Acta Biomaterialia 15 (2015) 254–265

Contents lists available at ScienceDirect

Acta Biomaterialia

journal homepage: www.elsevier .com/locate /ac tabiomat

Microstructure, mechanical properties, castability and in vitrobiocompatibility of Ti–Bi alloys developed for dental applications

http://dx.doi.org/10.1016/j.actbio.2015.01.0091742-7061/� 2015 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.

⇑ Corresponding author at: Department of Materials Science and Engineering,College of Engineering, Peking University, Beijing 100871, China. Tel./fax: +86 106276 7411.

E-mail address: yfzheng@pku.edu.cn (Y.F. Zheng).1 Co-corresponding author.

K.J. Qiu a, Y. Liu b, F.Y. Zhou a, B.L. Wang a, L. Li a, Y.F. Zheng a,b,⇑, Y.H. Liu c,1

a Center for Biomedical Materials and Engineering, Harbin Engineering University, Harbin 150001, Chinab Department of Materials Science and Engineering, College of Engineering, Peking University, Beijing 100871, Chinac General Dental Department, School of Stomatology, Peking University, Beijing 100081, China

a r t i c l e i n f o

Article history:Received 4 October 2014Received in revised form 19 December 2014Accepted 2 January 2015Available online 14 January 2015

Keywords:Ti–Bi alloysMechanical propertiesCastabilityCorrosion resistanceBiocompatibility

a b s t r a c t

In this study, the microstructure, mechanical properties, castability, electrochemical behaviors, cytotox-icity and hemocompatibility of Ti–Bi alloys with pure Ti as control were systematically investigated toassess their potential applications in the dental field. The experimental results showed that, except forthe Ti–20Bi alloy, the microstructure of all other Ti–Bi alloys exhibit single a-Ti phase, while Ti–20Bi alloyis consisted of mainly a-Ti phase and a small amount of BiTi2 and BiTi3 phases. The tensile strength, hard-ness and wear resistance of Ti–Bi alloys were demonstrated to be improved monotonically with theincrease of Bi content. The castability test showed that Ti–2Bi alloy increased the castability of pure Tiby 11.7%. The studied Ti–Bi alloys showed better corrosion resistance than pure Ti in both AS (artificialsaliva) and ASFL (AS containing 0.2% NaF and 0.3% lactic acid) solutions. The concentrations of both Tiion and Bi ion released from Ti–Bi alloys are extremely low in AS, ASF (AS containing 0.2% NaF) andASL (AS containing 0.3% lactic acid) solutions. However, in ASFL solution, a large number of Ti and Bi ionsare released. In addition, Ti–Bi alloys produced no significant deleterious effect to L929 cells and MG63cells, similar to pure Ti, indicating a good in vitro biocompatibility. Besides, both L929 and MG63 cellsperform excellent cell adhesion ability on Ti–Bi alloys. The hemolysis test exhibited that Ti–Bi alloys havean ultra-low hemolysis percentage below 1% and are considered nonhemolytic. To sum up, the Ti–2Bialloy exhibits the optimal comprehensive performance and has great potential for dental applications.

� 2015 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.

1. Introduction

The clinical trials of titanium (Ti) can be traced back to 1965 [1],when Brånemark et al. developed the titanium dental implant andsimultaneously introduced the concept of osseointegration. Sincethen, pure Ti and Ti alloys have been typically used in medicalapplications, especially in orthopedic implants and dentalimplants. Compared with other traditional metallic biomaterials,such as Co–Cr alloys and stainless steels, the Ti alloys possesslower modulus, superior biocompatibility and enhanced corrosionresistance [2]. These attractive performances promote the develop-ment and application of new orthopedic Ti alloys in the medicalfield.

Due to the low strength and the poor wear resistance, pure Ti islargely restricted in many medical applications. Therefore, Ti–6Al–4V and Ti–6Al–7Nb alloys have been introduced into the biomed-ical industry for their increased mechanical strength and enhancedwear resistance [3–5]. However, as reported in many research andclinical use, the long-term performance of Ti–6Al–4V and Ti–6Al–7Nb alloys has raised great concerns due to the release of Al and/or V ions [6–11]. Besides, as another promising biomedical mate-rial, Ti–Ni alloys have often been used as orthodontic wires andintravascular stents for their unique properties such as the super-elasticity and shape memory effect [12,13]. Although Ti–Ni alloyshave been stated to be safety for orthodontic wires [14,15], therelease of Ni ions and their potential toxic effects to tissues havebeen reported in other applications [16].

Therefore, except for the above mentioned alloys, there havebeen developed a large number of new binary Ti alloys in responseto those concerns in recent years [17–30]. In these binary Ti alloys,the addition of alloying elements such as Mo, Nb, Ta, Zr and Hf aregenerally used to enhance the strength and lower the elastic mod-ulus of Ti alloys. The addition of noble metal elements such as Au,

K.J. Qiu et al. / Acta Biomaterialia 15 (2015) 254–265 255

Ag, Pt and Pd mainly focus on improving the corrosion resistance ofTi alloys, and the fusible alloying elements Ga, Ge, In and Sn aremore of emphasizing the casting procedure of Ti alloys. Addition-ally, and importantly, all of these alloying elements reveal a goodbiocompatibility.

Like fusible alloying elements, bismuth (Bi) has a low meltingpoint (271 �C). Meanwhile, Bi possesses one of the lowest valuesof thermal conductivity among metals, 7.97 W m�1 K�1, which ismuch lower than that of Ti (21.9 W m�1 K�1) [31,32]. Actually ithas been much desired by the dentists and patients that the dentalmaterials have low thermal conductivities as the low thermal con-ductivity would protect the dental pulp from caloric stimulationand alleviate the patients’ pain to a great degree. With respect tothe effect of Bi on alloys’ castability, a few previous studies demon-strated that a minor amount (1 wt.%) of Bi could improve the cast-ability of Ti and Ti alloys significantly in a copper mold with aneedle-shaped cavity [33–35]. Moreover, Bi is one of the low toxicheavy metals and has minimum threat to the environment, and themedicinal application of Bi compounds has a history of over250 years [36,37]. Bi compounds were primarily used clinicallyfor the treatment of ulcers, and they were also considered to haveanti-cancer activity [38–40]. Ten years ago, Webster et al. [41]studied the osteoblast response to hydroxyapatite doped withmetal cations (in forms of 5 wt.% nitrate), the results demonstratedthat Bi3+ enhanced the long-term calcium-containing mineraldeposition of osteoblasts more than other cations. Thus, it is spec-ulated that Bi3+ release may have a positive influence on osteogen-esis. However, in some case reports, overdose of colloidal bismuthsubcitrate might increase Bi content in blood and cause nephrotox-icity, but fortunately this nephrotoxicity is reversible after drugwithdrawal in most cases [42,43]. Cengiz et al. [42] reported thata 16-year-old girl had tried to commit suicide by taking 60 tabletsof De-Nol (colloidal bismuth subcitrate 18 g) and 10 days later shehad been admitted to a hospital for treatment. The patient’s renalfunction gradually returned to normal over 64 days with her serumbismuth level decreased from 495 lg/L (12 days after overdose ofingestion) to 260 lg/L (64 days after ingestion). Islek et al. [43]reported another case that a 2-year-old boy had ingested 28 tabletsof De-Nol (colloidal bismuth subcitrate 8.4 g) and 6 h later he hadbeen admitted to a hospital for treatment. After 60 days, his bloodbismuth level had fallen to 96 lg/L, which was consistent with the16-year-old girl’s case. Therefore, considering the potential risk byaccumulation of Bi ions, the Bi ions released from Bi-containingalloys should be detected and given attention. In addition, as devel-oping for dental applications which might contact with blood inuse, the in vitro hemocompatibility is always needed to beevaluated.

Based on the above, by combining that the pharmacologicalfunctions of Bi with the biocompatibility of Ti, the study of Ti–Bialloys for dental application is promising. Therefore, in the presentstudy, four Ti–Bi binary alloys were designed and fabricated, withtheir castability being evaluated in a different way, as well asmicrostructure, mechanical properties, corrosion resistance,in vitro cytotoxicity and hemocompatibility being investigated tostudy their feasibility as dental applications, with pure Ti and Ti–6Al–4V alloy as controls.

2. Materials and methods

2.1. Alloys preparation

The as-cast pure Ti and Ti–Bi alloys with nominal compositions(Ti–2Bi, Ti–5Bi, Ti–10Bi and Ti–20Bi, wt.%) were prepared fromsponge Ti and Bi (both 99.5% in purity) in a non-consumable arcmelting furnace under an Ar atmosphere. Each ingot was re-melted

six times for homogenization. Commercial Ti–6Al–4V alloy was re-melted under the same condition for mechanical properties, cast-ability and electrochemical corrosion tests. The chemical composi-tions of the resulting Ti–Bi alloys were detected by energydispersive spectrum (EDS, Hitachi S-4800 SEM, Japan). The mea-sured Bi contents for Ti–2Bi, Ti–5Bi, Ti–10Bi, and Ti–20Bi alloysare 1.99(0.25), 3.98(0.54), 7.80(0.20), and 17.68(0.38) wt.%, respec-tively, values in parenthesis represent the standard error (hereinaf-ter the same). Different specimens were cut by electro-dischargemachining for various tests.

2.2. Microstructure characterization

The microstructure of Ti–Bi alloys was examined using an opti-cal microscope (OM, Olympus BX51M, Japan), after being polishedand etched via a standard metallographic procedure. The etchingsolution is a mixture of HF, HNO3 and H2O (5%:15%:80% in volume).An X-ray diffractometer (XRD, Rigaku DMAX 2400, Japan) with a Nifiltered Cu Ka radiation was employed to identify the phase consti-tution of the experimental Ti–Bi alloys.

2.3. Mechanical properties tests

The strip specimens (40 � 3 � 2 mm3) of Ti–Bi alloys were pre-pared for uniaxial tensile test, with pure Ti and Ti–6Al–4V alloy ascontrols. The tensile test was performed with an initial strain rateof 5 � 10�4 s�1 on a universal testing machine (Instron5969, USA)at room temperature. Young’s modulus was determined accordingto ASTM E111–97 [44], calculated from the ratio between the ten-sile stress and the corresponding strain up to the proportion limitof the alloys. Five duplicate specimens were tested for each alloy.The hardness of Ti–Bi alloys was measured on a digital microhard-ness tester (Shimadzu HMV-2T, Japan) at a load of 200 g for 15 s,repeating eight times in different positions for each alloy.

The wear test for Ti–Bi alloys was performed at room tempera-ture using a ball-on-flat tribometer, in dry condition and lubricatedcondition with artificial saliva (AS), respectively. The compositionof AS can be seen in Ref. [45]. The 15 � 15 � 1 mm3 square sheetspecimens were used in the wear test, with the commerciallyavailable silicon nitride ceramic ball (Si3N4, U 6 mm) as the matingball. The grinding ball ran in circles with a diameter of 6 mm, thefrequency of 2 Hz and the loading force of 4 N, so as to simulatethe wear conditions of human teeth [46]. Each specimen under-went the wear test for 0.5 h and then the wear tracks wereobserved by SEM (Hitachi S-4800 SEM, Japan). The weight losseswere recorded using an electric balance (Sartorius CP225D, Ger-many). Three duplicate specimens were tested for each alloy.

2.4. Castability test

The castability of Ti–Bi alloys was conducted by modified Whit-lock’s method [47], with pure Ti and Ti–6Al–4V alloy as controls.Details of this method can be found in our previous study [30].The schematic view of mesh-pattern wax mold used in this studyis shown in Fig. 5(a). A pure Ti casting system (SYMBION CAST, Nis-sin Dental Products INC., Japan) was used for casting the alloys andan argon pressure of 1.5 Pa was maintained during the casting.After casting, the castability value was obtained by using Whit-lock’s formula, as shown in Eq. (1). Here the number of total waxmold segments equals 200, and the cast segment is deemed ascomplete or incomplete according to the previous work [47]:

Castability value ð%Þ ¼ No: of complete cast segmentsNo: of total wax mold segments

� 100% ð1Þ

256 K.J. Qiu et al. / Acta Biomaterialia 15 (2015) 254–265

2.5. Electrochemical measurements

The electrochemical measurements of Ti–Bi alloys were con-ducted on an electrochemical working station (CHI660C, Chenhua,China) at 37 �C. Two kinds of electrolytes were prepared from theanalytic grade agents and de-ionized water. One is AS, while theother is ASFL (AS containing 0.2% NaF and 0.3% lactic acid, all addi-tives were in weight percentage). The specimens with an exposedarea of 1 cm2, in turn, were grounded, polished and ultrasonicallywashed. In the test, a platinum counter electrode and a saturatedcalomel electrode (SCE) reference electrode were used. Theopen-circuit potential (OCP) of each specimen was continuouslymonitored for 2 h in electrolytes. Afterward, the potentiodynamicpolarization test was measured up to 2.0 V with a scan rate of1 mV s�1. Corrosion parameters including corrosion potential (Ecorr)and corrosion current density (Icorr) can be estimated from thepolarization curves by Tafel analysis based on the polarization plots.After the potentiodynamic polarization test in ASFL, the surface cor-rosion products were detected by XRD for preliminary analysis.

2.6. Immersion test

The immersion test was carried out according to ASTM-G31-72[48]. Four kinds of solutions were prepared: (1) AS, (2) ASF (AS con-taining 0.2% NaF), (3) ASL (AS containing 0.3% lactic acid) and (4)ASFL. The pH values of these four solutions were 5.8, 5.8, 2.5 and4.0, respectively. In order to prevent dental caries effectively, fluo-rides have been widely added into toothpastes and mouthwash.Lactic acid was selected to resemble the pH of extremely acidicconditions, such as exposure to acidic beverages or regurgitationor the presence of dentobacterial plaque. Bacteria in active dentalplaque generate a powerful, tooth mineral destroying mixture oflactic, acetic and other metabolic acids, at a pH of 4.0 or lower[49]. The plate specimens of Ti–Bi alloys with the total exposedsurface area of 2.4 cm2, were immersed in 48 ml test solutionsfor 14 days at 37 �C. Concentrations of metal ions were detectedby inductively coupled plasma atomic emission spectrometry (Lee-man, Profile ICP-AES). Surface constitution of experimental speci-mens after immersion test was characterized by XRD.

2.7. Cytotoxicity test

The cytotoxicity test was carried out according to ISO 10993-5:2009 [50]. Murine fibroblast cells (L929) and human osteo-blast-like cells (MG63) were adopted to evaluate the cytotoxicityof Ti–Bi alloys by both indirect and direct cell assays. L929 andMG63 cells were cultured in Dulbecco’s modified eagle’s medium(DMEM) and minimum essential medium (MEM), respectively, ina humidified atmosphere with 5% CO2 at 37 �C. Both mediumswere supplemented with 10% fetal bovine serum (FBS), 100 U ml�1

penicillin and 100 lg ml�1 streptomycin.As for the indirect cell assay, extracts were prepared using a

serum-free medium as the extraction medium. The extraction ratiois 3 cm2/ml and the extraction was conducted for 72 h at 37 �C. Cellculture medium was used as a negative control and cell culturemedium containing 10% dimethylsulfoxide (DMSO) as a positivecontrol. Cells were seeded in 96-well plates at a density of 5 � 103 -cells per 100 ll medium and incubated for 24 h to allow attach-ment. Then the cell culture mediums were substituted byextracts, and incubated for 1, 2 and 4 days, respectively. After eachculture period, 10 ll 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltet-razolium bromide (MTT) was added into each well for 4 h of incu-bation. Then 100 ll formazan solubilization solution (10% sodiumdodecyl sulfate (SDS) in 0.01 M HCl) was added into each wellovernight in the incubator. The spectrophotometric absorbance of

the product in each well was measured with a microplate reader(Bio-RAD680) at 570 nm with a reference wavelength of 630 nm.

In the direct cell assay, L929 and MG63 cells with an initial den-sity of 5 � 104 cells/ml were seeded on the sterilized sheet speci-mens in 24-well plates and then cultured for 1, 2 and 4 days,respectively. After each culture period, the specimens with cellswere rinsed gently with phosphate buffered saline (PBS) for threetimes and fixed in 2.5% glutaraldehyde for 2 h, and then dehydratedin a gradient ethanol/distilled water mixture (50%, 60%, 70%, 80%,90% and 100%) for 10 min each, dried in the air. Cell attachmentand morphologies were observed by SEM after Au sputtering.

To measure alkaline phosphatase (ALP) activity, MG63 cellswere cultured in extracts for 7 days with an initial density of5 � 103 cells per 100 ll medium in 96-well plates. After the cultureperiod, the culture medium in each well was removed and 100 llof 1% Triton X-100 was added to obtain the cell lysates. ALP activitywas determined based on the principle that ALP hydrolyzes thephenylphosphate to phenol and phosphate at pH 10; the phenolreacted with 4-aminoantipyrine in the presence of potassium fer-ricyanide to form a red-colored chinone compound, and this absor-bance is proportional to the ALP activity. The reaction proceededfor 15 min at 37 �C and the absorbance of products was measuredat 545 nm using a microplate reader (Bio-RAD680).

2.8. Hemolysis test

Healthy human blood from a volunteer containing sodium cit-rate (3.8 wt.%) in the ratio of 9:1 was taken and diluted with nor-mal saline (4:5 ratio by volume). Ti–Bi alloys specimens(10 � 10 � 1 mm3) were dipped in a centrifuge tube containing10 ml of normal saline that were previously incubated at 37 �Cfor 30 min. Then 0.2 ml of diluted blood was added to the centri-fuge tube and the mixtures were incubated for 1 h at 37 �C. Simi-larly, normal saline solution was used as a negative control anddeionized water as a positive control. Then all the tubes were cen-trifuged for 5 min at 3000 rpm and the supernatant was carefullyremoved and transferred to the 96-well cell culture plates for spec-troscopic analysis. The optical density (OD) was measured by amicroplate reader (Bio-RAD680) at 545 nm. At last, the hemolysiswas calculated according to the Eq. (2):

Hemolysis ð%Þ ¼ ODðtest groupÞ�ODðnegative controlÞODðpositive controlÞ�ODðnegative controlÞ�100% ð2Þ

2.9. Platelet adhesion

Platelet-rich plasma (PRP) was prepared by centrifuging thewhole blood for 10 min at a rate of 1000 rpm/min. The PRP wasoverlaid atop specimens and incubated at 37 �C for 1 h. Sampleswere rinsed with PBS to remove the nonadherent platelets. Theadhered platelets were fixed in 2.5% glutaraldehyde solutions for1 h and dehydrated in a gradient ethanol/distilled water mixture(50%, 60%, 70%, 80%, 90%, 100%) for 10 min each and finally driedin air. Platelets number and morphologies were observed by SEM.Different fields were randomly counted and values were expressedas the average number of adhered platelets per mm2 of surface.

2.10. Statistical analysis

Statistical analysis was performed with SPSS 18.0 for Windowssoftware (SPSS Inc., Chicago, USA). All data were statistically ana-lyzed using a one-way analysis of variance (ANOVA), followed bythe Tukey post hoc tests. A p-value <0.05 was considered statisti-cally significant difference with pure Ti and the Ti–6Al–4V alloy,as indicated by an asterisk (⁄) and a pound sign (#) in the relevanttable and figures, respectively.

Fig. 2. XRD patterns of pure Ti and Ti–Bi alloys at room temperature.

K.J. Qiu et al. / Acta Biomaterialia 15 (2015) 254–265 257

3. Results

3.1. Microstructure of Ti–Bi alloys

The optical micrographs of Ti–Bi alloys are shown in Fig. 1. Ti–Bialloys exhibit a typical laminar microstructure, with lath martens-ites with an average thickness of �5 lm, which can be clearlyobserved in the Ti–5Bi alloy at high magnification. In addition, withthe increase of Bi content, Ti–Bi alloys gradually possess equiaxialgrains and their grain size becomes larger. Fig. 2 presents the XRDpatterns of Ti–Bi alloys at room temperature. Except for the Ti–20Bi alloy, all other Ti–Bi alloys show a single a-Ti phase with hex-agonal close packed (hcp) crystal structure. Compared with pure Tiand other Ti–Bi alloys, the intensity of a-Ti phase’s peaks becomesweaker and a few BiTi2 and BiTi3 phases’ peaks are detectable inthe Ti–20Bi alloy.

3.2. Mechanical properties of Ti–Bi alloys

3.2.1. Tensile properties and microhardnessFig. 3 shows the tensile properties and microhardness of Ti–Bi

alloys at room temperature, with pure Ti and Ti–6Al–4V alloy ascontrols. As seen in Fig. 3(a), compared with pure Ti, Bi additionimproves the yield strength (YS) and ultimate tensile strength(UTS) of Ti–Bi alloys, but weakens the elongation. This effectbecomes more distinct with the gradual increase of Bi content.Ti–10Bi alloy exhibits the optimal tensile property among all Ti–Bi alloys. When compared with Ti–6Al–4V alloy, Ti–10Bi alloyexhibits a much higher elongation but a significant lower strength.Besides, the microhardness of Ti–Bi alloys also increases monoto-nously with the Bi content, as shown in Fig. 3(b). Ti–20Bi alloy evenshows a significant higher microhardness than Ti–6Al–4V alloy.Young’s moduli of pure Ti, Ti–6Al–4V, Ti–2Bi, Ti–5Bi, Ti–10Bi andTi–20Bi alloys were measured accurately using a strain gaugeattached on the specimen, and their values are 103.1(1.6),113.7(2.0), 102.0(2.3), 100.1(2.4), 97.9(1.8) and 94.3(1.9) GPa,respectively. The Young’s moduli of Ti–Bi alloys are slightly lowerthan pure Ti. However, compared with the Ti–6Al–4V alloy, theydeclined significantly.

3.2.2. Wear propertiesThe weight losses of Ti–Bi alloys after sliding wear test in dry

condition and lubricated condition with AS are shown in

Fig. 1. Optical micrographs of pure Ti and Ti–Bi alloys: (a) pu

Fig. 4(a). It is obvious that weight loss decreases significantly withthe increase of Bi content in Ti–Bi alloys under both dry conditionand AS lubricated condition. Fig. 4(b) illustrates the typical weartracks of pure Ti and Ti–10Bi alloy specimens. The wear tracks ofother Ti–Bi alloys exhibit similar features with the Ti–10Bi alloy,hence not all of them are displayed here. It can be seen that aftersliding wear, the surface morphology of pure Ti is rougher thanthe Ti–10Bi alloy, and there are some turn-off areas and micro-plowing on the surface of pure Ti.

3.3. Castability of Ti–Bi alloys

Fig. 5 shows the castability of Ti–Bi alloys. The images of mesh-pattern castings of experimental pure Ti, Ti–6Al–4V alloy and Ti–Bialloys are shown in Fig. 5(b). It is obvious that castings of Ti–2Bialloy specimens have more complete casting segments than theothers. Fig. 5(c) lists their castability values, which of pure Ti, Ti–6Al–4V, Ti–2Bi, Ti–5Bi, Ti–10Bi and Ti–20Bi alloys are 79.7(3.1)%,76(3.2)%, 89.0(1.3)%, 82.2(6.3)%, 76.3(10.1)% and 75.0(6.1)%,respectively. Among them, the Ti–2Bi alloy displays the optimumcasting performance, its castability value is significantly higherthan that of both pure Ti and the Ti–6Al–4V alloy.

re Ti, (b) Ti–2Bi, (c) Ti–5Bi, (d) Ti–10Bi and (e) Ti–20Bi.

Fig. 3. Mechanical properties of pure Ti, Ti–6Al–4V alloy and Ti–Bi alloys at room temperature, (a) tensile properties and (b) microhardness.

Fig. 4. (a) Weight losses and (b) wear tracks of pure Ti and Ti–Bi alloys after sliding wear test in dry condition and lubricated condition.

258 K.J. Qiu et al. / Acta Biomaterialia 15 (2015) 254–265

3.4. Corrosion behavior of Ti–Bi alloys

3.4.1. Electrochemical corrosion behaviorFig. 6 shows the OCP and potentiodynamic polarization curves

of pure Ti, Ti–6Al–4V alloy and Ti–Bi alloys in AS and ASFL solu-tions. OCP is the potential of the working electrode relative tothe reference electrode without any applied current, reflectingthe thermodynamic equilibrium at the interface of metal and solu-tion as a function of time. As seen in Fig. 6(a), the OCP curves of allTi–Bi alloys take on an upward trend and become stable graduallyin AS solution. However, in ASFL solution, the OCP curves of all Ti–Bi alloys as well as pure Ti and Ti–6Al–4V alloy decline sharply atthe initial stage and are then stable at a low potential. With moreBi content in Ti–Bi alloys, more time is taken to reach a stablepotential at the initial stage and the more noble potentials are sta-bilized after 2 h immersion. In addition, the Ti–Bi alloys have morenoble OCP values than pure Ti and Ti–6Al–4V alloy in both solu-tions after immersion for 2 h. By comparing the two solutions,the addition of fluoride and lactic acid aggravates the corrosionbehaviors of Ti–Bi alloys seriously. After 2 h OCP tests, the poten-tiodynamic polarization curves were measured in AS and ASFLsolutions, as shown in Fig. 6(b). The Ti–Bi alloys exhibit polariza-tion behaviors similar to pure Ti and show broad passivationregions in AS solution. However, in the ASFL solution, their polari-zation behaviors are very different from pure Ti at the anodecurves, Ti–2Bi and Ti–5Bi alloys show similar trends with theincrease of scanning potential, while Ti–10Bi and Ti–20Bi alloysdisplay a much lower current density than others at the anodecurves. That is to say, Bi addition greatly improves the corrosionresistance of Ti–Bi alloys at the extreme conditions with both fluo-

ride and lactic acid. Besides, compared with the Ti–6Al–4V alloy, allTi–Bi alloys exhibit better corrosion resistance in both AS and ASFLsolutions. Their corrosion parameters are listed in Table 1.

3.4.2. Ion releaseFig. 7 shows the ion concentrations of Ti and Bi released from

pure Ti and Ti–Bi alloys after immersion in four solutions for14 days. As seen in Fig. 7(a). Ti ion concentration exhibits no signif-icant difference in AS, ASL, and ASFL solutions, respectively. How-ever, in ASF solution, with the increase of Bi content in Ti–Bi alloys,Ti ion concentration declines significantly, which connotes a betterresistance to fluoride. As shown in Fig. 7(b), Bi ion concentrationincreases slightly with the increase of Bi content in Ti–Bi alloysin AS, ASF, and ASL solutions, respectively. However, in ASFL solu-tion, the Bi ion concentration varies with an anomalous trend andreduces slightly with the increase of Bi content in Ti–Bi alloys.Besides, the concentrations of both Ti ion and Bi ion in ASFL solu-tion are much higher than those of in the other three kinds ofsolutions.

The corroded surfaces of pure Ti and Ti–Bi alloys after potentio-dynamic polarization and immersion test for 14 days in ASFL solu-tion have also been analyzed with XRD, as shown in Fig. 8. Afterpotentiodynamic polarization, a mixed layer of Ti compoundsand Bi compounds (such as TiO2, TiF2, Bi2O3 and BiCl3) has beendemonstrated to form on the surface of Ti–Bi alloys, especially onthe Ti–20Bi alloy surface, as shown in Fig. 8(a). However, afterimmersion test in ASFL solution for 14 days, there are only a fewTiO2 and TiF2 compounds detected on the surface of Ti–Bi alloys,which can be seen in Fig. 8(b).

Fig. 5. (a) Schematic view of mesh-pattern wax mold; (b) images of mesh-pattern castings and (c) castability value of pure Ti, Ti–6Al–4V alloy and Ti–Bi alloys.

Fig. 6. (a) OCP and (b) potentiodynamic polarization curves of pure Ti, Ti–6Al–4V alloy and Ti–Bi alloys tested in AS and ASFL solutions.

Table 1Corrosion parameters of pure Ti, Ti–6Al–4V alloy and Ti–Bi alloys obtained from electrochemical measurements in AS and ASFL solutions.

Materials OCP (V, vs SCE) Ecorr (V, vs SCE) Icorr (A cm�2)

AS ASFL AS ASFL AS ASFL

Pure Ti �0.312(0.012) �1.095(0.024) �0.489(0.007) �1.023(0.003) 3.56(0.70) � 10�8 5.45(0.30) � 10�4

Ti–6Al–4V �0.328(0.167) �1.157(0.007) �0.468(0.021) �1.151(0.006) 6.81(0.37) � 10�8 6.92(1.04) � 10�4

Ti–2Bi �0.246(0.011)*,# �1.051(0.017)# �0.439(0.004)* �1.015(0.002)# 4.77(0.99) � 10�8# 2.88(0.60) � 10�4*,#

Ti–5Bi �0.295(0.010) �1.045(0.041)# �0.544(0.004)*,# �1.068(0.008)*,# 2.11(0.72) � 10�8# 3.16(0.52) � 10�4*,#

Ti–10Bi �0.317(0.042) �0.992(0.045)*,# �0.556(0.011)*,# �0.986(0.006)*,# 3.62(0.37) � 10�8# 2.59(0.40) � 10�4*,#

Ti–20Bi �0.285(0.031) �0.979(0.024)*,# �0.527(0.009)*,# �0.976(0.007)*,# 5.34(0.63) � 10�8 2.54(0.26) � 10�4*,#

* Indicates the statistically significant difference (p < 0.05) with respect to pure Ti.# Indicates the statistically significant difference (p < 0.05) with respect the Ti–6Al–4V alloy.

K.J. Qiu et al. / Acta Biomaterialia 15 (2015) 254–265 259

Fig. 7. Ion concentrations of (a) Ti and (b) Bi released from pure Ti and Ti–Bi alloys after immersion in four solutions for 14 days.

Fig. 8. XRD patterns of pure Ti and Ti–Bi alloys after (a) potentiodynamic polarization and (b) immersion test for 14 days in ASFL solution.

260 K.J. Qiu et al. / Acta Biomaterialia 15 (2015) 254–265

3.5. Cytotoxicity of Ti–Bi alloys

Fig. 9 shows the relative viability of L929 and MG63 cells cul-tured in extracts of pure Ti and Ti–Bi alloys for 1, 2 and 4 days.As shown in Fig. 9(a), the cell viabilities of L929 cells cultured inTi–Bi alloys extracts are almost the same as in pure Ti after 4 daysexcept for Ti–20Bi alloy. After 4 days of culture, the relative viabil-ities of L929 cells cultured in Ti–Bi alloys extracts vary between84% (Ti–20Bi alloy) and 97% (Ti–2Bi alloy). The statistical analysisindicates no significant difference between Ti–Bi alloys and pureTi, representing an approximate non-toxicity feature, except for

Fig. 9. Cell viability of (a) L929 and (b) MG63 cells cultured i

the Ti–20Bi alloy exhibiting a slight toxicity. However, differingfrom L929 cells, the cell viabilities of MG63 cells in the Ti–Bi alloyextracts are lower than that of in pure Ti after each culture period.After 4 days of culture, the relative viabilities of MG63 cells cul-tured in Ti–Bi alloy extracts vary between 74% (Ti–10Bi alloy)and 88% (Ti–2Bi alloy), which indicates a slight toxicity, as shownin Fig. 9(b).

Fig. 10 illustrates the morphologies of L929 and MG63 cells cul-tured on pure Ti and Ti–10Bi alloy specimens for 1, 2 and 4 days,respectively. The morphologies of L929 and MG63 cells culturedon other Ti–Bi alloys exhibit similar features with the Ti–10Bi

n extracts of pure Ti and Ti–Bi alloys for 1, 2 and 4 days.

Fig. 10. The morphology of L929 and MG63 cells cultured on pure Ti and Ti–10Bi alloys for 1, 2 and 4 days.

Fig. 11. Normalized ALP activities of MG63 cells lysates after being cultured inextracts of pure Ti and Ti–Bi alloys for 7 days.

K.J. Qiu et al. / Acta Biomaterialia 15 (2015) 254–265 261

alloy, hence not all of them are displayed here. After 1 day of cul-ture, there are only a small number of cells attached on both pureTi and Ti–10Bi alloy specimens. The cells shrink slightly and exhibitpoor adhesion, especially for L929 cells on the Ti–10Bi alloy. Withincessant increase of culture time, the cells become denser andspread better. Both L929 and MG63 cells proliferate faster andadhere better on the surface of pure Ti than Ti–10Bi alloy, whichindicates that Bi has a slightly negative effect on cell proliferation.Good correspondence could be found between the direct observa-tion and the indirect cell viability evaluation.

Usually, the differentiated function of osteoblasts could be eval-uated by monitoring ALP activity of the cells. Fig. 11 shows the nor-malized ALP activities of MG63 cells lysates after being cultured in

extracts of pure Ti and Ti–Bi alloys for 7 days. The statistical anal-ysis indicated that there is no significant difference between pureTi and Ti–Bi alloys for ALP activities of MG63 cells. In combinationwith the results of the MTT assay, the conclusion could be drawnthat Ti–Bi alloys did not release toxic components into the med-ium, which inhibited notably the proliferation and weakened dif-ferentiated function of osteoblast-like cells, thereby showinggood, at least acceptable, in vitro cytocompatibility.

3.6. Hemocompatibility of Ti–Bi alloys

Fig. 12 shows the hemolysis percentage of pure Ti and Ti–Bialloys. It can be seen that all Ti–Bi alloys and pure Ti have an ultra-low hemolysis percentage below 1%, which indicates nonhemolyticproperties according to ASTM-F756-00 [51]. The morphologies ofadhered human platelets adhering on pure Ti and Ti–Bi alloys werecharacterized and the SEM images are shown in Fig. 13. The plateletspresent nearly spherical morphology and spread evenly on the sur-face of pure Ti and Ti–Bi alloys, which can be more clearly observedin the Ti–5Bi alloy at high magnification. The number of plateletsadhered on the surface of Ti–Bi alloys varies between1.15 � 104 mm�2 (Ti–5Bi alloy) and 1.44 � 104 mm�2 (Ti–20Bialloy), with the control of 1.36 � 104 mm�2 (pure Ti).

4. Discussion

4.1. Microstructure and mechanical properties

In this study, except for the Ti–20Bi alloy, all other Ti–Bi alloysexhibit a single a-Ti phase and show a typical laminar microstruc-ture at room temperature (as shown in Fig. 2), although the solubil-

Fig. 12. Hemolysis percentage of pure Ti and Ti–Bi alloys.

262 K.J. Qiu et al. / Acta Biomaterialia 15 (2015) 254–265

ity of Bi in a-Ti phase is just 1.8 at.% (7.4 wt.%) [52]. According to theequilibrium phase diagram, a BiTi3 phase should appear in the Ti–10Bi alloy, but there is no detectable trace of it. That is probablydue to the actual composition of Bi in Ti–10Bi alloy, 7.8 wt.%, whichis close to the max solubility of Bi in a-Ti phase. In contrast, a smallnumber of BiTi2 and BiTi3 phases appeared in the Ti–20Bi alloybecause of the high Bi content, but these phases are exactly the rea-son why the Ti–20Bi alloy exhibits such inferior ductility (1.49% oftensile elongation), as shown in Fig. 3(a). Generally speaking,mechanical properties of alloys depend on microstructure, whichis affected by the kind and amount of alloying elements. As seenin Fig. 1, the laminar martensitic structure become clearer and thegrain size grows larger from the Ti–2Bi alloy to the Ti–20Bi alloy.With this change, the Ti–2Bi alloy possesses the improved strengthand comparable elongation with pure Ti, whereas Ti–Bi alloys withmore than 5 wt.% Bi content exhibit much higher strength butweakened elongation than pure Ti. In addition, the improved tensilestrength and hardness of Ti–Bi alloys are largely derived from solidsolution strengthening of the Bi atoms in a-Ti phase. In Ti lattice, thelattice distortion can be easily caused and the motion of original dis-locations can be blocked due to the atomic size mismatch betweenBi and Ti, and therefore improves the strength of Ti–Bi alloys. How-ever, as shown in Fig. 3, Bi does not have a strong strengtheningeffect when its content is lower than 10 wt.%. In general, Ti–10Bialloy exhibits the optimal tensile property with the YS of428(9.5) MPa, UTS of 520(22.6) MPa, and elongation of 15.4(3.8)%,which is ascribed to the combined effect of relatively high contentof Bi and absence of Ti–Bi compounds phase. Although the Ti–

Fig. 13. SEM images of platelets adhering on pure Ti and Ti–Bi alloy

20Bi alloy possesses the highest strength and microhardness, thebrittleness limits it further biomedical applications. Compared withthe as-cast Ti–6Al–4V alloy, the YS and UTS of the Ti–10Bi alloy canreach as high as 60.2% and 61.8% of the Ti–6Al–4V alloy, respec-tively. Besides, the elongation of the as-cast Ti–6Al–4V alloy is just9.2(2.0)% in this study, which is consistent with the results of Nii-nomi et al. [53]. Therefore, the elongation of the Ti–10Bi alloy is169.2% of the Ti–6Al–4V alloy. Compared with as-cast Co–Cr–Moalloys [54], Ti–10Bi alloy is still showing an advantage of YS andelongation. According to ASTM F75-01 [55], the mechanical require-ments of the as-cast Co–28Cr–6Mo alloy (a current clinical dentalalloy for removable partial denture framework) are 450 MPa forYS, 650 MPa for UTS and 8% for elongation. From this perspective,the Ti–10Bi alloy is close to meet the application criteria. In addition,the Ti–10Bi alloy exhibits higher strength than Ti–10Au, Ti–10Ag,Ti–10Nb and Ti–10Hf alloys [25,56–58] but lower strength thanTi–7.5Mo, Ti–10Ge, Ti–10 Ga alloys [28,30,59]. It indicates that theBi element is a modest strengthening alloying element of Ti.

The occurrence of the wear and tear is inevitable when themetallic materials are used in the human body. In this study, thewear resistance of Ti–Bi alloys was tested in both dry and AS solu-tion lubricated conditions. In most cases, the wear resistance isconsidered to depend on hardness [53]. Therefore, the weight lossof Ti–Bi alloys decreases with the increase of Bi content in both dryand lubricated conditions, as shown in Fig. 4(a), which indicatesthat the wear resistance is improved with the increase of Bi con-tent. Fig. 4(b) representatively shows the wear tracks of pure Tiand the Ti–10Bi alloy, and it can be found that the adhesive wearhas occurred obviously, which includes scoring, galling, seizing,and scuffing. As for pure Ti, the severe micro-plowing and turn-off area left on its rough surface, are consistent with the previousstudies [28,30]. Micro-plowing usually occurs when the wear par-ticles break down from the alloys surface, and resulting in the for-mation of grooves on the alloy’s surface. Compared with pure Ti,the wear tracks of the Ti–10Bi alloy look smoother and more reg-ular in both dry and AS lubricated conditions, and there is no obvi-ous turn-off area left on the surface. Thus the Ti–10Bi alloypossesses better wear resistance because of the much higherstrength and microhardness than pure Ti. Besides, in the AS solu-tion lubricated condition, both pure Ti and the Ti–10Bi alloy exhi-bit fewer abrasive particles on their surface, indicating a goodlubricating effect of AS solution to reduce the degree of wear.Moreover, AS solution acts as a corrosive medium while lubricating

s: (a) pure Ti, (b) Ti–2Bi, (c) Ti–5Bi, (d) Ti–10Bi and (e) Ti–20Bi.

K.J. Qiu et al. / Acta Biomaterialia 15 (2015) 254–265 263

the wear surface, which would corrode the materials surface moreor less. That is, when sliding wear occurs in the AS solution, thelubrication and corrosion take place at the same time. In this study,Bi addition slightly improves the corrosion resistance of Ti–Bialloys in AS solution, as seen in Fig. 6. Therefore, Ti–Bi alloys exhi-bit a better wear resistance than pure Ti in AS solution.

4.2. Castability

For metallic biomaterials used in dentistry, the castability is animportant performance to fulfill the clinical use. The castability ofan alloy system is usually associated with its ability to fill the mold,which is affected by many factors, such as mold design, moltenmetal temperature, mold reaction, cooling rate, type of machine,applied pressure, surface tension (surface free energy) and oxidefilm of the alloy [60]. In this study, to assess the effect of Bi addi-tion on the castability of Ti–Bi alloys, the casting process parame-ters are precisely controlled in the tests. Therefore, it is more likelythat the factors affecting castability are mold reaction, surface ten-sion and oxide film of the alloy. In other words, the Ti–Bi alloy withslighter mold reaction, lower surface energy and more stable oxidefilm than pure Ti would have better castability.

As shown in Fig. 5, the Ti–2Bi alloy exhibits the highest castabil-ity value and improves the casting performance of pure Ti signifi-cantly. It is consistent with the previous studies that a minoramount (1–2 wt.%) of some alloying elements could improve thecastability of pure Ti and Ti alloys [30,33,34]. However, with thefurther increase of Bi content, the castability values of Ti–Bi alloysreduce inversely, and even are lower than those of pure Ti. Gener-ally speaking, the castability of dental alloys depends on their flu-idity. The better the metal fluidity, the better the castability.Besides, according to casting theory, the metal fluidity would bereduced by adding alloying elements, thereby the castability is alsodeclined. That is because when alloying elements are added in puremetal, the solidification would no longer occur in the solidificationinterface and the generation of the dendritic structure in the preli-minary stage of the solidification process would make the moltenmetal flow more difficult [34]. This may explain why the castabilityvalue decreases in Ti–10Bi, Ti–20Bi and Ti–6Al–4V alloys.

Another possible affecting factor of castability is that addingalloying elements could change the alloys surface oxide film charac-teristics. Usually the oxide film of the molten metal leads to high sur-face tension. Actually, by comparison, Bi has a much lower surfacetension than Ti [61]. Papworth et al. [62] reported that the additionof small amounts of Bi to Al alloys changed the surface morphologyand alleviated the formation of oxide defects within the casting,which is conducive to castability. Bi improves the castability of theAl alloy by being incorporated into the oxide, which weakens thestrength of the film and reduces the surface tension of the Al alloy.Cheng et al. [33] also proved that the addition of 1 wt.% Bi increasedthe castability value by 35% in a copper mold with a needle-shapedcavity. They believed that the addition of Bi decreased the surfacetension of Ti and thereby improved the castability but the real mech-anisms were not given. In this study, Ti–2Bi and Ti–5Bi alloysimproved the castability, and the reason may be that just asdescribed above. In conclusion, addition of Bi with the content nomore than 5 wt.% is beneficial for improving castability due to thelowered surface tension, while the content more than 10 wt.% wouldbe detrimental to castability based on the casting theory. That isprobably the reason why the castability of Ti–Bi alloys first increasesand then decreases with the increase of Bi content.

4.3. In vitro corrosion behaviors

In comparison with other dental alloys (such as Co–Cr–Moalloy, stainless steels), Ti and its alloys display a more outstanding

corrosion resistance arising from the spontaneous formation ofpassive film in air and aqueous solutions [63,64]. The increasedcorrosion rates are usually undesirable, since elements releasedmay have detrimental effects on local tissues or accumulate else-where in the body [65]. In this study, the corrosion resistance ofTi–Bi alloys improves with the increase of Bi content in both solu-tions, which indicates the addition of Bi helps to form the sponta-neous passive film on the metallic surface which is more stablethermodynamically. Especially in ASFL solution, Bi addition greatlyimproves the corrosion resistance of Ti–Bi alloys at extreme condi-tions with both fluoride and lactic acid, as shown in Fig. 6 andTable 1. Compared with the Ti–6Al–4V alloy, all Ti–Bi alloys exhibitsignificant nobler corrosion potentials and lower corrosion currentdensities in ASFL solution, which indicates that the Ti–Bi alloyspossess better corrosion resistance. Besides, it can be seen inFig. 8(a), after potentiodynamic polarization in ASFL solution, amixed layer of Ti compounds and Bi compounds (such as TiO2,TiF2, Bi2O3 and BiCl3) has been demonstrated to form on the surfaceof Ti–Bi alloys, especially on the Ti–20Bi alloy surface, which mayexplain the improved corrosion resistance of Ti–Bi alloys. Actually,Bi has a much higher standard electrode potential than Ti [66]. Inthe previous studies, the Ti alloys containing noble metal elements(such as Au, Ag, Pd and Pt) had been proven to exhibit an improvedcorrosion resistance [24,26,67]. Due to high corrosion resistance ofadded noble metals, Ti may dissolve preferentially during the ini-tial stage of corrosion and then the noble element atoms of thealloys surface may accumulate because of Ti loss. As a result, thepotential of the alloy became further nobler than the critical poten-tial for passivation of Ti, which improved the corrosion resistanceof alloys. Therefore, the influence of Bi on corrosion resistance ofTi may be analogous to those noble metals. Also, an early study[68] has demonstrated that a minor amount (1 wt.%) of Bi couldincrease the corrosion resistance of Zr.

The high corrosion resistance of Ti–Bi alloys was also proved bythe static immersion test. Concentrations of both the Ti ion and theBi ion are considerably low in AS, ASF and ASL solutions, in com-parison, their concentrations are sharply increased in ASFL solu-tion. Nonetheless, the Bi ion concentrations are below 1 lg/mlfor all Ti–Bi alloys after 14 days of immersion in the harshest envi-ronment (ASFL solution), which is also considered to be extremelylow. Take an implant with 3 cm2 area into consideration, which issufficient for a dental implant surface area, the Ti–2Bi alloy shouldhave a Bi ion release of no more than 3.90 lg/day, and even if allthese 3.90 lg Bi ions dissolve into the blood (for a healthy adult,the blood volume is about 4 L), the blood bismuth level(�0.97 lg/L) is much lower than 12 lg/L (105 days after ingestion,considered as an acceptable value) [43]. In addition, the normaldosage of De-Nol tablets for the treatment of peptic ulcer is nomore than 4 tablets (colloidal bismuth subcitrate 1.2 g, equals toBi2O3 0.48 g) per day [43,69]. From this point of view, the Bi ionsreleased from Ti–2Bi alloy in the ASFL solution is also far lowerthan the daily allowable intake of Bi. The Ti ion concentrationsare no more than 150 lg/ml (3 mg/cm2) in all Ti–Bi alloys, whichis consistent with the previous study [28]. Besides, there are onlya few TiO2 and TiF2 compounds detected on the surface of Ti–Bialloys, indicating a slight corrosion occurred in the static immer-sion test in ASFL solution. In conclusion, the results indicate thatthe Ti–Bi alloys exhibit the same or a better corrosion resistancethan pure Ti in all these simulated saliva solutions.

4.4. In vitro biocompatibility and hemocompatibility

A good biocompatibility with the surrounding tissue is usuallyexpected for biomaterials. Ti is known to be nontoxic and biocom-patible because a stable oxide film forms on its surface, which caneffectively inhibit the release of metal ions. Meanwhile, the oxide

264 K.J. Qiu et al. / Acta Biomaterialia 15 (2015) 254–265

itself shows low solubility and good biocompatibility. The alloyingelement Bi is also recognized as a less toxic heavy metal elementand has minimum threat to the environment [37]. In this study,the cytotoxicity test results showed that both L929 and MG63 cellsexhibited a good cell proliferation and attachment on the surface ofTi–Bi alloys. Besides, Ti–Bi alloys possessed ALP activities with nosignificant difference with pure Ti, which testified their contribu-tion to differentiation of osteoclasts. That is to say, all Ti–Bi alloysexhibited an acceptable in vitro biocompatibility. Besides, it hasbeen demonstrated that the Ti–15Mo–1Bi alloy showed superiorability to retain bone in the in vivo implantation experiment ascompared with the Ti–6Al–4V alloy [70]. After 26 weeks postim-plantation in rabbit femur, the sum of the new bone area aroundthe Ti–15Mo–1Bi alloy is �249% that of the Ti–6Al–4V alloy.Although the Ti–15Mo alloy had not been used as a control, theauthor still considered that the Ti–15Mo–1Bi alloy possessed bet-ter clinical applicability with the addition of the Bi element. Inaddition to that, the Bi ion release is considered to have a positiveinfluence on osteogenesis [41]. Hence, the Ti–Bi alloys may have anexcellent bone-bonding and the bone formation ability when theyare used as dental implants. However, limited studies about in vivoexperiments of Bi-containing alloys have been reported, and there-fore, the in vivo biocompatibility of Ti–Bi alloys requires furtherinvestigation.

Actually, the selection of the hemocompatibility test for theinteraction of biomaterials with the blood should take into consid-eration materials, clinical utility and usage environment. For exam-ple, interactions between blood (or blood component) andendosseous implants may affect the subsequent bone healingevents in the peri-implant healing compartment. Hemolysis cancause the release of hemoglobin, intracellular components andthromboplastic substances from ruptured erythrocytes [71]. Ti–Bialloys did not induce hemolysis since the Ti or Bi ions are almostimpossible to be released within 1 h in normal saline, which is con-sistent with the previous study [71]. In the platelet adhesion test, itcan be seen in the high magnification image of the Ti–5Bi alloy inFig. 13, most platelets are activated in all Ti–Bi alloys, which impliesthey can enhance bone-healing responses. Compared with pure Tiand other Ti–Bi alloys, there are fewer platelets adhered on the sur-face of the Ti–5Bi alloy, since the Ti–5Bi alloy exhibits the smooth-est surface than others. The modulation of platelet activity isrelated to both the alloys’ chemistry and surface topography. Tita-nium oxides are generally anti-hemolytic and are reported to havereduced platelet adhesion [72]. The crystalline titanium oxides hada slightly higher activation of the clotting cascade but lower plateletadhesion than nanocrystalline and amorphous titanium oxides. Inaddition, Mohan et al. [73] reported that the relatively thickerTiO2 layer (150–200 nm) obtained through hydrothermal process-ing in contrast to the native TiO2 layer (5–10 nm) would probablyprovide improved hemocompatibility. The rough topography couldresult in increased surface area, which in turn could lead toincreased plasma protein adsorption, and subsequently more plate-let adhesion and activation [74].

5. Conclusions

In this study, the Ti–Bi alloys with Bi contents of 2, 5, 10 and20 wt.% were investigated for potential biomedical applications.The Ti–Bi alloys with Bi content no more than 10 wt.% consistedof a single a-Ti phase, while Ti–20Bi alloy is composed of mainlya a-Ti phase and a minor amount of BiTi2 and BiTi3 phases. Thestrength, microhardness and wear resistance of Ti–Bi alloysincrease monotonically with the increase of Bi content due to thesolid solution strengthening. Ti–10Bi alloy exhibits the optimalmechanical property. By adding a minor amount of Bi (2 wt.%),the castability of the Ti alloy is improved significantly. Ti–Bi alloys

possess the better corrosion resistance than pure Ti, especially inASFL solution. Cytotoxicity test results show that both L929 andMG63 cells have a good proliferation, differentiation and attach-ment on the surface of Ti–Bi alloys. Besides, Ti–Bi alloys exhibitno nonhemolytic properties, indicating good hemocompatibility.In conclusion, the Ti–2Bi alloy exhibits the best castability, theacceptable mechanical property, the appropriate corrosion resis-tance, and the excellent in vitro biocompatibility, suggesting itspotential application for dental material.

Acknowledgements

This work was supported by the National Basic Research Pro-gram of China (973 Program) (Grant No. 2012CB619102 and2012CB619100), National Science Fund for Distinguished YoungScholars (Grant No. 51225101), National Natural Science Founda-tion of China (Grant Nos. 51431002 and 31170909), the NSFC/RGC Joint Research Scheme under Grant No. 51361165101, BeijingMunicipal Science and Technology Project (Z131100005213002)and Project for Supervisor of Excellent Doctoral Dissertation of Bei-jing (20121000101).

Appendix A. Figures with essential colour discrimination

Certain figures in this article, particularly Figs. 2–9 are difficultto interpret in black and white. The full colour images can be foundin the on-line version, at http://dx.doi.org/10.1016/j.actbio.2015.01.009.

References

[1] Odman J, Lekholm U, Jemt T, Branemark PI, Thilander B. Osseointegratedtitanium implants – a new approach in orthodontic treatment. Eur J Orthodont1988;10:98–105.

[2] Long M, Rack HJ. Titanium alloys in total joint replacement – a materialsscience perspective. Biomaterials 1998;19:1621–39.

[3] Lintner F, Zweymuller K, Brand G. Tissue reactions to titanium endoprostheses.Autopsy studies in four cases. J Arthroplasty 1986;1.

[4] Semlitsch MF, Weber H, Streicher RM, Schön R. Joint replacement componentsmade of hot-forged and surface-treated Ti–6Al–7Nb alloy. Biomaterials1992;13:781–8.

[5] Geetha M, Singh AK, Asokamani R, Gogia AK. Ti based biomaterials, theultimate choice for orthopaedic implants – a review. Prog Mater Sci2009;54:397–425.

[6] Perl DP, Brody AR. Alzheimer’s disease: X-ray spectrometric evidence ofaluminum accumulation in neurofibrillary tangle-bearing neurons. Science1980;208:297–9.

[7] Walker PR, LeBlanc J, Sikorska M. Effects of aluminum and other cations on thestructure of brain and liver chromatin. Biochemistry 1989;28:3911–5.

[8] Wapner KL. Implications of metallic corrosion in total knee arthroplasty. ClinOrthop Relat R 1991:12–20.

[9] Landsberg JP, McDonald B, Watt F. Absence of aluminium in neuritic plaquecores in Alzheimer’s disease. Nature 1992;360:65–8.

[10] Eisenbarth E, Velten D, Muller M, Thull R, Breme J. Biocompatibility of beta-stabilizing elements of titanium alloys. Biomaterials 2004;25:5705–13.

[11] Biesiekierski A, Wang J, Gepreel MA, Wen C. A new look at biomedical Ti-basedshape memory alloys. Acta Biomater 2012;8:1661–9.

[12] Thompson SA. An overview of nickel–titanium alloys used in dentistry. IntEndod J 2000;33:297–310.

[13] Stoeckel D, Pelton A, Duerig T. Self-expanding nitinol stents: material anddesign considerations. Eur Radiol 2004;14:292–301.

[14] Setcos JC, Babaei-Mahani A, Di Silvio L, Mjor LA, Wilson NHF. The safety ofnickel containing dental alloys. Dent Mater 2006;22:1163–8.

[15] Huang HH, Chiu YH, Lee TH, Wu SC, Yang HW, Su KH, et al. Ion release fromNiTi orthodontic wires in artificial saliva with various acidities. Biomaterials2003;24:3585–92.

[16] Heintz C, Riepe G, Birken L, Kaiser E, Chakfe N, Morlock M, et al. Corrodednitinol wires in explanted aortic endografts: an important mechanism offailure? J Endovasc Ther 2001;8:248–53.

[17] Ho WF, Ju CP, Chern Lin JH. Structure and properties of cast binary Ti–Moalloys. Biomaterials 1999;20:2115–22.

[18] Lee CM, Ju CP, Lin JHC. Structure–property relationship of cast Ti–Nb alloys. JOral Rehabil 2002;29:314–22.

[19] Ho WF, Cheng CH, Chen WK, Wu SC, Lin HC, Hsu HC. Evaluation of low-fusingporcelain bonded to dental cast Ti–Zr alloys. J Alloy Compd 2009;471:185–9.

K.J. Qiu et al. / Acta Biomaterialia 15 (2015) 254–265 265

[20] Okazaki Y, Rao S, Ito Y, Tateishi T. Corrosion resistance, mechanical properties,corrosion fatigue strength and cytocompatibility of new Ti alloys without Aland V. Biomaterials 1998;19:1197–215.

[21] Mareci D, Chelariu R, Gordin DM, Ungureanu G, Gloriant T. Comparativecorrosion study of Ti–Ta alloys for dental applications. Acta Biomater2009;5:3625–39.

[22] Cai Z, Koike M, Sato H, Brezner M, Guo Q, Komatsu M, et al. Electrochemicalcharacterization of cast Ti–Hf binary alloys. Acta Biomater 2005;1:353–6.

[23] Hsu HC, Wu SC, Hong YS, Ho WF. Mechanical properties and deformationbehavior of as-cast Ti–Sn alloys. J Alloy Compd 2009;479:390–4.

[24] Zhang BB, Zheng YF, Liu Y. Effect of Ag on the corrosion behavior of Ti–Agalloys in artificial saliva solutions. Dent Mater 2009;25:672–7.

[25] Takahashi M, Kikuchi M, Okuno O. Mechanical properties and grindability ofexperimental Ti–Au alloys. Dent Mater J 2004;23:203–10.

[26] Nakagawa M, Matono Y, Matsuya S, Udoh K, Ishikawa K. The effect of Pt and Pdalloying additions on the corrosion behavior of titanium in fluoride-containingenvironments. Biomaterials 2005;26:2239–46.

[27] Ohkubo C. Wear resistance of experimental Ti–Cu alloys. Biomaterials2003;24:3377–81.

[28] Lin WJ, Wang BL, Qiu KJ, Zhou FY, Li L, Lin JP, et al. Ti–Ge binary alloy systemdeveloped as potential dental materials. J Biomed Mater Res B Appl Biomater2012;100:2239–50.

[29] Wang QY, Wang YB, Lin JP, Zheng YF. Development and properties of Ti–Inbinary alloys as dental biomaterials. Mater Sci Eng C 2013;33:1601–6.

[30] Qiu KJ, Lin WJ, Zhou FY, Nan HQ, Wang BL, Li L, et al. Ti–Ga binary alloysdeveloped as potential dental materials. Mater Sci Eng C 2014;34:474–83.

[31] Bismuth. Available: http://en.wikipedia.org/wiki/Bismuth.[32] Titanium. Available: http://en.wikipedia.org/wiki/Titanium.[33] Cheng WW, Chern Lin JH, Ju CP. Effect of alloying addition on castability and

other properties of titanium. AFS Trans 2001:1–10.[34] Cheng WW, Chern Lin JH, Ju CP. Bismuth effect on castability and mechanical

properties of Ti–6Al–4V alloy cast in copper mold. Mater Lett 2003;57:2591–6.[35] Lin CW, Ju CP, Chern Lin JH. Comparison among mechanical properties of

investment-cast cp Ti, Ti–6Al–7Nb and Ti–15Mo–1Bi alloys. Mater Trans2004;45:3028–32.

[36] Briand GG, Burford N. Bismuth compounds and preparations with biological ormedicinal relevance. ChemInform 1999;30:2601–57.

[37] Udalova T, Logutenko O, Timakova E, Afonina L, Naydenko E, Yukhin YM.Bismuth compounds in medicine. Strategic Technologies, 2008 IFOST 2008Third International Forum on: IEEE; 2008. pp. 137–40.

[38] Kopfmaier P, Klapotke T. Antitumor activity of some organometallicbismuth(III) thiolates. Inorg Chim Acta 1988;152:49–52.

[39] Desoize B. Metals and metal compounds in cancer treatment. Anticancer Res2004;24:1529–44.

[40] Tiekink ER. Antimony and bismuth compounds in oncology. Crit Rev OncolHemat 2002;42:217–24.

[41] Webster TJ, Massa-Schlueter EA, Smith JL, Slamovich EB. Osteoblast responseto hydroxyapatite doped with divalent and trivalent cations. Biomaterials2004;25:2111–21.

[42] Cengiz N, Uslu Y, Gök F, Anarat A. Acute renal failure after overdose of colloidalbismuth subcitrate. Pediatr Nephrol 2005;20:1355–8.

[43] Islek I, Uysal S, Gok F, Dundaroz R, Kucukoduk S. Reversible nephrotoxicityafter overdose of colloidal bismuth subcitrate. Pediatr Nephrol2001;16:510–4.

[44] ASTM-E111-97. Standard test method for Young’s modulus, tangent modulusand chord modulus. Philadelphia, PA, USA: American Society for Testing andMaterials; 2004.

[45] Huang HH. Effect of fluoride and albumin concentration on the corrosionbehavior of Ti–6Al–4V alloy. Biomaterials 2003;24:275–82.

[46] Dowson D. History of tribology. 2nd ed. London: Professional EngineeringPublishing; 1998.

[47] Hinman RW, Tesk JA, Whitlock RP, Parry EE, Durkowski JS. A technique forcharacterizing casting behavior of dental alloys. J Dent Res 1985;64:134–8.

[48] ASTM-G31-72. Standard practice for laboratory immersion corrosion testing ofmetals. Philadelphia, PA, USA: American Society for Testing and Materials;2004.

[49] Zaura E, Ten Cate JM. Dental plaque as a biofilm: a pilot study of the effects ofnutrients on plaque pH and dentin demineralization. Caries Res 2004;38:9–15.

[50] ISO-10993-5. Biological evaluation of medical devices – Part 5: Tests forin vitro cytotoxicity. Arlington, VA: ANSI/AAMI, 2009.

[51] ASTM-E756-00. Standard practice for assessment of hemolytic properties ofmaterials. Philadelphia, PA, USA: American Society for Testing and Materials,2000.

[52] Okamoto H. Bi–Ti (Bismuth–Titanium). J Phase Equilib Diff 2010;31:314–5.[53] Niinomi M. Mechanical biocompatibilities of titanium alloys for biomedical

applications. J Mech Behav Biomed 2008;1:30–42.[54] Yoda K, Suyalatu, Takaichi A, Nomura N, Tsutsumi Y, Doi H. Effects of

chromium and nitrogen content on the microstructures and mechanicalproperties of as-cast Co–Cr–Mo alloys for dental applications. Acta Biomater2012;8:2856–62.

[55] ASTM-F75-01. Standard specification for cobalt-28 chromium-6 molybdenumalloy castings and casting alloy for surgical implants. Philadelphia, PA, USA:American Society for Testing and Materials, 2001.

[56] Takahashi M, Kikuchi M, Takada Y, Okuno O. Mechanical properties andmicrostructures of dental cast Ti–Ag and Ti–Cu alloys. Dent Mater J2002;21:270–80.

[57] Kikuchi M, Takahashi M, Okuno O. Mechanical properties and grindability ofdental cast Ti–Nb alloys. Dent Mater J 2003;22:328–42.

[58] Sato H, Kikuchi M, Komatsu M, Okuno O, Okabe T. Mechanical properties ofcast Ti–Hf alloys. J Biomed Mater Res B 2005;72:362–7.

[59] Ho WF. A comparison of tensile properties and corrosion behavior of cast Ti–7.5Mo with c.p. Ti, Ti–15Mo and Ti–6Al–4V alloys. J Alloy Compd2008;464:580–3.

[60] Oliveira PC, Adabo GL, Ribeiro RF, Rocha SS. The effect of mold temperature oncastability of CP Ti and Ti–6Al–4V castings into phosphate bonded investmentmaterials. Dent Mater 2006;22:1098–102.

[61] Keene BJ. Review of data for the surface-tension of pure metals. Int Mater Rev1993;38:157–92.

[62] Papworth A, Fox P. The disruption of oxide defects within aluminium alloycastings by the addition of bismuth. Mater Lett 1998;35:202–6.

[63] Karimi S, Nickchi T, Alfantazi A. Effects of bovine serum albumin on thecorrosion behaviour of AISI 316L, Co–28Cr–6Mo, and Ti–6Al–4V alloys inphosphate buffered saline solutions. Corros Sci 2011;53:3262–72.

[64] Niinomi M. Recent metallic materials for biomedical applications. MetallMater Trans A 2002;33:477–86.

[65] Khan MA, Williams RL, Williams DF. In-vitro corrosion and wear of titaniumalloys in the biological environment. Biomaterials 1996;17:2117–26.

[66] Vanysek P. Electrochemical series. 86th ed. Boca Raton, FL: CRC Press; 2005.[67] Nakagawa M, Matsuya S, Udoh K. Corrosion behavior of pure titanium and

titanium alloys in fluoride-containing solutions. Dent Mater J 2001;20:305–14.[68] Zhou FY, Qiu KJ, Li HF, Huang T, Wang BL, Li L, et al. Screening on binary Zr–1X

(X=Ti, Nb, Mo, Cu, Au, Pd, Ag, Ru, Hf and Bi) alloys with good in vitrocytocompatibility and magnetic resonance imaging compatibility. ActaBiomater 2013;9:9578–87.

[69] Dore MP, Graham DY, Mele R, Marras L, Nieddu S, Manca A, et al. Colloidalbismuth subcitrate-based twice-a-day quadruple therapy as primary orsalvage therapy for Helicobacter pylori infection. Am J Gastroenterol2002;97:857–60.

[70] Lee JW, Lin DJ, Ju CP, Yin HS, Chuang CC, Chern Lin JH. In-vitro and in-vivoevaluation of a new Ti–15Mo–1Bi alloy. J Biomed Mater Res B2009;91:643–50.

[71] Wang BL, Li L, Zheng YF. In vitro cytotoxicity and hemocompatibility studies ofTi–Nb, Ti–Nb–Zr and Ti–Nb–Hf biomedical shape memory alloys. BiomedMater 2010;5:044102.

[72] Maitz MF, Pham MT, Wieser E, Tsyganov I. Blood compatibility of titaniumoxides with various crystal structure and element doping. J Biomater Appl2003;17:303–19.

[73] Mohan CC, Chennazhi KP, Menon D. In vitro hemocompatibility and vascularendothelial cell functionality on titania nanostructures under static anddynamic conditions for improved coronary stenting applications. ActaBiomater 2013;9:9568–77.

[74] Park JY, Gemmell CH, Davies JE. Platelet interactions with titanium:modulation of platelet activity by surface topography. Biomaterials2001;22:2671–82.