Journal of Nanomaterials

Journal of Nanomaterials / 2016 / Article
Special Issue

Nanostructural Biomaterials and Applications

View this Special Issue

Research Article | Open Access

Volume 2016 |Article ID 9143969 | 13 pages |

Improved Bonding Strength of Hydroxyapatite on Titanium Dioxide Nanotube Arrays following Alkaline Pretreatment for Orthopedic Implants

Academic Editor: Lei Liu
Received22 Apr 2016
Accepted04 Jul 2016
Published11 Aug 2016


Hydroxyapatite (HA) is a bioactive bone substitute used in biomedical applications. One approach to use HA for bone implant application is to coat it on titanium (Ti) implant. However, adhesion of HA on Ti is major concern for their long-term use in orthopedic implants. To enhance the adhesion strength of HA coating on titanium (Ti), the surface of the Ti was anodized and alkaline pretreated prior to coating on Ti by electrodeposition. Alkaline pretreatment of titanium dioxide nanotubes (ATi) accelerated the formation of HA, which mimicked the features and structure of natural bone tissue. Nanostructured HA formed on the ATi and pretreated ATi (P-ATi), unlike on conventional Ti. This study is the first to show that the bonding of HA coating to a P-ATi substrate was stronger than those of HA coating to Ti and to ATi. The preosteoblast response tests were also conducted. The results indicated that HA coating improved preosteoblast proliferation after 3 days in standard cell culture.

1. Introduction

Recently, more than 90 percent of the elderly of world’s populations suffer from bone-related trauma, such as osteoporosis, bone cancers, rheumatoid arthritis, or accidents, which require spinal, hip, and knee replacements [1]. Thus, a huge success in bone implant development is anticipated in 2030 [2]. In Thailand, the total cost of orthopedic implants increased from $4.9 million in 2002 to $39.7 million in 2010.

Bone is composed of ceramic and biosubstance that draws attention among material engineers, who look for material to substitute for bone. This substituted material should possess both high strength and fracture toughness, appropriated for orthopedic implants [3]. Common biomaterials used for bone replacements are stainless steels, cobalt-chromium (Co-Cr) alloys, titanium (Ti), and Ti alloys. Among these metals, the tensile modulus of Ti-based materials is closest to that of bone. Additionally, wear and corrosion resistances of Ti-based materials are higher than those of the other two materials because of natural TiO2 formation on the Ti-based surface. Although TiO2 is bioinert in physiological environments, Ti-based implants often fail after 10–20 years of service. Failure modes include bone fracture around the implanted materials (due to lower elastic modulus of bone than that of the implant), wear or corrosion of the implant, inflammation, and infection [4]. Moreover, naturally formed TiO2 has low osteoconduction [5, 6], which causes implant loosening leading to failure [7]. To provide high bioactivity and to improve bone ingrowth, one approach is to use Ti with a coating of nanostructured hydroxyapatite (HA).

HA is a naturally derived ceramic found in bones and its calcium to phosphate atomic ratio (Ca/P) is 1.67 [8]. Various coating methods of hydroxyapatite (HA) on Ti surface have been investigated, such as Ti soaking in simulated body fluid (SBF) (mimicking naturally HA forming), plasma spraying, sol-gel deposition in HA particles solution, pulsed laser deposition, hot isostatic pressing, and electrochemical deposition [916]. HA coating by electrochemical deposition was shown to possess excellent corrosion protection and good biocompatibility under host environment [17]. Advantages of this technique are that it is relatively inexpensive and thickness of the coating can be easily controlled. Another concern point is that mechanical bonding strength of the HA coating should withstand the bone growth stresses, induced during the bone healing process [18]. The adhesion of electrodeposited HA on smooth Ti surface was observed to be very week, which may cause implant loosening [1922]. Therefore, to increase the adhesion strength between HA and Ti surface, one suggested method is to use TiO2 nanotube array as a substrate for HA electrodeposition. In addition to improving bonding strength, HA grown on TiO2 nanotube array is more favorable to osteoblast (or bone-forming cell) growth and differentiation when compared with naturally formed TiO2 layer on conventional Ti. It was also found that nanostructure surface could promote deposition of bone minerals, induced by osteoblasts [23].

TiO2 nanotubular can be formed by several methods, such as sol-gel method [24], electrophoretic deposition [25], and anodization [26]. These methods could form different tube alignment on surface. The vertically aligned TiO2 nanotube arrays can be achieved by anodization method with various parameters (electrolytes, voltages, etc.) [27, 28]. In the present study, anodization using pulse voltage and neutral-viscous, fluoride-containing electrolyte was investigated to achieve a uniform, well-oriented TiO2 nanotube arrays (ATi). The previous study shows that alkaline surface treatment on ATi and subsequently annealing of HA-electrodeposited ATi at high temperature (>300°C) improve the bonding strength between HA and ATi [29]. However, the results from another study suggested that annealed HA grown in simulated body fluid (SBF) at high temperature (>300°C) was not always necessary because HA grown in SBF still possesses the Ca/P ratio of 1.67, closely mimicking natural bone minerals [30, 31].

In the present study, HA coatings were coated onto alkaline pretreated TiO2 nanotubes. The anodization conditions included two different values of pulse voltage (+20/−4 and +35/−4 V) and temperatures (5 and 25°C). The influence of alkaline pretreatment on ATi and physical structures (e.g., wall thickness, diameter, and length) of TiO2 nanotubes on the growth of bioinspired HA was reported. The mechanical bonding strength of HA on its substrates was studied as an in vitro early assessment of long-term stability in the human body. The physical and chemical characteristics (e.g., structure, size, and a ratio of Ca/P) of HA were investigated to elucidate the effect of using this biomaterial to increase the preosteoblast viability under standard cell culture conditions.

2. Experimental Materials and Methods

2.1. Forming TiO2 Nanotube Arrays

Ti plate (Supra Alloys Inc., USA) of 99.9% purity was polished using silicon carbide papers (TOA, Thailand) of progressively decreasing roughness of 400, 600, 800, 1000, 1500, and 2000 grits. The samples were then polished with 0.05 μm alumina powder (Allied, USA) to achieve a mirror quality finish. Freshly polished Ti was then washed with deionized (DI) water and sonicated in acetone (Sigma-Aldrich, Thailand) and 95 vol% ethanol (Sigma-Aldrich, Thailand) for five minutes each before immediate anodization.

Anodization method was used to produce TiO2 nanotube arrays. Ti was used as a positive electrode, whereas platinum was used as a negative electrode. The electrolyte was a mixture of 90 vol% glycerol and 10 vol% ammonium fluoride (NH4F) (0.36 M) in water. Anodization was performed with the pulse voltage of either +20/−4 or +35/−4 V, the conditions from a previous study by Chanmanee et al. [32]. The anodization was carried out at either 5°C or 25°C for 90 minutes. Then anodized sample was washed with deionized water before it was dried with nitrogen gas at room temperature. All ATi samples were then annealed at 450°C for 30 minutes to obtain the anatase phase.

2.2. Electrodeposition of Hydroxyapatite

ATi samples were pretreated with 1 M NaOH at 50°C for two minutes. After the pretreatment, the HA deposition was conducted. The electrolyte for HA deposition was prepared by dissolving ammonium phosphate (NH4H2PO4, 1.67 mM) (Sigma-Aldrich, Thailand) and calcium nitrate (Ca(NO3)2, 2.5 mM) (Sigma-Aldrich, Thailand) in distilled water. To increase ionic conductivity and to buffer pH of electrolyte at 7.2, 0.15 M NaCl (Ajax Finechem, New Zealand), tris(hydroxyl aminomethane) (Ajax Finechem, New Zealand), and hydrochloric acid (Ajax Finechem, New Zealand) were mixed [29]. The pretreatment ATi (P-ATi) was used as a negative electrode, whereas platinum was used as a positive electrode. Electrodeposition of HA was processed at −2.5 V and at 80°C for 10 minutes.

2.3. Physical and Chemical Characterizations of Anodized Titanium and Hydroxyapatite Coating

Scanning electron microscopy (FE-SEM, CamScanMX2600, UK) was used to investigate surface morphology of ATi. The dimensions of the TiO2 nanotubes were measured using image analysis software (ImageJ version 1.32, NIH). The crystal structures of the surfaces of the NaOH-pretreated ATi and the HA-coated samples were investigated by X-ray diffractometer (Shimadzu Model: XRD 6000, Japan).

2.4. The Mechanical Properties of Coated Hydroxyapatite on Anodized Titanium

The mechanical bonding strength of coated HA on ATi was evaluated by tensile testing following an ASTM F 1044-05 standard [33]. Two HA-coated samples were glued together face-to-face using Plasmatex Klebbi adhesive (Plasma-Technik AG, Switzerland) and cured at 190°C for two hours. Tensile tests were conducted by a computer-controlled universal testing machine (Instron, Model 5500R, US) at a cross head speed of 0.25 centimeter per minute (0.1 inch/minute). The degree of bonding strength was calculated as shown in

2.5. Preosteoblast Cell Response
2.5.1. Cell Culture

Preosteoblasts (MC3T3-E, passage number = 10, MTEC, Thailand), or bone-forming cells, were used in the present study. Cells were cultured in alpha-modified minimal essential medium (alpha-MEM; Invitrogen Corporation, Paisley, UK) supplemented with 10 vol% fetal calf serum (Dominique Dutcher, Brumath, France) and 1 vol% penicillin/streptomycin (Invitrogen Corporation) at 37°C in humidified atmosphere of 5% CO2 in air. The cell culture media were replaced every three days. Cells were seeded and cultured on plastic polystyrene (control); Ti; ATi; P-ATi; HA-ATi; and HA-P-ATi at the cell density of 5 × 104 cells/cm2.

2.5.2. Cell Morphology

After three days in culture, cells were fixed with 2.5 vol% sodium phosphate buffered glutaraldehyde (Sigma-Aldrich, Thailand) at pH 7 for 20 minutes. After two washes, the cells were postfixed with 1 vol% osmium tetroxide (OsO4) (Sigma-Aldrich, Thailand) in saturated mercuric chloride (HgCl2) (Sigma-Aldrich, Thailand). The samples were then dehydrated with a series of ethanol washes (20, 30, 40, 50, 60, 70, 80, 90, and 100 vol%) at room temperature. All samples were subsequently critical-point dried (EMSCOPE CPD-750, Ashford, UK). The samples were coated with gold (EMSCOPE SC-500, UK) at a thickness of 10 nm before being characterized by SEM. The cell morphology on Ti, ATi, P-ATi, HA-coated on ATi (HA-ATi), and HA-coated on P-ATi (HA-P-ATi) was observed using SEM (JEOL J-SM-5300, Japan) at an accelerating voltage of 20 kV.

2.5.3. Cell Viability

Cells were seeded onto plastic polystyrene (control) and coated samples at a density of 5 × 104 cells·cm−2 in 24-well culture plates. Cell viability was tested using a commercial 3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyl tetrazolium bromide (MTT) assay (Sigma-Aldrich, Thailand). The 10 vol% MTT solution in 1x phosphate buffer saline was mixed with alpha-MEM without phenol red to form a yellowish solution before being added onto the cell-seeded samples at the day 3 of cultures. Mitochondrial dehydrogenases of viable cells cleave the tetrazolium ring, yielding purple formazan crystals on cell-seeded samples after incubation for an hour. The absorbance of purple solution was measured at 570 nm wavelength using a spectrophotometer (Synergy Mx Multimode Reader, US). A concomitant change in the amount of formazan formed correlates to the change in the number of viable cells on the samples. The percentage of viable cells on the samples was calculated as shown in

2.6. Statistical Analysis

Two-way analysis of variance (ANOVA) was used in the analysis of wall thickness, diameter, and length of TiO2 nanotubes, which were anodized with different voltage pulses and temperatures. ANOVA was used to statistically analyze the effects of materials on bonding strength of the HA coatings and cell viability. The statistical analysis was performed using Minitab 16 (Minitab Inc., USA) software. Significant level of 0.05 () was used for the test.

3. Results and Discussion

3.1. Characteristics of Anodized Titanium

Gong et al. reported that a certain anodization potential must be applied to yield ordered TiO2 nanotube arrays [26]. Therefore, in this study, to control the size of the nanotubes, anodization potentials were varied. Figure 1 showed top-view SEM images of ATi, which were anodized under two different temperatures and pulse voltages: 5°C, +20/−4 V (Figure 1(a)); 5°C, +35/−4 V (Figure 1(b)); 25°C, +20/−4 V (Figure 1(c)); and 25°C, +35/−4 V (Figure 1(d)) for 90 minutes. The insets in Figures 1(a)1(d) showed cross sections of TiO2 nanotube arrays. The SEM images demonstrated that the oxide nanotubes formed in a uniform shape under the conditions of +20/−4 V at both 5 and 25°C, while the nonuniform nanotubes were found for the sample prepared under the conditions of +35/−4 V at both temperatures. The formation of the nonhomogeneous nanotubes (at 35 V) may be because of the high electrical field inducing etching of TiO2 at higher anodization potential. While the electrochemical oxidation rate was increased at the higher anodization potential, longer tube lengths and formation oxide films were observed (inset of cross sections). At an anodization potential of +20 V, the formation and etching of TiO2 were slower than those formed at +35 V. Thus, the applied pulse voltage at +20/−4 V in anodization led to the formation of shorter and more uniform nanotube structures than that formed at +35/−4 V.

The wall thickness, tube diameter, and tube length of TiO2 nanotubes were measured using ImageJ and were plotted as shown in Figure 2. The effects of each anodization parameter on nanotube diameter and wall thickness were summarized in Table 1. The results suggested that temperature and pulse voltage in anodization had a remarkable effect on the TiO2 tubular features.


(i) 5°C to 25°C
At the same voltage
TW not changed
TD ↑
TL ↑

(ii) +20 V to +35 V
At the same temperature
A constant voltage of −4 V
TW not changed
TD ↑
TL ↑

Electrolytes containing fluorine are known as the most efficient electrolytes for anodic formation of TiO2 nanotube arrays. For electrochemical anodization with pulse potential of +20/−4 V or +35/−4 V, TiO2 nanotube layers were formed by self-organization of TiO2 as a result of the balance between electrochemical oxidation of Ti and TiO2 (reaction (3)). The induced electrical field caused dissolution of the TiO2 by fluorine ions (reactions (4) and (5)) [34, 35]. Dissolution of the TiO2 was partially suppressed by the binding of /NH3 species on the TiO2 surfaces (reaction (6)) as described by Chanmanee et al. [32]:

Reactions (3) and (5) occur during positive potential, whereas negative voltage electrostatically induces the binding of species with TiO2, as shown in reaction (6), to form adsorbing . The role of is to protect the nanotube walls against chemical etching by fluoride ions even in short times of negative potential (1 sec.) [32, 3638]:

In the present study, it was found that tube diameter and length of ATi strongly depended on temperature and applied pulse voltage (Figures 1 and 2). Figure 2 showed that nanotube diameter and tube length formed at various temperatures were significantly different with significant level of at least 0.05, while the wall thicknesses were comparable. The result suggested that the rate of TiO2 formation and chemical dissolution at 25°C were faster than those at 5°C, while the TiO2 formation rate and etching rate of TiO2 induced by the electric field were similar at both temperatures [28]. However, at low temperature, etching of TiO2 by fluorine ions is largely suppressed, leading to small tube diameter and similar wall thickness. The previous study by Wang and Lin found that, at room temperature, the TiO2 formation rate was higher than that in an ice bath [28], while the etching rate of TiO2 induced by electric field and fluorine ions remained similar at both temperatures. Their results showed a difference in inner diameter but not in the outer diameters. Their work was conducted in hydrofluoric electrolyte; thus, the anodization was not suppressed by neutral-viscous electrolyte. This phenomenon may be the reason to obtain different result from this present study. However, Wang and Lin also found that, at low temperature (~0°C), even though the potential was increased, the chemical dissolution of TiO2 by fluorine ions was still very low in the glycerol solution containing 0.25 wt% ammonium fluorides. Therefore, only a small difference in both wall thickness and diameter of the nanotube was observed at a constant temperature [28].

3.2. Hydroxyapatite Electrodeposition
3.2.1. Effect of Alkaline Pretreatment

From the previous work, the sodium titanate (Na2Ti5O11 or Na2Ti6O13) was deposited on the top of TiO2 nanotubes and acted as nucleation sites for nanoscale of HA electrodeposition [36]. The SEM images in Figures 3 and 4 showed the HA formation on ATi and P-ATi, respectively. From Figures 3(a)3(d), without alkaline pretreatment, the HA coatings appeared as an oriented rod-like crystal structure on the ATi substrates with the average HA crystal sizes of 175, 107, 155, and 316 nm, respectively. On the P-ATi substrates, the HA coatings formed as an unoriented rod-like crystal structure with the average HA crystal sizes of 171, 194, 105, and 151 nm, respectively, as shown in Figures 4(a)4(d). The average size of HA crystal on ATi (107 nm) in Figure 3(b) is smaller than the average HA crystal size on P-ATi (194 nm) in Figure 4(b). In this study, the results confirmed that the HA crystal was finer on ATi (5°C, +35/−4 V) without NaOH pretreatment.

3.2.2. Formation of Hydroxyapatite on Different Substrates

From Table 2, the EDS analysis of HA-ATi and HA-P-ATi revealed that the calcium to phosphate ratios, Ca/P, were in the range of 1.43 to 1.68 when ATi was used as a substrate. The HA on ATi and P-ATi had a Ca/P ratio closely mimicking that of natural bone. The HA coating on conventional Ti, however, had a Ca/P ratio of about 0.9. This ratio is close to that of dicalcium phosphate in bone [17]. The crystal structures of grown calcium phosphate minerals were analyzed by XRD, as shown in Figure 5. The XRD spectra and EDS analysis confirmed that HA (Ca10(PO4)6(OH)2) coating formed on ATi and P-ATi, but perhaps dicalcium phosphate formed on the conventional Ti.

Substrate The ratio of Ca and P
PretreatmentWithout pretreatment

ATi: 5°C/+20 V 1.561.44
ATi: 5°C/+35 V 1.461.67
ATi: 25°C/+20 V 1.661.68
ATi: 25°C/+35 V 1.431.61
Conventional Ti (control)0.890.97

The formation of HA crystals on Ti and ATi by electrodeposition in the solution containing calcium and phosphorus was investigated in the present study. Electrochemical reactions of HA formation are shown as follows [39]:

The reduction of water resulted in the release of H2 and an increase of hydroxide ions, as shown in reaction (7). As a result of reaction (7) [39], the pH between the cathode and electrolyte interface increases. When pH was higher than 7.64, the rates of the HA electrodeposition of reactions (8) and (9) were accelerated. When the interfacial pH was high, it provided an appropriate chemical environment to form HA, Ca10(PO4)6(OH)2, coatings on the cathode (reaction (10)) [40]. In addition, the electric field draws Ca2+ ions to the cathode surface causing that Ca/P ratio is closer to that of natural bone than the Ca/P ratio obtained by conservative methods (precipitation of calcium phosphates from aqueous solution).

The results from Figures 3 and 4 and Table 2 suggested that the chemical composition of HA depends on the surface structure of substrates. It was known that it is easier for electrons to travel through TiO2 nanotubes electrode than through TiO2 compact layers [34]. Therefore, feature of HA (Ca10(PO4)6(OH)2) coatings on ATi electrode was more orderly oriented than those on conventional Ti because HA layers on ATi possess Ca/P ratios near to that of natural bone due to Ca2+ ions enrichment on electrode (reaction (10)). Since ATi sample can carry more OH groups on its surface during NaOH pretreatment, it supports a very dense formation of apatite nuclei. Another study also found that HA deposition on conventional Ti from the solution containing Ca2+ and could obtain four different kinds of calcium phosphates, including Ca10(PO4)6(OH)2, Ca2(PO4)3·H2O, Ca8H2(PO4)6·5H2O, and CaHPO4·2H2O [41].

3.2.3. Mechanical Bonding Strength of Hydroxyapatite on Anodized Titanium

The bond strength between coated HA and the implant material is very important because the adhesive failure of implant after implantation easily takes place when the bond strength is low. Therefore, the present study aimed to study effect of morphology of substrates used for HA coatings on bond strength between them. As shown previously that only nanotube length and tube diameter varied with anodization conditions, length () to diameter () of TiO2 nanotube was used to test for this effect. Figure 6(a) showed ratio of TiO2 nanotube arrays for ATi and P-ATi, anodized at the following conditions: 5°C, +20/−4 V, ; 5°C, +35/−4 V, ; 25°C, +20/−4 V, ; and 25°C, +35/−4 V, . The results suggested that using either higher positive voltage or higher temperature increased ratios. It was found that bonding strength of the HA coating on P-ATi is higher than that of HA-ATi (such that, at 5°C, +35/−4 V with , found bonding strength of the HA coating on P-ATi is 21 MPa and via ATi is 12 MPa). It is due to the fact that the NaOH pretreatment increases the pH inside the titania nanopores, consequently, nucleation of HA crystals was enhanced during electrodeposition. As described by Kim et al., during the alkaline treatment, the protective oxide layer on Ti is dissolved into solution because of corrosive attack by hydroxyl groups [42]. Negatively charged hydrates, produced on the ATi substrate surface, combine with alkali ions from the aqueous solution to form a hydrogel layer of sodium titanate. After the hydrogel layer was exposed to high temperature, the layer was dehydrated and was densified to form a stable alkali titanate layer. The compact layer of alkali titanate facilitates good adhesion between the HA coating and Ti or ATi substrate. The bonding strength progressively decreased as increased when was higher than 2.92. At of 2.92, the highest bonding strength between HA and P-ATi was obtained. This phenomenon indicated that nanotube geometry is an important factor effecting on bonding strength between coated HA and substrate.

According to the ASTM F 1044-05 standard, there are three types of failures that can occur during the tensile test: adhesive, cohesive, and a combination of these two modes (Figures 6(b) and 6(c)). For the HA coated at 5°C, +35/−4 V (), during the shear bond tests, the fracture occurred only at the HA coating layers, as shown in Figure 6(c)(I). Accordingly, it can be stated that cohesion strength between the HA and TiO2 nanotube was stronger than that between the HA layers. The same fracture surface was found for the deposited HA prepared at 25°C, +20/−4 V (). For the HA coated at 25°C, +35/−4 V (), fracture occurred at the HA and TiO2 interface as well as at the HA layer as shown in Figure 6(c)(II). From Figure 6(c)(II), residues of TiO2 nanotube were found in both sides of the fracture surface. This failure indicated a combination of adhesive and cohesive failures and also occurred for the HA coated at 5°C, +20/−4 V (). Using both bonding strength and fracture surface information, the high bonding strength was found when cohesion between HA and TiO2 nanotube is stronger than adhesion between HA layers. The dependence of bonding strength on ratio may be described as the following. When is 1.27, the depth that deposited HA could form along nanotube is very shallow. As shown in Figure 2, the nanotube length in this case is only about 120 nm. Thus, nanotube surface area, which acts as nucleation site for HA formation, is low, leading to less HA-nanotube interface area. In addition, due to short nanotube length, both the nanotubes and Ti base may be subjected to the glue. Hence, the nanotube base was easily destroyed during testing. For the highest ratio of 7.03, the presence of nanotube residue indicated that TiO2 nanotubes were damaged. From Figure 2, tube length was approximately 1120 nm, which is almost four times longer than that of the sample whose ratio is 2.97. In this case, the failure of nanotube could be explained by the Euler equation; [43]. As shown in the equation, critical stress () that the column can bear is inversely proportional to the second order of column height, . The higher the height of the column, the lower the critical stress that the column can bear. As a result, tube strength of nanotube layer is inversely proportional to nanotube length.

For the sample with ratio of 5.96, the bonding strength is lower and the tube length is about two times higher than that for the sample with ratio of 2.97. The Euler equation may be applied for this case; nevertheless, the fracture surface of the two samples indicated the similar adhesion mode. Further study is needed to analyze the effect of geometry in this range. In case of HA on conventional Ti substrate (Figure 6(c)(III)) fracture during the test occurred at the calcium phosphate-Ti interface, which indicated that only adhesive failure occurred. From these results, it can be concluded that as ratio is approximately 3.00–6.00, the interfacial adhesion between HA and TiO2 array is stronger than cohesion between HA layers.

In conclusion, the topography and chemical composition of Ti substrates before the electrodeposition of bioinspired HA directly influence the crystal structure and phases as well as mechanical bonding strength of HA coating. Following ISO Standard 13779-4:2002, bonding strength between coating and substrate of apatite coating implant should be higher than 15 MPa [44]. With this result, ATi samples anodized at 5°C and +35/−4 V () were further studied for cell morphology and viability with preosteoblasts.

3.2.4. Responses of Preosteoblast Cells on Different Hydroxyapatite Coatings

ATi and P-ATi anodized at +35/−4 V and 5°C, as shown in Figures 3(b) and 4(b), respectively, were chosen for cell culture testing because the mechanical bonding strength of HA on these substrates was high, as shown in Figure 6. Figure 7 shows cell proliferation on Ti, ATi, P-ATi, HA-ATi, and HA-P-ATi. Previously, the samples were seeded with MC3T3-E1 preosteoblast cells and cultured for three days in a standard cell culture condition. The results of the MTT assay indicated that all of the samples had good cytocompatibility. The percentage of viable cells was the highest on HA-ATi. The electrodeposited HA, whose Ca/P ratio is similar to that of natural bone, exhibited the oriented rod-like crystal structure when it formed on ATi with an average crystal size of 107 nm. The results suggested that osteoblasts preferred to grow on HA-ATi compared to HA-P-ATi, whose average crystal size was larger, 194 nm. This result is in an agreement with other works [6, 23, 45]. Webster and Ejiofor suggested that a small HA crystal size improved the biocompatibility of implants and increases osteoblast adhesion [23]. Tsuchiya et al. reported that nanostructured materials significantly improved osteoblast growth while they inhibited cell apoptosis [6]. Moreover, the nanostructure of biomaterials promoted cell functions, such as the synthesis of extracellular matrix proteins and calcium mineral deposition [45]. Thus, the size of the HA crystals formed on ATi and P-ATi directly affects osteoblast proliferation in this study.

Adhesion of osteoblast cells is a crucial prerequisite to all subsequent cell functions. According to Webster et al., the use of nanostructured ceramics (such as alumina, titania, and HA) significantly improved osteoblast adhesion [46, 47]. Importantly, the results shown in Figure 7 are in good agreement with previous findings that finer HA crystals improve cell proliferation. A study by Chou et al. demonstrated that small plate-like HA had greater preosteoblast proliferation after 4 days than large plate-like HA. Large plate-like HA, however, induced higher expression of the mature osteogenic markers osteocalcin and bone sialoprotein in preosteoblasts after 21 days of culture compared to small plate-like HA, polystyrene, and conventional apatite [48]. Further studies are needed to confirm its effects on osteoblast differentiation and new bone formation juxtaposed to orthopedic implants.

Figures 7(b)–7(f) show that osteoblasts differed in their size, spreading, and adhesion on the samples. Figure 7(a) shows that although Ti is a biocompatible material, it is a bioinert material, which does not adequately increase osteoblast attachment, spreading, proliferation, or differentiation. In contrast, HA, as a major inorganic component of hard tissues, possesses excellent biocompatibility, bioactivity, and osteoconductivity [49]. Cells exhibited larger spreading on Ti than on all other samples, but cell peeling was observed due to the lower cell adhesion. Particularly noteworthy are HA-ATi and HA-P-ATi, which induced maximum cell adhesion with a smaller cell size. Cells appeared flattened and spread more widely on Ti, ATi, and P-ATi samples (Figures 7(b)–7(d)) than on the HA-coated samples (Figures 7(e) and 7(f)). Osteoblasts with no appendix on their substrates and cell peeling were observed on Ti, ATi, and P-ATi (Figures 7(b)–7(d)). This is perhaps due to the fact that the cell adhesion strength on the non-HA-coated samples was lower than that on HA-ATi and HA-P-ATi. HA improves osteoblast adhesion because it is a main component of the bone extracellular matrix. The HA coating reduced osteoblast spreading but improved the elongation of cell projections.

The effect of NaOH pretreatment on cell viability was also examined. Cell morphology of preosteoblasts on the ATi and P-ATi was similar (Figure 7). The ATi (without NaOH pretreatment), however, was more favorable for preosteoblast proliferation than P-ATi, possibly due to their different surface compositions. Sodium titanate was present on the P-ATi surface but not on the ATi surface. HA-P-ATi has a lower ratio of Ca/P (1.46) and an unoriented rod-like crystal HA coating (Figure 4(b)), while HA on ATi has a Ca/P of 1.67 and an oriented rod-like crystal structure. The percentage of viable cells after 3 days of culture was greater on HA-ATi than on HA-P-ATi, but the cell morphology of preosteoblasts on HA-ATi was similar to that on HA-P-ATi. A previous study suggested that osteoblast adhesion and metabolic activity were very sensitive to surface morphology and roughness [50]. Surface energy is also reported to be an influential surface characteristic on cellular proliferation [51]. Therefore, the higher cell viability on nanoscale HA-ATi could be due to the positive effects of the surface energy. Moreover, HA with low crystallinity may remarkably increase degradation rates in vitro or in vivo [49]. HA is the most stable and least soluble in aqueous media of all calcium phosphates [52]. It becomes more acidic and water-soluble as its Ca/P ratio decreases [53]. Thus, the dissolution of HA in HA-P-ATi may slow the growth of preosteoblast cells. Approximately 2–4 mM of Ca2+ is suitable for osteoblast proliferation and survival, and slightly higher concentrations (6–8 mM) favor osteoblast differentiation and mineralization. An excessively high concentration, >10 mM, can be cytotoxic to cells [54]. It is possible that the high concentration of Ca2+ in HA-P-ATi decreased preosteoblast proliferation in this study. On the other hand, with a higher Ca/P ratio (1.67) of HA-P-ATi, free phosphate level in the apatite microenvironment is also perhaps lower, and thus fewer phosphates participating in osteopontin regulation are transported into the cells [55]. This can cause lower cell proliferation on HA-P-ATi than on HA-ATi after 3 days of culture. Another possible mechanism is that high phosphate levels may induce the expression of calcium channels and phosphate transporters and activation of ERK1/2, but not MAPK, p38, or JNK signaling pathways [56]. This cascade participates in the regulation of a large variety of processes, including cell adhesion, cell cycle progression, cell migration, cell survival, differentiation, metabolism, proliferation, and transcription [57]. Therefore, we conclude that the size and Ca/P ratio of the HA crystal directly affect the cell proliferation, but not the cell morphology, of preosteoblasts.

Importantly, the nanotube formation of TiO2 is the predominant factor to increase the bonding strength between the HA coating and TiO2 nanotube substrate. The nanotube formation can promote a HA coating in the phase closely mimicking that found in natural bone. Therefore, the electrodeposited HA coating on ATi and P-ATi can illicit more favorable responses of preosteoblast morphology and proliferation than the samples without HA coating.

4. Conclusions

TiO2 nanotube arrays were anodized for 90 min under various conditions: 5°C, +20/−4 V; 5°C, +35/−4 V; 25°C, +20/−4 V; and 25°C, +35/−4 V. HA was then electrodeposited onto ATi and P-ATi (with NaOH pretreatment). Increasing the pulse-positive voltage (from +20 V to +35 V) and temperature (from 5°C to 25°C) during anodization increased the nanotube length. At 25°C, changing the anodization voltage did not significantly affect the nanotube wall thickness. The bonding strength of the HA coatings on ATi layers was stronger than that of the HA coating on conventional Ti. Importantly, the bonding strength of the HA coating was higher on HA-P-ATi than on HA-Ti and HA-ATi. The results suggested that the HA crystals on ATi had an oriented rod-like crystal structure with an average size of 107 nm and a Ca/P ratio of 1.67. The HA crystals on P-ATi had an unoriented rod-like structure with an average size of 194 nm and Ca/P ratio of 1.46. The percentage of osteoblast viability at day 3 of cell culture was significantly higher on HA-P-ATi (98%) than those on Ti (59%) and P-ATi (92%) but lower than those on ATi (108%) and HA-ATi (117%). Higher cell proliferation was observed on the smaller HA crystals of HA-ATi. The Ca/P ratio (1.67) of the HA crystals on ATi reduced the dissolution of calcium ions and phosphates, thus allowing for better cell proliferation. The electrodeposited HA coating improved preosteoblast adhesion as cell projections were elongated. In summary, P-ATi increases the mechanical bonding strength of an electrodeposited HA coating but decreases the Ca/P ratio of HA with a larger crystal size, thus leading to lower preosteoblast proliferation after 3 days of culture.

Competing Interests

The authors declare that there are no competing interests regarding the publication of this paper.


The authors would like to acknowledge financial supports from Commission on Higher Education (the Faculty Development Scholarships) and a Research Strengthening Project of the Faculty of Engineering, King Mongkut’s University of Technology Thonburi, Bangkok, Thailand. They also thank Associate Professor Dr. P. Kajitvichyanukul, Associate Professor Dr. K. Pasuwat, Mr. C. Manussapol, Mr. C. Jongwanasiri, and Ms. K. Siruksasin for their contributions and useful advice in this study; Dr. W. Chanmanee for his advice in anodization; and Ms. W. Chokevivat from National Metal and Materials Technology Center for her help in animal cell experiments. They acknowledge Dr. Carmen V. Gaines (Ph.D.) and Mr. Cristian Guajardo for her and his useful discussion.


  1. S. Ramakrishna, J. Mayer, E. Wintermantel, and K. W. Leong, “Biomedical applications of polymer-composite materials: a review,” Composites Science and Technology, vol. 61, no. 9, pp. 1189–1224, 2001. View at: Publisher Site | Google Scholar
  2. J. B. Park and J. D. Bronzino, Biomaterials: Principles and Applications, CRC Press, Boca Raton, Fla, USA, 2003.
  3. M. J. Olszta, X. Cheng, S. S. Jee et al., “Bone structure and formation: a new perspective,” Materials Science and Engineering R: Reports, vol. 58, no. 3–5, pp. 77–116, 2007. View at: Publisher Site | Google Scholar
  4. R. S. Park and J. B. Lakes, Biomaterials: An Introduction, Plenum Press, New York, NY, USA, 1992.
  5. P. Li and P. Ducheyne, “Quasi-biological apatite film induced by titanium in a simulated body fluid,” Journal of Biomedical Materials Research, vol. 41, no. 3, pp. 341–348, 1998. View at: Publisher Site | Google Scholar
  6. H. Tsuchiya, J. M. Macak, L. Müller et al., “Hydroxyapatite growth on anodic TiO2 nanotubes,” Journal of Biomedical Materials Research—Part A, vol. 77, no. 3, pp. 534–541, 2006. View at: Publisher Site | Google Scholar
  7. G. Balasundaram and T. J. Webster, “A perspective on nanophase materials for orthopedic implant applications,” Journal of Materials Chemistry, vol. 16, no. 38, pp. 3737–3745, 2006. View at: Publisher Site | Google Scholar
  8. M. S. Johnsson and G. H. Nancollas, “The role of brushite and octacalcium phosphate in apatite formation,” Critical Reviews in Oral Biology & Medicine, vol. 3, pp. 61–82, 1992. View at: Google Scholar
  9. T. Kokubo, T. Matsushita, and H. Takadama, “Titania-based bioactive materials,” Journal of the European Ceramic Society, vol. 27, no. 2-3, pp. 1553–1558, 2007. View at: Publisher Site | Google Scholar
  10. A. R. Boccaccini, J. Cho, T. Subhani, C. Kaya, and F. Kaya, “Electrophoretic deposition of carbon nanotube-ceramic nanocomposites,” Journal of the European Ceramic Society, vol. 30, no. 5, pp. 1115–1129, 2010. View at: Publisher Site | Google Scholar
  11. J. D. Haman, L. C. Lucas, and D. Crawmer, “Characterization of high velocity oxy-fuel combustion sprayed hydroxyapatite,” Biomaterials, vol. 16, no. 3, pp. 229–237, 1995. View at: Publisher Site | Google Scholar
  12. H. Hero, H. Wie, R. B. Jorgensen, and I. E. Ruyter, “Hydroxyapatite coatings on Ti produced by hot isostatic pressing,” Journal of Biomedical Materials Research, vol. 28, no. 3, pp. 343–348, 1994. View at: Publisher Site | Google Scholar
  13. R. Hu, C.-J. Lin, and H.-Y. Shi, “A novel ordered nano hydroxyapatite coating electrochemically deposited on titanium substrate,” Journal of Biomedical Materials Research Part A, vol. 80, no. 3, pp. 687–692, 2007. View at: Publisher Site | Google Scholar
  14. A. Montenero, G. Gnappi, F. Ferrari et al., “Sol-gel derived hydroxyapatite coatings on titanium substrate,” Journal of Materials Science, vol. 35, no. 11, pp. 2791–2797, 2000. View at: Publisher Site | Google Scholar
  15. V. Nelea, C. Ristoscu, C. Chiritescu et al., “Pulsed laser deposition of hydroxyapatite thin films on Ti-5Al-2.5Fe substrates with and without buffer layers,” Applied Surface Science, vol. 168, no. 1–4, pp. 127–131, 2000. View at: Publisher Site | Google Scholar
  16. Y. C. Tsui, C. Doyle, and T. W. Clyne, “Plasma sprayed hydroxyapatite coatings on titanium substrates part 2: optimisation of coating properties,” Biomaterials, vol. 19, no. 22, pp. 2031–2043, 1998. View at: Publisher Site | Google Scholar
  17. X. Zhu, D. W. Son, J. L. Ong, and K. Kim, “Characterization of hydrothermally treated anodic oxides containing Ca and P on titanium,” Journal of Materials Science: Materials in Medicine, vol. 14, no. 7, pp. 629–634, 2003. View at: Publisher Site | Google Scholar
  18. R. Rodriguez, K. Kim, and J. L. Ong, “In vitro osteoblast response to anodized titanium and anodized titanium followed by hydrothermal treatment,” Journal of Biomedical Materials Research Part A, vol. 65, no. 3, pp. 352–358, 2003. View at: Google Scholar
  19. K. De Groot, R. Geesink, C. P. A. T. Klein, and P. Serekian, “Plasma sprayed coatings of hydroxylapatite,” Journal of Biomedical Materials Research, vol. 21, no. 12, pp. 1375–1381, 1987. View at: Publisher Site | Google Scholar
  20. H. Kurzweg, R. B. Heimann, and T. Troczynski, “Adhesion of thermally sprayed hydroxyapatite-bond-coat systems measured by a novel peel test,” Journal of Materials Science Materials in Medicine, vol. 9, no. 1, pp. 9–16, 1998. View at: Publisher Site | Google Scholar
  21. C.-M. Lin and S.-K. Yen, “Characterization and bond strength of electrolytic HA/TiO2 double layers for orthopedic applications,” Journal of Materials Science: Materials in Medicine, vol. 15, no. 11, pp. 1237–1246, 2004. View at: Publisher Site | Google Scholar
  22. R. McPherson, N. Gane, and T. J. Bastow, “Structural characterization of plasma-sprayed hydroxylapatite coatings,” Journal of Materials Science: Materials in Medicine, vol. 6, no. 6, pp. 327–334, 1995. View at: Publisher Site | Google Scholar
  23. T. J. Webster and J. U. Ejiofor, “Increased osteoblast adhesion on nanophase metals: Ti, Ti6Al4V, and CoCrMo,” Biomaterials, vol. 25, no. 19, pp. 4731–4739, 2004. View at: Publisher Site | Google Scholar
  24. B. B. Lakshmi, C. J. Patrissi, and C. R. Martin, “Sol-gel template synthesis of semiconductor oxide micro- and nanostructures,” Chemistry of Materials, vol. 9, no. 11, pp. 2544–2550, 1997. View at: Publisher Site | Google Scholar
  25. Z. Miao, D. Xu, J. Ouyang, G. Guo, X. Zhao, and Y. Tang, “Electrochemically induced sol-gel preparation of single-crystalline TiO2 nanowires,” Nano Letters, vol. 2, no. 7, pp. 717–720, 2002. View at: Publisher Site | Google Scholar
  26. D. Gong, C. A. Grimes, O. K. Varghese et al., “Titanium oxide nanotube arrays prepared by anodic oxidation,” Journal of Materials Research, vol. 16, no. 12, pp. 3331–3334, 2001. View at: Publisher Site | Google Scholar
  27. L.-K. Tsui and G. Zangari, “Water content in the anodization electrolyte affects the electrochemical and electronic transport properties of TiO2 nanotubes: a study by electrochemical impedance spectroscopy,” Electrochimica Acta, vol. 121, pp. 203–209, 2014. View at: Publisher Site | Google Scholar
  28. J. Wang and Z. Lin, “Anodic formation of ordered TiO2 nanotube arrays: effects of electrolyte temperature and anodization potential,” The Journal of Physical Chemistry C, vol. 113, no. 10, pp. 4026–4030, 2009. View at: Publisher Site | Google Scholar
  29. A. Kar, K. S. Raja, and M. Misra, “Electrodeposition of hydroxyapatite onto nanotubular TiO2 for implant applications,” Surface and Coatings Technology, vol. 201, no. 6, pp. 3723–3731, 2006. View at: Publisher Site | Google Scholar
  30. H.-M. Kim, H. Kaneko, M. Kawashita, T. Kokubo, and T. Nakamura, “Mechanism of apatite formation on anodically oxidized titanium metal in simulated body fluid,” Key Engineering Materials, vol. 254–256, pp. 741–744, 2004. View at: Publisher Site | Google Scholar
  31. H.-M. Kim, T. Himeno, M. Kawashita, J.-H. Lee, T. Kokubo, and T. Nakamura, “Surface potential change in bioactive titanium metal during the process of apatite formation in simulated body fluid,” Journal of Biomedical Materials Research Part A, vol. 67, no. 4, pp. 1305–1309, 2003. View at: Publisher Site | Google Scholar
  32. W. Chanmanee, A. Watcharenwong, C. R. Chenthamarakshan, P. Kajitvichyanukul, N. R. de Tacconi, and K. Rajeshwar, “Titania nanotubes from pulse anodization of titanium foils,” Electrochemistry Communications, vol. 9, no. 8, pp. 2145–2149, 2007. View at: Publisher Site | Google Scholar
  33. ASTM International, “Standard test method for shear testing of calcium phosphate coatings and metallic coatings,” ASTM F1044-05, ASTM International, West Conshohocken, Pa, USA, 2005, View at: Google Scholar
  34. A. Ghicov and P. Schmuki, “Self-ordering electrochemistry: a review on growth and functionality of TiO2 nanotubes and other self-aligned MOx structures,” Chemical Communications, no. 20, pp. 2791–2808, 2009. View at: Publisher Site | Google Scholar
  35. P. Roy, S. Berger, and P. Schmuki, “TiO2 nanotubes: synthesis and applications,” Angewandte Chemie-International Edition, vol. 50, no. 13, pp. 2904–2939, 2011. View at: Publisher Site | Google Scholar
  36. Y. Parcharoen, P. Kajitvichyanukul, S. Sirivisoot, and P. Termsuksawad, “Hydroxyapatite electrodeposition on anodized titanium nanotubes for orthopedic applications,” Applied Surface Science, vol. 311, pp. 54–61, 2014. View at: Publisher Site | Google Scholar
  37. R. M. Pittman and A. T. Bell, “Raman investigations of NH3 adsorption on TiO2, Nb2O5, and Nb2O5/TiO2,” Catalysis Letters, vol. 24, no. 1-2, pp. 1–13, 1994. View at: Publisher Site | Google Scholar
  38. J. Ahdjoudj and C. Minot, “Adsorption of H2O on metal oxides: a periodic ab-initio investigation,” Surface Science, vol. 402-404, pp. 104–109, 1998. View at: Publisher Site | Google Scholar
  39. Y.-M. Liao, Z.-D. Feng, and S.-W. Li, “Preparation and characterization of hydroxyapatite coatings on human enamel by electrodeposition,” Thin Solid Films, vol. 516, no. 18, pp. 6145–6150, 2008. View at: Publisher Site | Google Scholar
  40. J. M. Zhang, C. J. Lin, Z. D. Feng, and Z. W. Tian, “Mechanistic studies of electrodeposition for bioceramic coatings of calcium phosphates by an in situ pH-microsensor technique,” Journal of Electroanalytical Chemistry, vol. 452, no. 2, pp. 235–240, 1998. View at: Publisher Site | Google Scholar
  41. G. H. Nancollas and B. Tomažič, “Growth of calcium phosphate on hydroxyapatite crystals. Effect of supersaturation and ionic medium,” The Journal of Physical Chemistry, vol. 78, no. 22, pp. 2218–2225, 1974. View at: Publisher Site | Google Scholar
  42. H.-M. Kim, F. Miyaji, T. Kokubo, and T. Nakamura, “Preparation of bioactive Ti and its alloys via simple chemical surface treatment,” Journal of Biomedical Materials Research, vol. 32, no. 3, pp. 409–417, 1996. View at: Google Scholar
  43. B. I. Yakobson, M. P. Campbell, C. J. Brabec, and J. Bernholc, “High strain rate fracture and C-chain unraveling in carbon nanotubes,” Computational Materials Science, vol. 8, no. 4, pp. 341–348, 1997. View at: Publisher Site | Google Scholar
  44. ISO, “Implants for surgery (Hydroxyapatite) Part 4: determination of coating adhesion strength,” Tech. Rep. ISO/CD 13779-4, The International Organization for Standardisation, Geneva, Switzerland, 2002, View at: Google Scholar
  45. K. C. Popat, L. Leoni, C. A. Grimes, and T. A. Desai, “Influence of engineered titania nanotubular surfaces on bone cells,” Biomaterials, vol. 28, no. 21, pp. 3188–3197, 2007. View at: Publisher Site | Google Scholar
  46. T. J. Webster, C. Ergun, R. H. Doremus, R. W. Siegel, and R. Bizios, “Enhanced functions of osteoblasts on nanophase ceramics,” Biomaterials, vol. 21, no. 17, pp. 1803–1810, 2000. View at: Publisher Site | Google Scholar
  47. T. J. Webster, L. S. Schadler, R. W. Siegel, and R. Bizios, “Mechanisms of enhanced osteoblast adhesion on nanophase alumina involve vitronectin,” Tissue Engineering, vol. 7, no. 3, pp. 291–301, 2001. View at: Publisher Site | Google Scholar
  48. Y.-F. Chou, W. Huang, J. C. Y. Dunn, T. A. Miller, and B. M. Wu, “The effect of biomimetic apatite structure on osteoblast viability, proliferation, and gene expression,” Biomaterials, vol. 26, no. 3, pp. 285–295, 2005. View at: Publisher Site | Google Scholar
  49. W. Chen, B. Tian, Y. Lei, Q.-F. Ke, Z.-A. Zhu, and Y.-P. Guo, “Hydroxyapatite coatings with oriented nanoplate and nanorod arrays: fabrication, morphology, cytocompatibility and osteogenic differentiation,” Materials Science and Engineering: C, vol. 67, pp. 395–408, 2016. View at: Publisher Site | Google Scholar
  50. N. Ribeiro, S. R. Sousa, and F. J. Monteiro, “Influence of crystallite size of nanophased hydroxyapatite on fibronectin and osteonectin adsorption and on MC3T3-E1 osteoblast adhesion and morphology,” Journal of Colloid and Interface Science, vol. 351, no. 2, pp. 398–406, 2010. View at: Publisher Site | Google Scholar
  51. B. Feng, J. Weng, B. C. Yang, S. X. Qu, and X. D. Zhang, “Characterization of surface oxide films on titanium and adhesion of osteoblast,” Biomaterials, vol. 24, no. 25, pp. 4663–4670, 2003. View at: Publisher Site | Google Scholar
  52. J. C. Elliott, Structure and Chemistry of the Apatites and Other Calcium Orthophosphates, vol. 18 of Studies in Inorganic Chemistry, Elsevier, Amsterdam, The Netherlands, 2003.
  53. V. Nelea, I. N. Mihailescu, and M. Jelínek, “Biomaterials: new issues and breakthroughs for biomedical applications,” in Pulsed Laser Deposition of Thin Films: Applications-Led Growth of Functional Materials, R. Eason, Ed., chapter 18, John Wiley & Sons, Hoboken, NJ, USA, 2006. View at: Publisher Site | Google Scholar
  54. S. Maeno, Y. Niki, H. Matsumoto et al., “The effect of calcium ion concentration on osteoblast viability, proliferation and differentiation in monolayer and 3D culture,” Biomaterials, vol. 26, no. 23, pp. 4847–4855, 2005. View at: Publisher Site | Google Scholar
  55. G. R. Beck Jr., E. C. Sullivan, E. Moran, and B. Zerler, “Relationship between alkaline phosphatase levels, osteopontin expression, and mineralization in differentiating MC3T3-E1 osteoblasts,” Journal of Cellular Biochemistry, vol. 68, no. 2, pp. 269–280, 1998. View at: Publisher Site | Google Scholar
  56. G. R. Beck Jr., “Inorganic phosphate as a signaling molecule in osteoblast differentiation,” Journal of Cellular Biochemistry, vol. 90, no. 2, pp. 234–243, 2003. View at: Publisher Site | Google Scholar
  57. R. Roskoski Jr., “ERK1/2 MAP kinases: structure, function, and regulation,” Pharmacological Research, vol. 66, no. 2, pp. 105–143, 2012. View at: Publisher Site | Google Scholar

Copyright © 2016 Yardnapar Parcharoen et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

More related articles

2073 Views | 888 Downloads | 10 Citations
 PDF  Download Citation  Citation
 Download other formatsMore
 Order printed copiesOrder

Related articles

We are committed to sharing findings related to COVID-19 as quickly and safely as possible. Any author submitting a COVID-19 paper should notify us at to ensure their research is fast-tracked and made available on a preprint server as soon as possible. We will be providing unlimited waivers of publication charges for accepted articles related to COVID-19. Sign up here as a reviewer to help fast-track new submissions.