown, J. Biomed. Mater. Res. 36 (1997) 306–314. A. Bigi, G. Falini, E. Foresti, M. Gazzano, A. Ripamonti, N. Roveri, J. Inorg. Biochem. 49 (1993) 69–78. V.C. West, Calcif. Tissue Res. 7 (1971) 212. M.J. Root, Calcif. Tissue Int. 47 (1990) 112–116. M. Pourbaix, Atlas of Electrochemical Equilibria in Aqueous Solutions, 1st English Ed, Pergamon Press, Oxford, New York, 1966. V.F. Lvovich, Impedance Spectroscopy: Applications to Electrochemical and Dielectric Phenomena, Wiley, Somerset, NJ, USA, 2012.
Contents lists available at ScienceDirect
Materials Science and Engineering C journal homepage: www.elsevier.com/locate/msec
Protective layer formation on magnesium in cell culture medium V. Wagener, S. Virtanen ⁎ Chair for Surface Science and Corrosion, Department Materials Science, University of Erlangen-Nuremberg, Martensstr. 7, 91058 Erlangen, Germany
a r t i c l e
i n f o
Article history: Received 27 October 2015 Received in revised form 18 February 2016 Accepted 1 March 2016 Available online 3 March 2016 Keywords: Magnesium Calcium phosphate coating Protein influence Corrosion protection electrochemical impedance spectroscopy
a b s t r a c t In the past, different studies showed that hydroxyapatite (HA) or similar calcium phosphates can be precipitated on Mg during immersion in simulated body fluids. However, at the same time, in most cases a dark grey or black layer is built under the white HA crystals. This layer seems to consist as well of calcium phosphates. Until now, neither the morphology nor its influence on Mg corrosion have been investigated in detail. In this work commercially pure magnesium (cp) was immersed in cell culture medium for one, three and five days at room temperature and in the incubator (37 °C, 5% CO2). In addition, the influence of proteins on the formation of a corrosion layer was investigated by adding 20% of fetal calf serum (FCS) to the cell culture medium in the incubator. In order to analyze the formed layers, SEM images of cross sections, X-ray Photoelectron Spectroscopy (XPS), Xray diffraction (XRD), Energy Dispersive X-ray Spectroscopy (EDX) and Fourier Transformed Infrared Spectroscopy (FTIR) measurements were carried out. Characterization of the corrosion behavior was achieved by electrochemical impedance spectroscopy (EIS) and by potentio-dynamic polarization in Dulbecco's Modified Eagle's Medium (DMEM) at 37 °C. Surface analysis showed that all formed layers consist mainly of amorphous calcium phosphate compounds. For the immersion at room temperature the Ca/P ratio indicates the formation of HA, while in the incubator probably pre-stages to HA are formed. The different immersion conditions lead to a variation in layer thicknesses. However, electrochemical characterization shows that the layer thickness does not influence the corrosion resistance of magnesium. The main influencing factor for the corrosion behavior is the layer morphology. Thus, immersion at room temperature leads to the highest corrosion protection due to the formation of a compact outer layer. Layers formed in the incubator show much worse performances due to completely porous structures. The existence of proteins in DMEM seems to hinder the formation of a corrosion layer. However, protein adsorption leads to similar results as concerns corrosion protection as the formed calcium phosphate layer. © 2016 Elsevier B.V. All rights reserved.
1. Introduction The scientific focus on magnesium as bio-degradable implant material in the orthopedic field is not only based on the non-toxicity of magnesium and its mechanical properties, but also on the positive influence of magnesium ions on bone formation [1–3]. Magnesium is a trace element in the human body and thus Mg ions exist abundantly in blood and several tissues. Indeed, half of the Mg ions in the body are stored in bone tissue [1]. Several studies showed that bone formation can be stimulated by the presence of Mg implants [4]. It is assumed that released Mg ions trigger the formation of calcium phosphate compounds on the implant surface that may be pre-stages to hydroxyapatite (HA), which is a main component of human bone [5,6]. The pH increase due to Mg dissolution favors the formation of calcium phosphate as well [7]. ⁎ Corresponding author. E-mail address: [email protected] (S. Virtanen).
http://dx.doi.org/10.1016/j.msec.2016.03.003 0928-4931/© 2016 Elsevier B.V. All rights reserved.
It is as well reported that calcium phosphates and especially HA act as good corrosion protective coatings on magnesium [7–10]. Thus, in several studies HA was deposited on Mg or on Mg alloys in order to increase corrosion resistance for biomedical application [9, 11–13]. For the application as bone implant not only the formation of calcium phosphate in vivo but also the pre-coating with HA is of interest. Blind et al. [14] reported improved osseo-integration due to HA coatings, for example. Different studies showed that precipitation of calcium phosphate compounds on magnesium in simulated body fluids or cell culture media can occur [1,15–20]. At the same time, however, a dark grey or black layer is formed under the white calcium phosphate precipitations in most cases [17,21,22]. Degner et al. propose a corrosion protective effect of this black layer [21]. Willumeit et al. [17] found out that those layers as well mainly consist of calcium phosphate compounds. However, the morphology of the layers was not investigated in detail. In this study, magnesium was immersed in cell culture medium under different conditions. Immersion time, temperature and CO2 concentration
342
V. Wagener, S. Virtanen / Materials Science and Engineering C 63 (2016) 341–351
Fig. 1. SEM images after immersion of Mg in DMEM at room temperature; a)–c) immersion for one day; d)–f) immersion for three days; g)–i) immersion for five days.
were varied. The formed layers were analyzed concerning their thickness, morphology and composition and corrosion resistance. In addition, the influence of proteins in cell culture medium on the formation of a corrosion protective layer was investigated. 2. Materials and methods 2.1. Sample preparation Samples were cut from a cp magnesium rod (25.4 mm diameter, 99.9% purity, Chempur Feinchemikalien GmbH & Co. KG), and afterwards deburred and ground with SiC abrasive paper up to a grid size of 1200. After sonication in ethanol and drying, samples were immersed in ca. 50 ml of cell culture medium for one, three and five days at room temperature and in the incubator at 37 °C (5% CO2, ~95% humidity). As cell culture medium, Dulbecco's Modified Eagle's Medium with 1 g glucose and 3.7 g NaHCO3 (DMEM, Biochrom AG) was used for the immersion at room temperature. For the immersion in the incubator 1% Penicillin-Streptomycin-Glutamine (PSG, Sigma) was added to the described DMEM, in order to prevent growth of microorganisms. For a third testing row, 20% of fetal calf serum (FCS, Gibco) was added to the cell culture medium in the incubator, in order to investigate the influence of serum proteins on the layer formation on magnesium in DMEM. After immersion in cell culture medium for the desired time, samples were rinsed with water and dried with nitrogen. 2.2. pH measurements The pH of the cell culture medium was measured before and after immersion for all parameters. As reference, pH was measured for DMEM at room temperature, DMEM (+ PSG) in the incubator and
DMEM (+PSG) + 20% FCS in the incubator without immersed samples after one, three and five days. 2.3. SEM Cross sections were prepared with help of an ion mill (IM4000, Hitachi), using an acceleration voltage of 6 V and a discharging voltage of 1.5 kV. Samples were cut in the “cross milling” mode with three repetitions per minute with an angle of ± 30°. Afterwards, the cross sections were characterized with a scanning electron microscope (FESEM S4800, Hitachi), using an acceleration voltage between 3 kV and 10 kV. The working distance was varied between 4 and 6 mm. 2.4. XPS XPS measurements were carried out with a high-resolution X-ray photoelectron spectrometer (PHI 5600 USA) using aluminum Kα radiation (1486.6 eV, 300 W) for excitation. The binding energy of the target elements was determined at a pass energy of 23.5 eV with a resolution of b0.4 eV (values measured every 0.2 eV for the high resolution spectra and 0.8 for survey) and a takeoff angle of 45° with respect to the surface normal. The binding energy of the C1s signal was used to correct the spectra for charging. The background was subtracted using the Shirley method in all spectra. To obtain the molar fractions of each species, the peak areas of the measured XPS spectra were corrected with the photoionization cross sections of Scofield [23] σ and the asymmetry parameter β (orbital geometry) [24], which are contained in the sensitivity factors of the acquisition software (MultiPakV6.1A, 99 June 16, copyright Physical Electronics Inc.,1994–1999). For each immersion parameter a survey spectrum and high resolution spectra for C1s, O1s, N1s, Ca2p, P2p and Mg2p signals were measured. C1s narrow scans for one day immersion were fitted with Origin (Gaussian multi-peak fitting).
V. Wagener, S. Virtanen / Materials Science and Engineering C 63 (2016) 341–351
343
Fig. 2. SEM images after immersion of Mg in DMEM in the incubator (37 °C, 5% CO2); a)–c) immersion for one day; d)–f) immersion for three days; g)–i) immersion for five days.
2.5. XRD X-ray diffraction (XRD) spectra were recorded for the different immersion parameters for one day, in order to gain information concerning crystallinity of the formed layers. Therefore, an X-ray diffractometer (X'PERT PW 3040 MPD, Panalytical) with a monochromatic Cu-Kα radiation (λ = 1.54 Å) was used, the divergence slit was put to 0.19 mm. Diffraction patterns were calculated from 2θ (10°–80°) for every sample, with a step width of 0.03° and an incident angle of 1°.
Potential (OCP) was recorded for 30 min to guarantee a rather stable potential for the EIS measurements (not shown). Afterwards, the impedance was measured at the particular OCP of the samples in the range of 100 kHz to 10 mHz with an excitation amplitude of ±10 mV. In addition, polarization curves were measured after EIS. Therefore, voltage was varied within a range of −300 mV relative to the particular OCP and 0 V with a scan rate of 3 mV/s and with a 10 mA current limit in the anodic region. 3. Results and discussion
2.6. FTIR Fourier Transformed Infrared Spectroscopy (FTIR) was conducted with an infrared spectrometer (Thermo Scientific Nicolet 6700/Smart iTR) with a diamond reference crystal for the immersion parameters for one day. Measurements took place in a wave number range of 500–4000 cm−1 with a penetration depth of 2 μm. 2.7. Electrochemistry Electrochemical impedance spectroscopy (EIS) was used to characterize the different Mg samples concerning corrosion resistance. All measurements were conducted using an electrochemical workstation “IM6eX”, a potentiostat “XPot” and the corresponding “Thales”-Software (Zahner-Elektrik GmbH & Co. KG, Kronach, Germany). The experiments were carried out in an electrochemical cell with a three electrode set-up, in which the particular Mg samples acted as working electrodes. As counter electrode a platinum sheet and as reference an Ag/AgCl electrode with 3 M KCl were used. As electrolyte 100 ml of DMEM was chosen, measuring temperature was kept constant at 37 °C (±1 °C). The surface exposed to the electrolyte was 0.7854 cm2. The Open Circuit
3.1. Layer morphology Cross sections were prepared and analyzed by SEM in order to gain detailed information on morphology and thickness of the layers formed during immersion in DMEM under different conditions. Fig. 1 depicts the SEM images of the cross sections for the immersion of magnesium in DMEM at room temperature. All three immersion times show the formation of layers with similar morphology. Layers can be divided into two areas with different structures. While the inner layer is porous, the outer regions show compact structures (see Fig. 1b, e and h). In general, a decrease of porosity in the inner layer from substrate to the outer compact region can be observed. The high porosity near the substrate may be due to hydrogen evolution caused by initial strong magnesium dissolution (see Eq. (1)). Mg þ 2H2 O→Mg2þ þ 2OH− þ H2
ð1Þ
The formation of gas bubbles leads to the formation of a porous initial layer. With the growth of the layer during immersion, Mg corrosion slows down due to isolating effects of the layer and a pH increase (compare Fig. 5) [25], resulting in a decreased hydrogen evolution and thus in
344
V. Wagener, S. Virtanen / Materials Science and Engineering C 63 (2016) 341–351
Fig. 3. SEM images after immersion of Mg in DMEM + 20% FCS in the incubator (37 °C, 5% CO2); a)–c) immersion for one day; d)–f) immersion for three days; g)–i) immersion for five days.
a decrease of porosity. The overview pictures (Fig. 1a, d and g) show crack formation for all immersion times. Cracks extend to the whole layer thickness and may be caused by internal stresses during immersion in DMEM or during drying. For 5d immersion, the formation of structures beneath the layer is found (see Fig. 1i), indicating corrosion. These structures seem to begin directly at the substrate-layer interface and spread into the substrate. Fig. 2 depicts the SEM images for layer formation on magnesium in DMEM in the incubator. Compared to DMEM immersion at room
Fig. 4. Average thicknesses of layers built after immersion of Mg in DMEM at room temperature, in DMEM in the incubator (37 °C, 5% CO2) and in DMEM + 20% FCS in the incubator (37 °C, 5% CO2) with standard deviations (n = 4).
temperature, the morphology of the formed layers differs for the different immersion times. The immersion for one day shows a porous structure with no gradient in pore size or density over the layer thickness. No compact outer layer can be observed (see Fig. 2b and d). In comparison, the 3d and 5d immersion show again different regions with a variation in pore density (Fig. 2e and h). For 3d, pore density decreases from inside to the outer areas. Nevertheless, no completely compact outer layer is formed, as can be seen in Fig. 2e. 5d immersion shows as well areas with different pore density. However, again no completely compact outer layer is
Fig. 5. Development of pH during immersion of magnesium in DMEM at room temperature (RT) in the incubator at 37 °C (Inc) and in DMEM + 20% FCS in the incubator (20FCS). Reference measurements of medium without samples showed no significant differences in pH values.
V. Wagener, S. Virtanen / Materials Science and Engineering C 63 (2016) 341–351
345
Fig. 6. XPS results after immersion of Mg in DMEM for 1, 3 and 5 days; a) atomic concentrations of several signals after immersion in DMEM at room temperature; b) atomic concentrations of several signals after immersion in DMEM in the incubator (37 °C, 5% CO2); c) atomic concentrations of several signals after immersion in DMEM + 20% FCS in the incubator (37 °C, 5% CO2); d) Ca2p/P2p ratios for the DMEM immersion with the different parameters.
formed (Fig. 2i). The reason for the missing compact outer layer for the immersion in DMEM in the incubator may be the rather small increase in pH due to the better buffered system caused by the high CO2 concentration in the environment (see Fig. 5). Thus, magnesium dissolution is not slowed down due to a rise in pH as it may be the case for the immersion at room temperature. Crack formation seems to be more distinctive than for the immersion in DMEM at room temperature, especially after three and five days, where layer thicknesses increase (Fig. 2d, g and h). This may be explained by increasing internal stresses in the thicker layers. The SEM images of the cross sections for the immersion of Mg in DMEM + 20% FCS in the incubator are shown in Fig. 3. Compared to the immersion in DMEM without proteins, rather thin layers are formed and the border between substrate and layer seems to be rougher. While three days immersion shows a similar pore structure as the layers built in DMEM without proteins (see Fig. 3e, d and f), the formed structures after 1d and 5d passivation seem to be more compact. Regularly cracks were only detected after three days, where layers reached a thickness of about 1 μm. This confirms the assumption that cracks originate from internal stresses for sufficient thick layers. Layer thicknesses were taken from the particular cross section SEM images. Mean values were calculated from four different positions in the layer on each sample. The results can be seen in Fig. 4. For the immersion in DMEM at room temperature a slight decrease in layer thickness over time is observed. The maximum values are reached for 1d (3.5 μm). The low standard deviations indicate a homogeneous layer formation. In comparison, the layers formed in DMEM in the incubator show a strong increase in average thickness with immersion time. The thickest layers are achieved after five days (up to 24 μm). Nevertheless, deviations increase strongly for longer immersion times indicating
inhomogeneous layer growth, as it can be seen in Fig. 2g. The addition of 20% FCS to DMEM in the incubator causes a significant drop in layer thicknesses to values of 1 μm and lower. It is assumed that the proteins in the cell culture medium hinder the formation of the corrosion layer. Layer growth at room temperature is for all immersion times very similar and seems to be self-limited. This may be due to the formation of a compact outer layer. Further surface analysis indicates that the formed layers mainly consist of calcium phosphates (see Section 3.2). It is reported that the formation of calcium phosphate layers is triggered by the release of Mg ions and favored by an increase of pH values [7]. The reduced corrosion during layer growth and pH increase as well as the formation of a compact outer layer (compare Fig. 1) decreases the release of Mg ions strongly. As a consequence, calcium phosphate layers do not extend further after the compact layer is built. SEM images show that this stage is already reached after one day. Another indicator for layer growth limitation is the development of pH in the only weakly buffered electrolyte. As Fig. 5 shows, pH values increase within the first day of immersion for the passivation at room temperature. For three and five days, pH values stay more or less on the same level, pointing out that Mg corrosion reaches a plateau. For the immersion in DMEM in the incubator, a more or less steady increase of layer thickness over time occurs. SEM images show that porous layers without completely compact outer regions are formed in the incubator. Thus a limitation like it is observed for the immersion at room temperature does not exist. The pH values show only a small rise due to the better buffering in the incubator. It is assumed that corrosion and thus Mg release occur during the whole immersion time. The large variation in layer thickness may be caused by impurities in the Mg substrate or slightly different pH values in the cell culture medium. As layer formation seems not to be limited within the investigated period
346
V. Wagener, S. Virtanen / Materials Science and Engineering C 63 (2016) 341–351
Fig. 7. C1s narrow scans with fitted curves for 1d immersion in a) DMEM at room temperature, b) in DMEM in the incubator and c) in DMEM + 20% FCS in the incubator.
due to buffering, slightly different immersion conditions may have an influence on the growth. Layer formation in DMEM + 20% FCS seems again to underlie a limitation. As already mentioned, proteins seem to hinder the formation of a calcium phosphate layer.
3.2. Layer composition In order to analyze the composition of the formed layers, XPS spectra and narrow scans were measured. Atomic concentrations for C1s, N1s, O1s, Ca2p, P2p and Mg2p are shown in Fig. 6. In addition, C1s narrow
scans for 1d immersion under the three different conditions can be found in Fig. 7. The layers formed on Mg in DMEM at room temperature (Fig. 6a) show for all immersion times' similar values for the selected signals. The largest variations for the different times can be observed for C1sand O1s-signals. While C1s decreases significantly after three days (~ 20.2 at.%), O1s reaches maximum values (~ 55.5 at.%). N1s signals show values b1.5 at.% for all immersion times with peaks being overlaid by noise. This leads to the assumption that no nitrogen compounds are included in the outer layer (as detected by XPS). Ca2p signals lie around 12 at.%, P2p signals around 7 at.% for all immersion times. Mg2p signals are with values smaller than 3 at.% comparably low, showing that magnesium is not included as a main component in the formed (outer) layer. The high O1s, Ca2p and P2p signals indicate the formation of a calcium phosphate layer. C1s signals may be caused by atmospheric contamination. However, the relatively high values hypothesize the incorporation of carbonate ions in the calcium phosphate layer [26, 27]. The analysis of C1s narrow scan (Fig. 7a) shows that indeed most of the carbon signal seems to originate from atmospheric contamination (C–Hx) or carboxyl or carbonate groups [28]. As for the DMEM immersion at room temperature, the DMEM immersion in the incubator shows no significant variations of the selected signals for the different immersion times (Fig. 6b). Compared to the immersion at room temperature, C1s signals decrease slightly, while O1s signals show an increase. N1s signals show again values b1.5 at.% leading as well to the assumption that no nitrogen compounds are included in the (outer) layer. Ca2p and P2p signals lie in the same range as for the immersion in DMEM in the incubator. However, Mg2p signals show an increase (6 at.%). The higher concentration of Mg in the outer layer may be due to higher Mg dissolution because of the better buffering in the incubator, leading to an incorporation of Mg in the whole layer. The analysis of the C1s narrow scan (Fig. 7b) shows similar results as for the immersion in DMEM at room temperature. Again, the carbon signal is mainly caused by contamination and carboxyl and carbonate groups. However, relative to the peak at about 284.5 eV (C–Hx) the peak at about 289.4 eV (O = C–O, CO2– 3 ) is higher. This may indicate that more carbonate groups are incorporated in the layer due to the significantly higher CO2 concentration in the incubator. The addition of 20% FCS to the cell culture medium in the incubator leads to a significantly different chemical composition of the outer layer as observed by XPS: A strong increase of C1s and N1s signals for all immersion times is seen, while O1s, Ca2p and P2p signals show a decrease (Fig. 6c). The increase of C1s and N1s signals indicates adsorption of proteins on the surface, whereas the lower amounts of Ca and P detected verify the assumption that proteins hinder the formation of calcium phosphate layers on the surface. The C1s narrow scan shows as well that proteins are adsorbed on the surface. Compared to the immersion of Mg in DMEM without FCS, the shape of the overall C1s signal changes due to an additional distinctive peak caused by C–N bindings (Fig. 7c). Mg2p reaches the comparably highest values (7–8 at.%) for this system. The buffering of the system seems to be even better than for DMEM in the incubator because of the additional buffering effect of albumin and other proteins contained in FCS (see Fig. 5). Thus it is probable that Mg is strongly incorporated in the layers due to continuing Mg dissolution. The clear Ca2p and P2p signals for all parameters hypothesize the formation of calcium phosphate layers on the Mg surfaces after immersion in cell culture medium. This assumption is supported by several studies, describing the formation of calcium phosphates on magnesium in SBF or cell culture medium [1,15–17]. To gain more information about the formed calcium phosphate compounds, Ca2p/P2p ratios were calculated for each immersion parameter (Fig. 6d). For the layer formation in DMEM at room temperature, all Ca2p/P2p ratios show values around 1.67, indicating the formation of hydroxyapatite (HA). The ratio of 1.67 can be deduced from the molecular formula of HA, Ca5(PO4)3OH [14,29]. Small deviations can be caused by substitution
V. Wagener, S. Virtanen / Materials Science and Engineering C 63 (2016) 341–351
347
Fig. 8. Results of EDX line scans after 1d immersion of Mg in DMEM and the corresponding SEM images showing the position of the lines; a) immersion of Mg in DMEM at room temperature; b) immersion of Mg in DMEM in the incubator (37 °C, 5% CO2); c) immersion of DMEM + 20% FCS in the incubator (37 °C, 5% CO2).
of Ca ions by Mg ions and/or phosphate groups by carbonate groups [7, 26]. For the immersion in DMEM in the incubator, a strong time dependence is observed. With increasing immersion time an almost linear drop of the Ca2p/P2p ratio from ca. 1.53 for 1d to 1.29 for 5d occurs. Such big deviations from the Ca2p/P2p ratio of HA cannot be exclusively explained by substitution of Ca and phosphate groups. It is assumed that the formed layers mainly consist of pre-stage calcium phosphates such as amorphous calcium phosphate (Ca/P: ~1.00) [30], tri-calcium phosphate (TCP, Ca/P: ~ 1.50) or octa-calcium phosphate (OCP, Ca/P: ~1.33) [29]. As well considerable is a mixture of different calcium phosphate compounds [29]. The decrease of the Ca/P ratio with immersion time may be due to progressing Ca substitution by Mg ions that are
available as well in the outer areas of the layer because of on-going corrosion in the buffered system. It has been described that in the beginning of immersion pre-stage calcium phosphate compounds are formed on a surface. With time and rising pH the transformation into HA can occur [7].Thus, especially the smaller increase in pH for the immersion in the incubator may explain why no HA is built. The addition of 20% FCS to DMEM in the incubator leads to a further decrease of the Ca/P ratio. For 1d and 3d, values are about 1.31, for 5d the Ca/P reaches minimal values (1.25) compared to all other parameters. In addition to the very surface sensitive XPS analysis, EDX line scans were conducted to gain information not only about the outer regions but also about the whole formed layers. Exemplary, 1d passivation in
348
V. Wagener, S. Virtanen / Materials Science and Engineering C 63 (2016) 341–351
Fig. 9. XRD spectra after one day immersion of Mg in DMEM at room temperature, in the incubator (37 °C, 5% CO2) and in DMEM + 20% FCS in the incubator (37 °C, 5% CO2).
DMEM at room temperature, in the incubator and in DMEM + 20% FCS in the incubator were analyzed. Fig. 8 depicts the results of the EDX line scans. Lines were placed in the way that information from substrate, interface and layer could be gained (see Fig. 8d, e and f). Analyzed were Mg, O, P and Ca signals. For none of the samples C and N could be detected. For all immersions the substrate shows high Mg signals, while signals for O and especially P and Ca approach zero. At the transition from substrate to layer a drop in Mg signals and an increase of oxygen occurs. For the immersion in DMEM at room temperature and in the incubator, as well a rise in P and Ca can be observed. While Mg is decreasing continuously from inner to outer layer regions, O, P and Ca increase more or less continuously over the layer thickness. After the addition of FCS to DMEM, neither Ca nor P can be detected in the substrate or in the layer. The results show that not only the outer regions of the formed layers consist mainly of Ca and phosphates. However, an increase of Ca and P content from inner to outer layer regions can be observed. At the same time the Mg concentration is decreasing. Due to the depth dependent variation in the layer composition, different calcium phosphate compounds may be found. However, caution has to be paid to EDX data due to limited resolution. In order to gain information about crystallinity of the formed layers, XRD spectra were recorded for the different DMEM immersion treatments for 1d (Fig. 9). All three curves show that no crystalline calcium phosphates were formed during immersion in DMEM. The peaks for RT, Inc and 20FCS are all caused by the crystalline Mg substrate. However, especially for the immersion in DMEM at room temperature an elevation in the
Fig. 10. FTIR spectra after one day immersion of Mg in DMEM at room temperature, in the incubator (37 °C, 5% CO2) and in DMEM + 20% FCS in the incubator (37 °C, 5% CO2).
range of 25° to 40° can be observed, indicating the existence of amorphous species. Most of the characteristic peaks for HA lie between 25° and 55° [31]. Accordingly, it is assumed that for the immersion in DMEM at room temperature amorphous HA is built. This can be due to incorporation of Mg in the layers. It has been reported that crystallization of HA and other calcium phosphates is hindered by the presence of Mg ions due to the substitution of Ca by Mg [7,32–36]. For the immersion at room temperature, a peak for Mg(OH)2 is visible which is not present for the other samples. It seems that magnesium hydroxide is only built in DMEM at room temperature, possibly due to the increase of pH values up to about 9 in this case. According to the Pourbaix diagram for Mg, Mg(OH)2 is only formed at pH values higher than 8.5 [37]. FTIR results for the three different 1d immersion treatments are shown in Fig. 10. For the immersion in DMEM at room temperature as well as for the immersion in cell culture medium with and without FCS in the incubator, phosphate groups and carbonate groups are detected, supporting the assumption that during immersion in cell culture medium calcium phosphate compounds with incorporated carbonates are deposited on the Mg surface. Positive ions such as Ca2+ and Mg2+ cannot be detected with FTIR. Influence of formed layers on the corrosion behavior of Mg. The influence of the layer formation in DMEM on corrosion resistance was investigated using EIS and I/E measurements. As a reference, electrochemical measurements for cp Mg without passivation in DMEM were conducted (Fig. 11). Fig. 12 depicts the results for the EIS measurements after immersion of Mg in DMEM. One representative Nyquist plot is shown for every sample. Figs. 11a and 12a–c show that the shapes of all curves are very similar, consisting of two semi-circles (compare as well Bode plots in the supporting information) whereat the first semi-circle is much smaller than the second one at lower frequencies. In general, two semi-circles in the Nyquist plot indicate the formation of a porous layer on the surface, with the second semi-circle representing the resistance and the capacity of the layer [38]. In addition, inductive loops occur for most of the curves, indicating dissolution of the Mg substrate under the formed layer [38]. For comparison of the results of the different immersion parameters, the total charge transfer resistance (Rct) was calculated by subtraction of Z′ values for the first intersection with the y-axis at high frequencies from Z′ values for the second intersection at low frequencies. Results are displayed in Fig. 12d. The orange hatched area marks the range of Rct values for cp Mg without DMEM treatment. The highest total charge transfer resistances and thus the best corrosion resistances are achieved by the immersion in DMEM at room temperature, whereat three days show maximum Rct values (~ 32 kΩ ∗ cm2). Compared to the immersion at room temperature, the layer formation in the incubator leads for all immersion times to a decrease in the corrosion resistance. With Rct values between 2 and 4 kΩ ∗cm2, the immersion in the incubator shows results that lie in the range of cp Mg without DMEM treatment. Noticeable are the rather good results for untreated Mg. This indicates the formation of a protective layer on the sample surface already during the electrochemical measurement in DMEM. The results for the polarization measurements are shown in Fig. 13. The layer formation at room temperature (Fig. 13a) shows for all immersion times similar curves. After the OCP a short increase in current density with a following plateau occurs, indicating passivity in this region. Between − 1.29 V and − 1.17 V, a breakdown of passivity (i.e., an abrupt increase of the anodic current density) can be observed for all samples. A difference for the immersion times occurs in the position of the OCP, in the level of the passive current density and the breakdown potential. Best results are again achieved by 3d immersion (most positive OCP, lowest passive current density, most positive breakdown potential). Immersion in DMEM in the incubator leads to similar polarization curves as the immersion at room temperature. However, curves are shifted to more negative potentials and to higher current densities, indicating a decrease in corrosion resistance. In addition, no real passive
V. Wagener, S. Virtanen / Materials Science and Engineering C 63 (2016) 341–351
349
Fig. 11. Representative Nyquist plot (a) and polarization curve (b) for cp Mg without passivation.
region can be observed for the DMEM passivation in the incubator. 1d and 3d immersion time show very similar curves, for 5d corrosion resistance is decreased. Addition of FCS during immersion leads to no significant changes in OCP. However, a passive region is again observed for three and five days of immersion. Nevertheless, passive current densities lie with values of about 40 μA/cm2 four to 20 times higher than passive current densities for DMEM immersion at room temperature. A comparison with the polarization measurement of cp Mg shows a very similar curve progression as after immersion at RT with a small plateau in the anodic region. This again indicates the formation of a protective layer during the electrochemical measurements in DMEM that take all in all about 1 h. The comparison of the corrosion current densities for the different treatments as determined from polarization curves is shown in
Fig. 13d. The orange hatched area marks the range of icorr for cp Mg. It is visible that the lowest icorr is achieved by 3d DMEM immersion at room temperature (b1 μA/cm2). In general, immersion in DMEM at room temperature leads to the best results. However, a significant increase of icorr occurs for five days. Here the deviations are comparably high. The layers formed in the incubator lie in the same region for 1d and 3d, considering standard deviations. While corrosion current densities increase to an overall maximum of about 9 μA/cm2 for DMEM, icorr decreases for DMEM + 20% FCS and reaches the level of 5d DMEM at room temperature. Nevertheless, DMEM immersion in the incubator does not lead to a decrease of icorr compared to cp Mg. Corrosion resistance even seems to be much worse for 5d in the incubator. In general, results of EIS and potentio-dynamic polarization show a good correlation. However, a positive effect of proteins during immersion
Fig. 12. EIS results of the DMEM immersion of Mg; a) representative Nyquist-plots after immersion for 1, 3 and 5 days in DMEM at room temperature; b) representative Nyquist-plots after immersion for 1, 3 and 5 days in DMEM in the incubator (37 °C, 5% CO2); c) representative Nyquist-plots after immersion for 1, 3 and 5 days in DMEM + 20% FCS in the incubator (37 °C, 5% CO2); d) mean total charge transfer resistances (Rct) of all immersion parameters with according standard deviations (n = 3). The orange hatched area marks the range of Rct for cp Mg.
350
V. Wagener, S. Virtanen / Materials Science and Engineering C 63 (2016) 341–351
Fig. 13. IE results of the immersion of Mg in DMEM; a) representative polarization curves after immersion for 1, 3 and 5 days in DMEM at room temperature; b) representative polarization curves after immersion for 1, 3 and 5 days in DMEM in the incubator (37 °C, 5% CO2); c) representative polarization curves after immersion for 1, 3 and 5 days in DMEM + 20% FCS in the incubator (37 °C, 5% CO2); d) mean corrosion current densities (icorr) of all immersion parameters with according standard deviations (n = 3). The orange hatched area marks the range of icorr for cp Mg.
is only visible for the polarization curves. For EIS no significant difference between the immersion in DMEM and in DMEM + 20% FCS can be detected. A correlation between layer thickness and corrosion resistance cannot be found. This may be due to crack formation that is increasing with increasing layer thickness and porosity. Main influencing factor on corrosion resistance seems to be the morphology of the formed layer. Best results concerning corrosion resistance are achieved by the immersion of Mg in DMEM at room temperature. The formed layers possess a compact outer region that may hinder ion transfer and thus reduce corrosion. The Ca/P ratio of the outer layer region that is very close to the value for HA indicates as well a strong reduction in Mg corrosion (i.e., less Mg2+ available for the layer formation). For the immersion in the incubator corrosion resistance decreases significantly, as already described. SEM images show that porous layers are formed and a completely compact outer layer does not exist. Thus, ion transfer is not reduced to the same degree as for the DMEM immersion at room temperature. The addition of 20% FCS to DMEM during immersion does not lead to an increase in charge transfer resistance. The formed layers are comparably thin and as well porous. However, a positive influence of proteins on the corrosion resistance can be observed for the polarization curves. Hence, proteins in FCS hinder the formation of a calcium phosphate layer on the Mg surface, but protein adsorption on the surface can at least partially replace the corrosion protective effect. 4. Conclusion It could be shown that immersion of magnesium in DMEM with varying parameters can lead to the formation of a corrosion layer consisting mainly of calcium phosphate compounds that can strongly enhance corrosion resistance of magnesium. Best results concerning corrosion protection are achieved by the immersion of Mg in DMEM
at room temperature; in this case, the corrosion rate is decreased by a factor of ca. 3–10 compared with bare Mg. Main influencing factor on the corrosion resistance is the layer morphology. The formation of a compact outer layer in DMEM at room temperature reduces corrosion significantly. Immersion of Mg in DMEM in the incubator does not improve corrosion resistance of Mg, in spite of a formation of thicker layers. However, due to stronger buffering in the incubator, Mg dissolution is not slowed down as strongly as for immersion at room temperature (with stronger pH increase), and completely porous layers are formed due to continuous hydrogen gas formation. The addition of FCS to DMEM shows similar results for electrochemical measurements as for passivation in the incubator without proteins. Surface analysis indicates that proteins are adsorbed on the surface but the formation of the calcium phosphate layer is hindered. Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.msec.2016.03.003. Acknowledgments The authors thank Martin Weiser, Sarah Höhn, Helga Hildebrand and Ulrike Marten for technical support with surface analysis and DFG (VI 350/7-1) for funding. References [1] M.P. Staiger, A.M. Pietak, J. Huadmai, G. Dias, Biomaterials 27 (2006) 1728–1734. [2] F. Witte, V. Kaese, H. Haferkamp, E. Switzer, A. Meyer-Lindenberg, C.J. Wirth, H. Windhagen, Biomaterials 26 (2005) 3557–3563. [3] R. Zeng, W. Dietzel, F. Witte, N. Hort, C. Blawert, Adv. Eng. Mater. 10 (2008) B3–B14. [4] F. Witte, J. Reifenrath, P.P. Muller, H.A. Crostack, J. Nellesen, F.W. Bach, D. Bormann, M. Rudert, Mater. Werkst. 37 (2006) 504–508. [5] L. Yang, N. Hort, D. Laipple, D. Höche, Y. Huang, K.U. Kainer, R. Willumeit, F. Feyerabend, Acta Biomater. 9 (2013) 8475–8487. [6] D. Williams, Med. Device Technol. 17 (2006) 9–10.
V. Wagener, S. Virtanen / Materials Science and Engineering C 63 (2016) 341–351 [7] S. Hiromoto, A. Yamamoto, Electrochim. Acta 54 (2009) 7085–7093. [8] D. MacConnell, Apatite: Its Crystal Chemistry, Mineralogy, Utilization, and Geologic and Biologic Occurrences, Springer-Verlag, New York, 1973. [9] J.X. Yang, Y.P. Jiao, F.Z. Cui, I.S. Lee, Q.S. Yin, Y. Zhang, Surf. Coat. Technol. 202 (2008) 5733–5736. [10] Y.F. Zheng, X.N. Gu, F. Witte, Mater. Sci. Eng. R 77 (2014) 1–34. [11] S. Hiromoto, M. Tomozawa, Surf. Coat. Technol. 205 (2011) 4711–4719. [12] Y.W. Song, D.Y. Shan, E.H. Han, Mater. Lett. 62 (2008) 3276–3279. [13] S. Shadanbaz, G.J. Dias, Acta Biomater. 8 (2012) 20–30. [14] O. Blind, L.H. Klein, B. Dailey, L. Jordan, Dent. Mater. 21 (2005) 1017–1024. [15] L.C. Li, J.C. Gao, Y. Wang, Surf. Coat. Technol. 185 (2004) 92–98. [16] H. Kuwahara, Y. Al-Abdullat, N. Mazaki, S. Tsutsumi, T. Aizawa, Mater. Trans. 42 (2001) 1317–1321. [17] R. Willumeit, J. Fischer, F. Feyerabend, N. Hort, U. Bismayer, S. Heidrich, B. Mihailova, Acta Biomater. 7 (2011) 2704–2715. [18] S. Virtanen, Mater. Sci. Eng. B Adv. 176 (2011) 1600–1608. [19] Y. Xin, T. Hu, P.K. Chu, Acta Biomater. 7 (2011) 1452–1459. [20] X.N. Gu, Y.F. Zheng, L.J. Chen, Biomed. Mater. 4 (2009) 065011. [21] J. Degner, F. Singer, L. Cordero, A.R. Boccaccini, S. Virtanen, Appl. Surf. Sci. 282 (2013) 264–270. [22] F. Singer, T. Distler, S. Virtanen, Int. J. Electrochem. Sci. 9 (2014) 7965–7976. [23] J.H. Scofield, J. Electron Spectrosc. Relat. Phenom. 8 (1976) 129–137. [24] R.F. Reilman, A. Msezane, S.T. Manson, J. Electron Spectrosc. Relat. Phenom. 8 (1976) 389–394.
[25] [26] [27] [28]
[29] [30] [31] [32] [33] [34] [35] [36] [37] [38]
351
G.L. Makar, J. Kruger, Int. Mater. Rev. 38 (1993) 138–153. R. Astala, M.J. Stott, Chem. Mater. 17 (2005) 4125–4133. H. El Feki, J.M. Savariault, A. Ben Salah, J. Alloys Compd. 287 (1999) 114–120. J.F. Moulder, J. Chastain, Handbook of X-ray Photoelectron Spectroscopy: A Reference Book of Standard Spectra for Identification and Interpretation of XPS Data, Physical Electronics Division, Perkin-Elmer Corporation, 1992. O. Suzuki, Acta Biomater. 6 (2010) 3379–3387. R. Rettig, S. Virtanen, J. Biomed. Mater. Res. A 88A (2009) 359–369. F. Witte, F. Feyerabend, P. Maier, J. Fischer, M. Stormer, C. Blawert, W. Dietzel, N. Hort, Biomaterials 28 (2007) 2163–2174. I.V. Fadeev, L.I. Shvorneva, S.M. Barinov, V.P. Orlovskii, Inorg. Mater. 39 (2003) 947–950. K.S. TenHuisen, P.W. Brown, J. Biomed. Mater. Res. 36 (1997) 306–314. A. Bigi, G. Falini, E. Foresti, M. Gazzano, A. Ripamonti, N. Roveri, J. Inorg. Biochem. 49 (1993) 69–78. V.C. West, Calcif. Tissue Res. 7 (1971) 212. M.J. Root, Calcif. Tissue Int. 47 (1990) 112–116. M. Pourbaix, Atlas of Electrochemical Equilibria in Aqueous Solutions, 1st English Ed, Pergamon Press, Oxford, New York, 1966. V.F. Lvovich, Impedance Spectroscopy: Applications to Electrochemical and Dielectric Phenomena, Wiley, Somerset, NJ, USA, 2012.