Materials Science and Engineering C 33 (2013) 1654–1661

Contents lists available at SciVerse ScienceDirect

Materials Science and Engineering C

j ourna l homepage: www.e lsev ie r .com/ locate /msec

Chemical characterization of hydroxyapatite obtained by wet chemistry in thepresence of V, Co, and Cu ions

Claus Moseke a,⁎, Michael Gelinsky b, Jürgen Groll a, Uwe Gbureck a

a Department for Functional Materials in Medicine and Dentistry, University of Würzburg, Pleicherwall 2, 97070 Würzburg, Germanyb Centre for Translational Bone, Joint and Soft Tissue Research, University of Dresden, Fetscherstr. 74, 01307 Dresden, Germany

⁎ Corresponding author. Tel.: +49 931 201 73710; faE-mail address: claus.moseke@fmz.uni-wuerzburg.d

0928-4931/$ – see front matter © 2012 Elsevier B.V. Allhttp://dx.doi.org/10.1016/j.msec.2012.12.075

a b s t r a c t

a r t i c l e i n f o

Article history:Received 15 November 2012Accepted 20 December 2012Available online 27 December 2012

Keywords:MineralizationApatitic phaseAmorphous calcium phosphateTricalcium phosphate

Amodel system for the precipitation of hydroxyapatite (HA) from saturated solutions at basic pH was utilizedto investigate the effects of V, Co, and Cu ions on crystallography and stoichiometry of the produced apatites.X-ray diffraction (XRD) was applied to analyze phase composition and crystallinity of powders obtained withdifferent metal ion concentrations and annealed at different sintering temperatures. This procedure used thetemperature-dependent phase transitions and decompositions of calcium phosphates to analyze the partic-ular influences of the metal ions on apatite mineralization. Comparative XRD measurements showed thatall metal ion species reduced crystallinity and crystallite size of the produced apatites. Furthermore the trans-formation of amorphous calcium phosphate (ACP) to HA was partially inhibited, as was deduced from theformation of α-tricalcium phosphate (α-TCP) peaks in XRD patterns of the heated powders as well as fromthe reduced intensity of the OH stretch vibration in FTIR spectra. The thermally induced formation ofβ-TCP indicated a significantly reduced Ca/P ratio as compared to stoichiometric HA. This effect was morepronounced with rising metal ion content. In addition, the appearance of metal oxides in the XRD patternsof samples heated to higher temperatures indicated the incorporation of metal ions in the precipitated apa-tites. Peak shifts showed that both the apatitic as well as the β-TCP phase apparently had incorporated metalions.

© 2012 Elsevier B.V. All rights reserved.

1. Introduction

Recent approaches to equip materials for their application as or-thopedic implants with bioactive and osteointegrative propertiesare based on the modification with low doses of inorganic bioactiveions, which provides the possibility of a controlled manipulationand support of implant-specific tissue reactions [1,2]. Many metalions are present as low concentrated trace elements in the humanbody and play an essential role in biological processes, e.g. the regula-tion of cell functions, the activation or inhibition of enzymaticprocesses [3] or the triggered expression of specific proteins [4–6].Vanadium, for example, supports osteogenesis by mitogenetic effectson osteoblasts and the inhibition of osteoclast formation [7,8] and hasalso been proven to show antitumoural activity [9]. However, theosteoclast inhibition must be regarded as a sensitive parameter, asosteoclasts also play a major role in the initialization of bone-remodeling. Copper promotes angiogenesis [10,11] and osteoblast ac-tivity [12] and has also antimicrobial properties in the μM range,whilst cobalt also shows angiogenetic effects [13–15] and stimulatesosteoclast formation in concentrations of 0.01 to 1 μM [16,17]. An

x: +49 931 201 73500.e (C. Moseke).

rights reserved.

application of these ions in subtoxic concentrations excludes detri-mental effects on cells such as mutagenic and carcinogenic effects[18]. In comparison to organic molecules metal ions have a lot of ad-vantages: they can be stored for practically unlimited time, are easilysterilized, and last but not least they are inexpensive and easy tohandle. Studies carried out so far merely concentrated on metal ionmodified polymeric biomaterials, whilst ceramic materials based oncalcium phosphates have been modified almost exclusively with sili-con, magnesium, strontium or zinc ions [19–23]. However, when therelease of metal ions from an orthopedic implant or a bone fillingcement into adjacent tissue occurs, possible effects on the formationof new bone tissue have to be taken into account.

Bone mineralization is a complex process which is mainly initiatedby the interaction of osteoblasts and collagen fibrils. The osteoblasts se-crete matrix vesicles containing calcium complexes of phospholipids,basic proteins and alkaline phosphatase. After bursting of the vesiclemembrane the calcium phosphate crystallites formed inside act as ex-tracellular nucleation seeds, with the actual precipitation starting atthe collagen fibrils serving as the organic template [24]. The mineraliza-tion process may be understood as a chemical precipitation from anaqueous solution locally oversaturated with Ca2+ and PO4

3− ions,followed by transformation of calcium phosphate to HA crystallites,which finally align along the collagen fibrils to form a solid but elasticcomposite material of inorganic bone mineral and collagen.

1655C. Moseke et al. / Materials Science and Engineering C 33 (2013) 1654–1661

To understand these processes Blumenthal et al. studied severalprecipitation reactions for the synthesis of HA to examine howbiomineralization is regulated by molecules occurring in blood plasma[25]. Pyrophosphate ions (P2O7

4−) and adenosine triphosphate (ATP)contain P–O–P bounds and can significantly reduce the formation rateof HA. Depending on its concentration ATP inhibits respectively delaysthe transformation of amorphous calciumphosphate (ACP) toHA. It is as-sumed that the molecules adsorb to the newly formed HA seeds bymeans of the P–O–P bounds and hence inhibit further crystal growth.Of particular interest for the study on hand were Blumenthal's experi-ments with the application of metal ions instead of inhibiting biomole-cules [25,26], which showed inhibiting effects on HA formation indifferent mineralization model systems. Aluminium ions for exampleact in a similar way like ATPmolecules and repress crystal growth by ad-sorption to active growth sites of the initially formed apatite seeds[25,27]. Experiments with titanium and vanadium ions which were car-ried out only for the direct crystallization of HAwith andwithout the ap-plication of seed crystals showed similar results. However, in any case theformation of HA began immediately after mixing of the reactants. Obvi-ously the metal ions only inhibited the progress of crystal growth, butnot the formation of the first crystal seeds. The aim of the present studywas to investigate in more detail the influences of vanadium, cobalt andcopper ions on the mineralization of apatites from oversaturated solu-tions of Ca2+ and PO4

3− ions and – by thermal treatment at differentsintering temperatures – the crystallographic effects on the products.For this purpose a simplified model system for the formation ofphase-pure HAwas established, which thenwas contaminatedwith sev-eral different concentrations of the tested metal ions. The precipitatedproducts were characterized by X-ray diffraction, FTIR spectroscopy andICP-MS analysis with regard to their chemical and crystallographic com-position. XRD was also applied to heat treated samples in order to studyinfluences of themetal ions on thermally induced phase transformations.It should be stated that themetal ion concentrations applied in this studywere much higher than the concentrations that would be expected fromleaching processes or that would be suitable for biomaterial applications;they were chosen in order to obtain significant effects in the measure-ments, which could help to understand the interaction of HAmineraliza-tion with the metal ions at clinically relevant concentrations.

2. Materials and methods

2.1. Mineralization

Hydroxyapatite powders were obtained by precipitation ac-cording to the stoichiometric reaction of calcium nitrate tetrahydrate(Ca(NO3)2·4H2O, Riedel-de Haën, Seelze, Germany) and diammoniumhydrogen phosphate ((NH4)2HPO4), Sigma-Aldrich,Munich, Germany):

5CaðNO3Þ2 þ 3ðNH4Þ2HPO4 þ H2O→Ca5ðPO4Þ3OHþ 6NH4NO3 þ 4HNO3 ð1Þ

Two mother solutions of the reactants were prepared by dissolutionin deionized water in portions of 150 ml each. Solution 1 contained0.025 mol Ca(NO3)2 and solution 2 0.015 mol (NH4)2HPO4. Thus, theconcentrations of the reacting ions in the resulting mixture of solutions1 and 2 were [Ca2+]=83.34 mmol/l and [PO4

3−]=50 mmol/l.The mixing of the solutions was carried out by slow addition of so-

lution 2 to solution 1. This so-called “inverse” precipitation method[28] maintains a Ca/P ratio above 1.67, which promotes the formationof stoichiometric HA [29]. The addition of solution 2 occurred in threeportions of 50 ml each. The first portion was added dropwise, the sec-ond one slowly flowing, and the last one was poured within a fewseconds. Between the three steps the mixture was shaken very care-fully to avoid the destruction of initial crystallization seeds. Immedi-ately after the addition of the third portion of phosphate solution20 ml of 25% ammonia solution (Grüssing GmbH Analytika, Filsum,

Germany) were added to elevate the pH value to 12. This was neces-sary to counteract the pH reduction caused by the formation of nitricacid on the product side of Eq. (1). Furthermore the solubility of HA inaqueous solutions decreases with increasing temperature [30], there-fore the mixture was heated for 10 min with a maximum tempera-ture of about 80 °C and was then allowed to cool down and rest.After 20 h the precipitate was poured through a folded filter andwashed several times to separate the insoluble product from thenon-adsorbed ions remaining in the solution. Then the precipitateswere stored on the filter paper in a furnace at 37 °C until they werecompletely dried. The solid productswere crushed in amortar and divid-ed into portions. One of these portionswas heated separately in unglazedporcelain crucibles in a sintering furnace to 1000 °C. The heating-up took60 min to reach themaximum temperature, whichwas thenmaintainedfor 60 min. After thermal treatment the samples were allowed to cooldown to about 100 °C with closed furnace door, before they weretaken out. Additional heating experiments were carried out for selectedpowder samples at lower temperatures (600 or 900 °C).

For the introduction of V, Co and Cu ions into the model mineraliza-tion system metal salts with high solubility were used, namelyvanadium(III) chloride 99% (VCl3, metals basis, Alfa Aesar, Germany),cobalt(II) nitrate hexahydrate 98% (Co(NO3)·6H2O, Aldrich, Germany)and copper sulphate pentahydrate (CuSO4·5H2O, Merck, Germany).Prior to the mixing of solutions 1 and 2 a certain amount of one of themetal salts was weighed in and added to solution 1. The added concen-trations were calculated in percentage related to the mass of calciumatoms in the expected reaction product according to the stoichiometricprecipitation. This was done under the assumption that the incorpora-tion of metal ions into the HA lattice occurred by substitution of Ca2+

ions. Therefore the amountms of metal salt to be added in order to sub-stitute 1% of Ca2+ ions was calculated by the following equation:

ms ¼Ms

Mi⋅0:01 g ð2Þ

Here Ms and Mi are the molar masses of the metal salt and theconcerning metal ion respectively. Hence the addition of 1% to theprecipitation reaction corresponded to 30.9 mg for VCl3, 49.4 mg forCo(NO3)2·6H2O, and 39.3 mg for CuSO4·5H2O. Based on these values re-actions with various additions of metal salts were carried out, wherebythe percentages of ions were 1, 5, and 10%. The addition of the differentamounts ofmetal salts occurred 10 min beforemixing of the twomothersolutions to ensure complete dissolution of all compounds. The mixingitself and the following treatment of the reaction mixtures were carriedout according to the production of pure HA, as described above.

2.2. X-Ray diffraction

Phase composition of the obtained powders was determined byX-ray diffraction (XRD) in Bragg–Brentano geometry with a SiemensD5005 X-ray diffractometer (Bruker AXS, Karlsruhe, Germany) usingCu-Kα radiation with a voltage of 40 kV and a tube current of 40 mA.Prior to measurement, all samples were finely ground in a mortar andthe resulting powders were filled into powder cuvettes. Diffraction pat-terns were recorded in a 2Θ range from 20 to 40° with a step size of0.02° and scan speeds of 3 s/step. The DiffracPlus evaluation software(Bruker AXS) was utilized for both qualitative and semi-quantitativeanalyses, the latter one being carried out by evaluation of the net areasof the (0 0 2) diffraction peak, which was the only distinct peak in thepatterns of unheated samples, as well as by estimation of crystallitesize according to Scherrer's formula [31].

2.3. Fourier transform infrared spectroscopy

FTIR measurements were performed with a Nicolet iS10 FT-IRspectrometer (Thermo Fisher Scientific Inc., Waltham, MA, USA).

1656 C. Moseke et al. / Materials Science and Engineering C 33 (2013) 1654–1661

Spectra were recorded in attenuated total reflexion (ATR) mode with16 scans each and with a resolution of 4 cm−1. To ensure comparabil-ity of the samples independently of water adsorbed from the labora-tory environment, all powders were stored in a drying chamber at45 °C and taken out only immediately before measurement. Further-more, prior to every single measurement a background spectrum wasrecorded and subtracted from the subsequently obtained samplespectrum.

2.4. ICP-MS analysis

Ion concentrations (V, Co, and Cu) in the dried precipitates weredetermined using inductively-coupled-plasma mass-spectrometry(ICP-MS, Varian, Darmstadt, Germany). The powders were dissolvedin 1 ml nitric acid (Suprapur) and subsequently diluted with ultrapurewater to a total volume of 10 ml. The quantitative measurement wascarried out against standard solutions (Merck, Darmstadt, Germany)containing defined concentrations of all ions of interest.

2.5. Release study

For an estimation of the release behaviour of the metal ionsadsorbed to the precipitated HA powders, unheated samples withthe highest concentration of 10% of V, Co, and Cu as well as pure HAas a reference were immersed in sterile phosphate buffered saline(PBS, 137 mmol/l NaCl, 2.7 mmol/l KCl, 7 mmol/l Na2HPO4×2H2O,1.5 mmol/l KH2PO4). 100 mg of each powder was mixed with 2 mlPBS in centrifuge tubes. These were thoroughly shaken and placedon an orbital shaker which was operated at 100 rpm. After 1 h, 3 h,1 d, 3 d, and 6 d the tubes were centrifuged for 5 min with4500 rpm. The liquid supernatants were detracted and stored in sep-arate sample flasks, whilst the remaining precipitate was providedwith fresh PBS and put back on to the orbital shaker. Finally the con-centrations in the elution liquids were determined by ICP-MS.

3. Results and discussion

3.1. Metal-free model mineralization

Fig. 1 shows the XRD patterns of HA powder obtained by precipita-tion in the model system without the addition of metal ions. One frac-tion of the sample was measured after drying at 37 °C, whilst theother three fractions were heated for 1 h to 600 °C, 900 °C, and1000 °C respectively. The unheated fraction and the fraction heated at

Fig. 1. XRD patterns of precipitates obtained without metal ions after drying at 37 °Cand heating for 1 h at 600, 900, and 1000 °C respectively.

600 °C showed the typical peak broadening of HA with low crystallitesize. Due to the high peak width adjacent peaks could hardly be distin-guished. Heating at 900 °C increased crystallite size significantly andallowed the identification of all peaks listed in the powder diffractionfile (PDF) from the JCPDS data base for synthetic HA (PDF-No. 09-432)[32]. Non-stoichiometric apatite that may contain noticeable concentra-tions of CO3

2− or HCO3− ions – occasionally referred to as basic calcium

phosphate (bCaP) in literature [33] – decomposes to stoichiometric HA(s-HA) and β-tricalcium phosphate by prolonged heating at tempera-tures above 900 °C according to the following reaction [34]:

bCaP→x Ca10ðPO4Þ6ðOHÞ2 þ y β� Ca3ðPO4Þ2 þ water þ CO2 ð3Þ

The Ca/P ratio in the bCaP can be calculated to (10 x/y+3)/(6 x/y+2)). As also the heating at 1000 °C did not result in theformation of β-TCP peaks in the XRD pattern, the concerningprecipitates could be considered as phase-pure HA with a Ca/P ratio ofabout 1.67, a value which was verified by determination of the Ca andP concentrations in dissolved powder samples using ICP-MS analysis.However, in comparison to the patterns obtained from samples heatedat only 900 °C a few very small peaks appeared, which could be attribut-ed to α-TCP. The observation of this phase below its usual transitiontemperature of 1125 °C indicated the presence of small amounts ofremaining ACP in the unheated powders, which was thermallytransformed during heating [35].

Fig. 2 shows the various solid phases that may occur during theexperiments with the model HA synthesis reaction in the unheatedpowders and in the samples treated at different heating tempera-tures. It also illustrates the possible pathways on which phases canbe formed by crystallization and thermal transformation from othersand where the influence of metal ions in the starting material canaffect these. This simplified scheme may be utilized to extrapolatethe properties of calcium phosphates precipitated in the presence ofmetal ions by evaluation of XRD patterns that can be obtained fromthe analysis of thermally treated powders but not from unheatedsamples.

3.2. Mineralization in the presence of V, Co, and Cu

Fig. 3 shows the results of the ICP-MS analysis of the powders pre-cipitated with different additions of metal ions. Apparently the affin-ity of V and Co to the solid reaction product was high even after thethorough rinsing procedure, whilst only a relatively small fraction of

Fig. 2. In vitro mineralization of HA in the basic model system and the phases formedby thermal treatment for 1 h at the denoted temperatures. Black and white arrowsmark the pathways of phase formation in the metal-free system; white arrow path-ways are directly inhibited by metal ions. The grey arrows lead to phases which onlyappear in the presence of certain metal ion species.

Fig. 3. Metal ion to calcium ratio in the dried precipitates, determined by ICP-MS mea-surement of samples dissolved in nitric acid, vs. the relative calcium content in the re-action mixture (in relation to calcium mass provided in the starting materials).

1657C. Moseke et al. / Materials Science and Engineering C 33 (2013) 1654–1661

the added copper could be detected in the rinsed powders. Besidesthe concentrations of the metal ions also the relative concentrationsof Ca and P in the dissolved powders were determined to calculatethe Ca/P ratios. For V and Co these values varied in a broad rangefrom 1.2 to 1.75 without an apparent correlation with the amountof metal added to the reaction mixture. Only for Cu, all determinedCa/P ratios were found in a range of 1.04 to 1.35, which was wellbelow the value for stoichiometric HA. These strong deviations maypartially be explained by the presence of untransformed ACP in theprecipitates. The composition of ACP, which is often found as an HAprecursor in aqueous precipitations, is given by the indefinite formulaCax(PO4)y·nH2O. Its CaP ratio is very sensitive to Ca and P concentra-tions in the mother liquid and the pH value and can vary in a widerange from 1.18 up to 2.5 [36], hence its presence in the precipitatesmay lead to total Ca/P ratios below as well as above the stoichiometricvalue for HA. In the following the different effects on the precipitationreaction as well as on the behaviour of the resulting reaction productsunder thermal treatment are presented and discussed for each of thetested metal ions.

3.3. Vanadium

The effects of V addition on the model synthesis of HA alreadyshowed in the reaction vessel a significant reduction of the reactionproduct volume with rising V content. Instantly with the start of theprecipitation the initial brown colour of the mixture faded as well inthe precipitate as in the aqueous supernatant. After 20 h rest therinsed precipitate and the obtained filtrate were totally decolourized,which suggests that all V3+ ions had participated in the reaction tothe insoluble product. In fact there are various pentavalent vanadiumcompounds that are colourless at basic pH value, but these do notoccur in aqueous solutions [37]. When the drying process was com-pleted, all powder samples showed a light grey colour that bright-ened up by thermal treatment and turned to pale yellow at 900 °Cfor V contents ≥5%. The XRD patterns obtained from unheated pow-der samples did not allow any specific phase analysis due to the lowcrystallite size and the resulting peak overlap. Nevertheless the ob-servable area in the 2Θ range from 32 to 38° mainly consisting ofthe peaks referring to the (211), (112), (300) and (202) lattice planesof HA decreased noticeably with rising V content in the reaction mix-ture (Fig. 4b).

This effect could also be found in the patterns that were heated for1 h at 600 °C (not shown here), although this thermal treatment didnot significantly enhance crystallite size. However, at high V contents(5 and 10%) distinct peaks appeared in the XRD patterns of samplesheated at 600 °C that could be attributed to the formation of β-TCP.Heating up to 900 and 1000 °C led to the formation of small fractionsof α-TCP even for the lowest V content of 1%. With rising V contentβ-TCP also appeared in fast increasing fractions and practically elim-inated the HA phase for a V content of 10% (Fig. 4b). In parallel theaverage crystallite size of the β-TCP phase increased with rising Vcontents, which was not only indicated by the decreasing peakwidth in the XRD patterns, but also by a considerable increase ofhardness, particularly for samples heated at 1000 °C. At the highestV addition of 10% the sample volume decreased rapidly after thermaltreatment. In addition the colour of the powder changed to a dark yel-low with brown parts. This indicated the formation of an additionalphase, which could be detected in the XRD patterns and was identi-fied as calcium vanadate (Ca(VO3)2, PDF-No. 73-0971). As this com-pound is pale-yellow in the hydrated form and red-brown in thedehydrated state [38], the thermally induced change in sample colourcan be attributed to the formation of hydrated Ca(VO3)2 at 900 °C anddehydration of this phase at 1000 °C.

The examination of the peak positions in the diffraction patternsof the apatitic phase did not show any measurable deviation whencompared to metal-free apatites. If V ions were incorporated intothe HA lattice, they were most probably transferred to the β-TCPphase during thermal decomposition. Actually, some samples showedslight peak shifts for the formed β-TCP in comparison to the purewhitlockite structure, indicating a marginal lattice distortion. The oc-currence of Ca(VO3)2 in the samples that contained 10% V and wereheated at 1000 °C, leads to the assumption that at this temperaturethe incorporated V ions migrated to the surface and reacted with ae-rial oxygen. FTIR analysis of unheated powders revealed a significant re-duction of the OH-stretch vibration at awavenumber of 3572 cm−1. Forthe highest V content of 10% the stretch vibrational signalwas complete-ly extinguished (Fig. 5a), which supports the assumption that V ionsinhibited the transformation of ACP to HA.

3.4. Cobalt

The addition of cobalt nitrate to the calcium nitrate solution led toan intensive red colouring. The precipitate forming after the additionof phosphate solution consisted of flaky agglomerates and changed itscolour to violet after ammonia addition and heating. After filteringand drying the samples showed an increasingly powdery consistencewith rising cobalt addition, indicating a cobalt induced reduction ofthe crystallinity respectively increasing nanocrystallinity. This wassupported by the reduced intensities detected in the XRD patternsof unheated powders, which are shown in Fig. 6a. With rising Co con-tent a proportionally increasing background could be detected, due tothe X-ray fluorescence of the incorporated Co excited by the Cu-Kαradiation of the X ray tube. This proved the affinity of the Co ions tothe rinsed solid reaction product and was in accordance with the re-sults from the ICP-MS measurements (Fig. 3). However, a significantline broadening with increasing Co content could be observed afterbackground subtraction, indicating a reduction of crystallite size. Sim-ilar to the results of the experiments with V addition, a suppression ofthe OH-stretch vibration could be found in the FTIR spectra of theunheated powders, which increased with rising Co content (Fig. 5b).

Thermal treatment at 900 resp. 1000 °C of the samples againresulted in the formation of additional phases. In the samples with aCo content of 1% small amounts of α-TCP could be detected afterheating at 900 °C. With rising Co content resp. heating temperatureanother phase appeared with a diffraction pattern very similar toβ-TCP. The significant peak shifts at high Co content and respectivecomparisons with patterns from the JCPDS data base led to the

Fig. 4. X-ray diffraction patterns of calcium phosphate powders precipitated with the addition of 1, 5, and 10 wt.% vanadium to the reaction mixture. a: unheated; b: heated for 1 hat 1000 °C. The most prominent peaks of the occurring phases have been marked as follows: O: hydroxyapatite, +: β-TCP, Δ: α-TCP, #: Ca(VO3)2.

1658 C. Moseke et al. / Materials Science and Engineering C 33 (2013) 1654–1661

assumption that the new phase consisted of calcium cobalt phosphate(CCP, Ca19Co2(PO4)14), a violet salt with the same rhomboedric crys-tal structure and similar lattice parameters as β-TCP. The chemicalformula corresponds to 7 formula units of β-TCP with 2 Ca atomsreplaced by Co atoms. The diffraction patterns of β-TCP and CCP arevirtually identical and can only be distinguished by means of a smallpeak shift due to the smaller ionic radius of Co2+ as compared toCa2+, which results in decreased lattice parameters. As in mostcases the measured peak positions were found somewhere betweenthe values taken from the JCPDS data of β-TCP and CCP, it may be as-sumed that the additional phase consisted of a mixed phase with thechemical formula:

Ca19þxCo2−x PO4ð Þ14;0≤x≤2: ð4Þ

However, a clear correlation between peak positions and Co contentcould not be observed. The fact that the Co ions which adsorbed to theprecipitate in the reaction vessel appeared as an incorporated fractionof the β-TCP formed in the thermally induced decomposition leads tothe assumption that the ions show a preferential affinity to thenon-stoichiometric phase in the unheated powder rather than to theapatitic phase. As part of the more soluble fraction of the precipitatedpowder the Co ions may dissolve much more easily than other metalions which are preferentially incorporated into the apatitic phase.

3.5. Copper

When copper sulphate was added to the model reaction, the mostobvious difference to the experiments with V and Co additions wasthe fact that the supernatant of the reaction mixture maintained an

Fig. 5. FTIR spectra for the OH-stretch vibration at 3572 cm−1 of unheated p

intensive blue colour even after aging overnight, whilst the colourof the precipitate was white for a Cu content of 1% and only lightblue for the highest Cu addition of 10%. Apparently only a small frac-tion of the Cu ions from the mother solution was firmly adsorbed tothe precipitate, which was also supported by the results of theICP-MS measurements (Fig. 3). Also the OH-stretch vibration sup-pression in the FTIR spectra was much less pronounced than for thesame amounts of V or Co ions in the reaction mixture. XRD measure-ments of unheated powders showed only marginal reductions ofcrystallite size and crystallinity, as the diffraction patterns shown inFig. 7a suggest.

To further illustrate this, Fig. 8 presents an overview of the crystal-lite sizes and crystallinities with respect to the line width (Fig. 8a)and the net area (Fig. 8b) of the (0 0 2) HA peak of samples preparedwith the different concentrations of all three ion species. Apparentlythe effect of Cu on the crystallographic properties of the precipitatewas considerably lower than for the other metals. Thermal treatmentof the samples led to the formation of comparatively small amountsof β-TCP, as the XRD patterns show (Fig. 7b). This would indicatethat the stoichiometry of the apatitic phase in the precipitates shouldonly marginally deviate from 1.67. Surprisingly, the Ca/P ratio calcu-lated from ICP-MS analyses of dissolved Cu containing powders wasconsiderably reduced even for low Cu concentrations, as was men-tioned before. Up to now, this apparent contradiction could not beexplained; however it is possible that surface-adsorbed Cu ionsmay have an influence on the thermal behaviour respectively thedecomposition of the phases appearing in the aqueous precipita-tion reaction.

Fig. 9 shows the cumulative release curves determined by immer-sion of powders precipitated with 10% of V, Co and Cu ions in PBS. To

owders with different metal contents a: vanadium, b: cobalt, c: copper.

Fig. 6. X-ray diffraction patterns of calcium phosphate powders precipitated with the addition of 1, 5, and 10 wt.% cobalt to the reaction mixture. a: unheated; b: heated for 1 h at1000 °C. The most prominent peaks of the occurring phases have been marked as follows: O: hydroxyapatite, #: calcium cobalt phosphate, Δ: Co3O4.

1659C. Moseke et al. / Materials Science and Engineering C 33 (2013) 1654–1661

show the three curves in one diagram in order to compare the releasekinetics, the cumulative concentrations of Co and Cu had to be multi-plied by the factors 100 and 1000. After 144 h of immersion about40% of the vanadium had been released from the calcium phosphatepowder, whilst after the same period only 0.12% of Co had been re-leased and the total amount of detected Cu was virtually negligiblewith only 0.002%. Apparently V ions can be eluted from doped calci-um phosphate over a long period, whilst Co and Cu show a high affin-ity to the powders they are incorporated in. This has to be taken intoaccount in the design of experiments with HA cements doped withthese ions for the controlled release in therapeutically relevantdoses. However, more experiments will be conducted to explorealternative methods of the introduction of metal ions into HA cementpowders. In particular, thorough long-term release studies in variouselution media (e.g. cell culture medium and fetal calf serum) will benecessary to provide for more information on the behaviour of theseions in contact to the physiological environment.

4. Conclusions

The introduction of V, Co, and Cu ions into the model system forthe precipitation of HA from an aqueous solution significantly

Fig. 7. X-ray diffraction patterns of calcium phosphate powders precipitated with the additi1000 °C. The most prominent peaks of the occurring phases have been marked as follows:

influenced the phase composition as well as the crystallographicstructure of the reaction products. Crystallinity and crystallite sizewere affected by all three ion species, as indicated by diminishedpeak intensities and peak broadening in the diffraction patterns ofmetal-containing powders. The reduced crystallinity was attributedto the inhibition of ACP-to-HA transformation, whilst the formationof large crystallites was apparently restrained by the adsorption ofmetal ions to active growth sites. Furthermore, the presence of in-creasing amounts of β-TCP with rising metal ion contents in the pow-ders heated at high temperatures was a strong indication for thepreferred formation of non-stoichiometric HA in the unheated pre-cipitates. Whilst V and Co ions had significant effects on the precipi-tates obtained even in the presence of low metal concentrations, theinfluences of Cu were much weaker, as far as they were determinedby crystallographic analysis. These results for the in vitro mineraliza-tion of HA strongly suggest that metal ions may take part in thebiomineralization of HA when emitted from a medical implant andthus influence solubility and degradability of newly formed bonemineral. As the ions are incorporated into the lattice respectivelyadsorbed to the surface of mineralized apatites, they may stay innewly formed bone tissue for long periods and maintain their in-fluence during various bone remodelling cycles. Whilst the total

on of 1, 5, and 10 wt.% copper to the reaction mixture. a: unheated; b: heated for 1 h atO: hydroxyapatite, +: β-TCP.

Fig. 8. Semi-quantitative analysis of unheated calcium phosphate powders precipitated with the addition of different metal contents in the reaction mixture. a: crystallite size or-thogonal to the (0 0 2) lattice plane, calculated with Scherrer's equation from the (0 0 2) peak width, b: net area of the (0 0 2) peak.

1660 C. Moseke et al. / Materials Science and Engineering C 33 (2013) 1654–1661

concentrations of released metal ions in the volume around metalimplants may be expected to be much lower than in this basic re-search study, effects of localized high concentrations may have tobe taken into account. On the other hand, the results provide valu-able information about possible side-effects that may occur in thebioactive modification of HA based bone replacement materials,which are loaded with subtoxic additions of biologically relevantmetal ions. In this case, the addition of V and Co ions to the ce-ment powder will most probably lead to localized increased solu-bilities of the precipitation products and to increasing settingtimes, and will possibly deteriorate the mechanical properties ofthe set cements. Due to the unexpected behaviour of Cu in theheating experiments its actual influence on cement reactions isdifficult to predict. Hence this ion species will be in the focus offorthcoming studies, particularly regarding the possible effects ofCu introduction into the starting materials of cement systems un-dergoing sintering treatment at high temperatures.

Fig. 9. Cumulative release of V, Co, and Cu ions from unheated powder samples. Con-centrations are related to the total amount of metal ions adsorbed to the precipitatedcalcium phosphate. In order to show all three release curves in one diagram, the con-centrations of Co and Cu have been upscaled by factors of 100 respectively 1000.

Acknowledgements

The authors would like to acknowledge the financial support fromthe Deutsche Forschungsgemeinschaft (DFG) for the projects GB 1/12-1and Ge1133/13-1.

References

[1] C. Ergun, T.J. Webster, R. Bizios, R.H. Doremus, J. Biomed. Mater. Res. 59 (2002)312.

[2] C.M. Serre, M. Papillard, P. Chavassieux, J.C. Voegel, G. Boivin, J. Biomed. Mater.Res. 42 (1998) 626.

[3] W. Kaim, B. Schwederski, Bioanorganische Chemie. Zur Funktion chemischerElemente in Lebensprozessen, Teubner, Wiesbaden 2005.

[4] J.D. Bumgardner, L.C. Lucas, J. Dent. Res. 74 (1995) 1521.[5] A. Schedle, P. Samorapoompichit, X.H. Rausch-Fan, A. Franz, W. Fureder, W.R.

Sperr, W. Sperr, A. Ellinger, R. Slavicek, G. Boltz-Nitulescu, P. Valent, J. Dent. Res.74 (1995) 1513.

[6] M.D. Waters, D.E. Gardner, C. Aranyl, D.L. Coffin, Environ. Res. 9 (1975) 32.[7] A.V. Rousselle, D. Heymann, V. Demais, C. Charrier, N. Passuti, M.F. Baslé, Histol.

Histopathol. 17 (2002) 1025.[8] A.M. Cortizo, M.S. Molinuevo, D.A. Barrio, L. Bruzzone, Int. J. Biochem. 38 (2006)

1171.[9] D.A. Barrio, S.B. Etcheverry, Can. J. Physiol. Pharmacol. 84 (2007) 677.

[10] R. Barbucci, A. Magnani, Macromol. Symp. 156 (2000) 239.[11] E.D. Harris, Nutr. Rev. 62 (2004) 60.[12] C. Käppel, U. Gbureck, E. Vorndran, C. Moseke, A. Ewald, J. Biomed. Mater. Res.

100A (2012) 2392.[13] H.H. Kim, S.E. Lee, W.J. Chung, Y. Choi, K. Kwack, S.W. Kim, M.S. Kim, H. Park, Z.H.

Lee, Cytokine 17 (2002) 14.[14] A. Loboda, A. Jazwa, B. Wegiel, A. Jozkowicz, J. Dulak, Cell. Mol. Biol. 51 (2005)

347.[15] T. Tanaka, I. Kojima, T. Ohse, J.R. Ingelfinger, S. Adler, T. Fujita, M. Nangaku, Lab.

Invest. 85 (2005) 1292.[16] Y. Niki, H. Matsumoto, Y. Suda, T. Otani, K. Fujikawa, Y. Toyama, N. Hisamori, A.

Nozue, Biomaterials 24 (2003) 1447.[17] S. Patntirapong, P. Habibovic, P.V. Hauschka, Biomaterials 30 (2009) 548.[18] Z. Nackerdien, K.S. Kasprzak, G. Rao, B. Halliwell, M. Dizdarogul, Cancer Res. 51

(1991) 5837.[19] W. Paul, C.P. Sharma, J. Mater. Sci. Mater. Med. 18 (2007) 699.[20] A. Bandyopadhyay, S. Bernard, W.C. Xue, S. Bose, J. Am. Ceram. Soc. 89 (2006)

2675.[21] A. Ito, H. Kawamura, M. Otsuka, M. Ikeuchi, H. Ohgushi, K. Ishikawa, K. Onuma, N.

Kanzaki, Y. Sogo, N. Ichinose, Mater. Sci. Eng. C22 (2002) 21.[22] Y.Z. Wan, Y. Huang, F. He, Y.L. Wang, Z.G. Zhao, H.F. Ding, Surf. Coat. Technol. 201

(2006) 2904.[23] W. Xue, J.L. Moore, H.L. Hosick, S. Bose, A. Bandyopadhyay, W.W. Lu, K.M.C.

Cheung, K.D.K. Luk, J. Biomed. Mater. Res. 79A (2006) 804.[24] S. Marks, S.T. Popoff, Am. J. Anat. 183 (1988) 1.[25] N.C. Blumenthal, A.S. Posner, Colloid Surf. 26 (1987) 123.[26] N.C. Blumenthal, V. Cosma, J. Biomed. Mater. Res. 23A (1989) 13.[27] I. Mayer, H. Cohen, J.C. Voegel, F.J.G. Cuisinier, J. Cryst. Growth 172 (1997) 219.[28] M. Vignoles, G. Bonel, D.W. Holcomb, R.A. Young, Calcif. Tissue Int. 43 (1988) 33.

1661C. Moseke et al. / Materials Science and Engineering C 33 (2013) 1654–1661

[29] Y. Song, H.H. Hahn, E. Hoffmann, Proceedings of the 10th Gothenburg Symposium,Sweden, IWA Publishing, London, 2002, p. 349.

[30] M. Okido, K. Kuroda, M. Ishikawa, R. Ichino, O. Takai, Solid State Ionics 151 (2002) 47.[31] S.N. Danilchenko, O.G. Kukharenko, C. Moseke, I.Y. Protsenko, L.F. Sukhodub, B.

Sulkio-Cleff, Cryst. Res. Technol. 37 (2002) 1234.[32] JCPDS, International Center For Diffraction Data: Powder Diffraction File 1–40,

1990.[33] J.C. Elliott, Structures and chemistry of the apatites and other calcium orthophos-

phates, Elsevier, Amsterdam-London-New York-Tokyo 1994.

[34] J. Barralet, J.C. Knowles, S. Best, W. Bonfield, J. Mater. Sci. Mater. Med. 13 (2002)529.

[35] E.D. Eanes, Calcif. Tissue Res. 5 (1970) 133.[36] S.V. Dorozhkin, M. Epple, Angew. Chem. Int. Ed. 41 (2002) 3130.[37] H.P. Latscha, H.A. Klein, Anorganische Chemie, 6th edition Springer, Berlin-

Heidelberg-New York-Tokyo 1994.[38] R.J. Meyer, E. Pietsch, Gmelins Handbuch der anorganischen Chemie–Vanadium–

Teil B–Lieferung 2–Band 48, Chemie GmbH, Weinheim/Bergstraße 1973.