Tricalcium Phosphate: an X-Ray Diﬀraction, Solid State NMR and ATR-FTIR Study

Journal of Functional Biomaterials

Article Strontium and Zinc Substitution in β-Tricalcium Phosphate: An X-ray Diﬀraction, Solid State NMR and ATR-FTIR Study

Elisa Boanini 1 , Massimo Gazzano 2,* , Carlo Nervi 3 , Michele R. Chierotti 3 , Katia Rubini 1, Roberto Gobetto 3 and Adriana Bigi 1

1 Department of Chemistry “Giacomo Ciamician”, Alma Mater Studiorum-University of Bologna, 40126 Bologna, Italy; [email protected] (E.B.); [email protected] (K.R.); [email protected] (A.B.) 2 ISOF-CNR, via Gobetti 101, 40129 Bologna, Italy 3 Department of Chemistry, University of Torino, via P. Giuria 7, 10125 Torino, Italy; [email protected] (C.N.); [email protected] (M.R.C.); [email protected] (R.G.) * Correspondence: [email protected]; Tel.: +39-051-2099552

Received: 27 March 2019; Accepted: 28 April 2019; Published: 5 May 2019

Abstract: β-tricalcium phosphate (β-TCP) is one of the most common bioceramics, widely applied in bone cements and implants. Herein we synthesized β-TCP by solid state reaction in the presence of increasing amounts of two biologically active ions, namely strontium and zinc, in order to clarify the structural modifications induced by ionic substitution. The results of X-ray diffraction analysis indicate that zinc can substitute for calcium into a β-TCP structure up to about 10 at% inducing a reduction of the cell parameters, whereas the substitution occurs up to about 80 at% in the case of strontium, which provokes a linear increase of the lattice constants, and a slight modification into a more symmetric structure. Rietveld refinements and solid-state 31P NMR spectra demonstrate that the octahedral Ca(5) is the site of β-TCP preferred by the small zinc ion. ATR-FTIR results indicate that zinc substitution provokes a disorder of β-TCP structure. At variance with the behavior of zinc, strontium completely avoids Ca(5) site even at high concentration, whereas it exhibits a clear preference for Ca(4) site. The infrared absorption bands of β-TCP show a general shift towards lower wavenumbers on increasing strontium content. Particularly significant is the shift of the infrared 1 symmetric stretching band at 943 cm− due to P(1), that is the phosphate more involved in Ca(4) coordination, which further supports the occupancy preference of strontium.

Keywords: β-tricalcium phosphate; zinc; strontium; ionic substitutions; Rietveld reﬁnement; solid-state NMR; ATR-FTIR

1. Introduction

β-tricalcium phosphate, β-Ca3(PO4)2 (β-TCP) is the stable polymorph of tricalcium phosphate at temperatures lower than 1125 ◦C. At higher temperatures, it undergoes a thermal transition into the α form [1]. β-TCP is synthesized by solid state reaction at high temperature, alternatively it can be prepared by thermal transition of other calcium phosphates. Although this phase is not present in physiologically calcified biological tissues, it has been found in pathological calcifications and it can be obtained as a product of thermal conversion of the poorly crystalline hydroxyapatite (HA), which constitutes the inorganic phase of bone [2]. Actually, β-TCP both alone and in combination with synthetic HA, is widely applied as a biomaterial for the treatment of defects of the skeletal system [3]. The properties, as well as the relative stability, of calcium phosphates can be modified by functionalization with foreign ions [4]. The number of studies on ionic substitution into HA structure

J. Funct. Biomater. 2019, 10, 20; doi:10.3390/jfb10020020 www.mdpi.com/journal/jfb J. Funct. Biomater. 2019, 10, x FOR PEER REVIEW 2 of 15

J. Funct. Biomater. 2019, 10, 20 2 of 15 structure is particularly relevant, due to its high flexibility, which supports a great variety of cationic, as well as anionic, substitutions [4–8]. is particularlyThe rhombohedral relevant, due structure to its high of flexibility,β-TCP, space which group supports R3c, is a without great variety any doubt of cationic, less flexible as well than as anionic,HA lattice; substitutions however, [4 –it8 ].can host several ionic replacements, especially the replacement of calcium ionsThe with rhombohedral bivalent cations structure [9–14]. of βCalcium-TCP, space ions group in β-TCP R3c, occupy is without five any different doubt lesscation flexible sites: thanCa(1), HACa(2) lattice; and however, Ca(3) are it on can general host several positions ionic with replacements, eight to nine especially coordinated the replacement oxygens, whereas of calcium both ions Ca(4) withand bivalent Ca(5) are cations on special [9–14 ].positions, Calcium with ions inanβ effective-TCP occupy multiplicity five diff erentof 1/3 cationof the sites:other Ca(1),cation Ca(2)sites. andCa(5) Ca(3)exhibits are on an general approximately positions octahedr with eightal coordination, to nine coordinated Ca(4) oxygens,site is only whereas half occupied both Ca(4) and and exhibits Ca(5) a arequite on special distorted positions, nine coordination with an effective (Figure multiplicity 1) [15]. ofPhosphorus 1/3 of the other exhi cationbits three sites. crystallographically Ca(5) exhibits an approximatelydifferent positions, octahedral with coordination,the multiplicity Ca(4) of P(1) site isequal only to half 1/3 occupied of those andof P(2) exhibits and P(3), a quite since distorted it is on a ninespecial coordination position. (Figure1)[ 15]. Phosphorus exhibits three crystallographically different positions, with the multiplicity of P(1) equal to 1/3 of those of P(2) and P(3), since it is on a special position.

Figure 1. Coordination geometry of the five different cation sites of β-TCP [15]. Bonds at 3.0 0.1 Å Figure 1. Coordination geometry of the five different cation sites of β-TCP [15]. Bonds at 3.0± ± 0.1 Å length are dotted. length are dotted. In this paper we investigated the structural modifications induced on β-TCP structure by functionalizationIn this paper with we two investigated bivalent cations, the structural namely strontium modifications and zinc. induced Both theseon β-TCP ions arestructure of great by biologicalfunctionalization interest:strontium with two hasbivalent been cations, reported namely to display strontium a beneficial and zinc. action Both in the these treatment ions are of of bone great diseasesbiological characterized interest: strontium by abnormally has been high reported mass to resorption, display a beneficial such as osteoporosis action in the [ 16treatment,17]; zinc of hasbone beendiseases shown characterized to inhibit osteoclast by abnormally proliferation high mass [18] and resorption, it has been such reported as osteoporosis to display [16,17]; antibacterial zinc has propertiesbeen shown [19]. to Moreover, inhibit osteoclast their beneficial proliferation effect on [18] bone and cells it has has been also beenreported observed to display when coupledantibacterial to β-TCPproperties [20–24 [19].]. In Moreover, particular, their strontium-substituted beneficial effect onβ bone-TCP cells (SrTCP) has also added been with observed mesenchymal when coupled stem cellsto β has-TCP been [20–24]. shown In toparticular, improve strontium-substituted bone formation and fusionβ-TCP across(SrTCP) the added transverse with mesenchymal processes, both stem incells osteoporotic has been andshown non-osteoporotic to improve bone animal formation models and [ 25fusion]. The across possible the transverse substitution processes, of these both two in cationsosteoporotic into β-TCP and non-osteoporotic structure has been animal previously models investigated. [25]. The possible Heat treatmentsubstitution of poorlyof these crystalline two cations HAinto synthesized β-TCP structure in the presence has been of zincpreviously produced investigat Zn substituteded. Heatβ -TCPtreatment (ZnTCP) of uppoorly to a Zncrystalline content ofHA aboutsynthesized 9 at% [26 in]. the The presence results of of Rietveld zinc produced refinement Zn indicatedsubstituted a linearβ-TCP decrease (ZnTCP) of up the to lattice a Zn constantcontent of andabout a preference 9 at% [26]. of The Zinc results for the of Ca(5) Rietveld site. refinement Evidences ofindicated preference a linear for the decrease octahedral of the Ca(5) lattice site constant were obtainedand a preference also by combination of Zinc for ofthe near-edge Ca(5) site. X-ray Evidences absorption of preference fine structure for the (NEXAFS) octahedral measurements Ca(5) site were andobtainedab-initio alsocalculations by combination with X-rayof near-edge refinement X-ray of absorption ZnTCP synthesized fine structure by (NEXAFS) solid state measurements reaction up toand a zinc ab-initio content calculations of about 3.3with at% X-ray [27 ].refinement These data of areZnTCP in agreement synthesized with by thesolid results state reaction of a study up ofto a first-principleszinc content calculationsof about 3.3 which at% [27]. indicated These that data Ca(4) are and in Ca(5)agreement are the with favored the results substitution of a sitesstudy of of divalentfirst-principles cations: thecalculations lower coordination which indicated number that and Ca bond(4) and lengths Ca(5) ofare Ca(5), the favored are preferred substitution by smaller sites of sizeddivalent cations cations: such as the Zn lower and Mg, coordination whereas larger-sized number and ones, bond such lengths as Sr of and Ca(5), Ba, favor are preferred the Ca(4) by site smaller [28]. Onsized the othercations hand, such on asthe Zn basisand Mg, of the whereas refinement larger-siz resultsed ofones, zinc such doped as biphasicSr and Ba, calcium favor the phosphate Ca(4) site

J. Funct. Biomater. 2019, 10, 20 3 of 15

(BCP) Gomes et al. (2011) [29] concluded that Zn is incorporated into β-TCP structure by substituting calcium ions both in Ca(5) and in Ca(4) sites. Strontium substitution to calcium into β-TCP has been obtained by thermal treatment of calcium deficient apatite up to 20 at% and by solid state reaction up to 80 at% [9,30]. However, detailed structural analysis on strontium substituted β-TCP (SrTCP) was performed only up to a Sr substitution of about 33 at% and indicated a preferential occupation of the Ca(4) site [31,32]. Herein, we investigated the substitution of zinc and strontium to calcium into β-TCP structure by comparing solid-state NMR, ATR-FTIR and Rietveld refinement data of samples at different degree of cationic substitution.

2. Materials and Methods

2.1. Synthesis β-TCP was obtained by solid-state reaction of CaCO and CaHPO 2H O (DCPD) in the molar 3 4· 2 ratio of 1:2 at 1000 ◦C for 12 h. The solid product was ground into a mortar before being submitted to further heat treatment. In particular, two diﬀerent materials were prepared following the procedure reported above: one was prepared using commercial DCPD (Sigma-Aldrich, Milano, Italy) and labeled as β-TCP-C; the other one (labeled just as β-TCP) was synthesized using freshly-prepared DCPD. The synthesis of DCPD was carried out using 600 mL of a phosphate solution containing 0.08 mol of Na HPO 12H O (Carlo Erba, 2 4· 2 Milano, Italy) and 0.08 mol of NaH PO H O (Carlo Erba), pH 4 adjusted with glacial CH COOH. 2 4· 2 3 The solution was heated at 37 C and 200 mL of solution containing 0.16 mol of Ca(CH COO) H O ◦ 3 2· 2 (Carlo Erba) was added drop-wise over a period of about 60 min, under continuous stirring. Afterwards the precipitate was stored in contact with the mother solution for 10 min, ﬁltered, repeatedly washed with bidistilled water and dried at 37 ◦C. For the preparation of Sr-substituted β-TCP, α-Sr3(PO4)2 was also prepared by solid-state reaction at 1200 ◦C of SrCO3 and (NH4)2HPO4 in the molar ratio of 3:2 for 12 h. Sr-substituted β-TCP samples with a Sr content of 10, 20, 40, 60 and 80 atom% (with respect to total cations Ca + Sr) were prepared by heat treatment of appropriate stoichiometric mixture of β-TCP and α-Sr3(PO4)2 at 1000 ◦C for 12 h. Samples were labeled 10SrTCP, 20SrTCP, 40SrTCP, 60SrTCP and 80SrTCP, respectively. For the preparation of Zn-substituted β-TCP, α-Zn3(PO4)2 was also prepared by solid-state reaction at 800 ◦C of (ZnCO3)2[Zn(OH)2]3 and (NH4)H2PO4 in the molar ratio of 3:10 for 12 h. Zn-substituted β-TCP samples with a Zn content of 5, 10 and 15 atom% (with respect to total cations Ca + Zn) were prepared by heat treatment of appropriate stoichiometric mixture of β-TCP and α-Zn3(PO4)2 at 1000 ◦C for 8 h. Samples were labeled 5ZnTCP, 10ZnTCP and 15ZnTCP, respectively.

2.2. Characterization X-ray diffraction analysis was carried out by means of a PANalytical X’Pert PRO powder diffractometer equipped with a fast X’Celerator detector. Cu Kα radiation was used (40 mA, 40 kV). The 2θ range was from 10◦ to 60◦ with a step size of 0.05◦ and time/step of 150 s. Data used for structural investigation were collected in the 2θ range 10◦–100◦ with a 0.026◦ step and counting time of 500 s/step. MAUD program (Material Analysis Using Diffraction) [33] was used for structural refinements 2+ 2+ 2+ 2 employing scattering factors for Ca , Sr , Zn and O − ions, as well as for phosphorus atom. The background was fitted as a polynomial function. The process started with the refinement of scale factor, background coefficients and 2θ shift. In the further cycles cell axes, peak widths, their dependence on 2θ, asymmetry and Gaussianity fraction of the peaks were kept as free variables. Finally the structural parameters (occupancy factors and coordinates) were refined using rigid body constrain for phosphate tetrahedra. The reported lattice parameters resulted from refinements. J. Funct. Biomater. 2019, 10, 20 4 of 15

Solid-state NMR spectra were acquired with a Jeol ECZR 600 instrument, operating at 242.95 MHz forJ. theFunct.31 Biomater.P nucleus. 2019, 10 Powder, x FOR PEER samples REVIEW were packed into cylindrical zirconia rotors with a 4 3.2 of 15mm o.d. and a 60 µL volume and spun at 20 kHz. A certain amount of sample was collected and used o.d. and a 60 µL volume and spun at 20 kHz. A certain amount of sample was collected and used 31 withoutwithout further further preparations preparations fromfrom allall samplessamples toto fillfill the rotor. 31PP MAS MAS spectra spectra were were acquired acquired at at roomroom temperature temperature for for all all samples. samples. TheThe spectraspectra were collected collected using using a awith with the the direct direct excitation excitation pulse pulse 31 sequencesequence with with a a 90 90°◦ 31PP pulsepulse ofof 1.31.3 µµs,s, anan optimized recycle recycle delay delay of of 1800 1800 s sand and a anumber number of ofscans scans 31 betweenbetween 4–8, 4–8, depending depending on on the the sample. sample. TheThe 31PP chemical chemical shift shift scale scale was was calibrated calibrated through through a 85% a 85% H3HPO3PO4 solution4 solution as as external external standard.standard. ForFor infrared infrared absorption absorption analysis analysis in attenuated in attenuat totaled total reflection reflection (ATR) (ATR) mode, mode, samples samples were analyzed were usinganalyzed a Bruker using ALPHA a Bruker FT-IR ALPHA spectrometer FT-IR spectromet equippeder withequipped a diamond with a unit,diamond to collect unit, 64to scanscollect in 64 the 1 −1 1 −1 rangescans 400–4000 in the range cm− 400–4000at a resolution cm at of a 4resolution cm− . Data of 4 analysis cm . Data was analysis operated was with operated OPUS software.with OPUS software.Morphological investigation was performed using a Hitachi S-2400 scanning electron microscope operatingMorphological at 15 kV. Sputter-coating investigation withwas gold performed was performed using a before Hitachi examination. S-2400 scanning electron microscope operating at 15 kV. Sputter-coating with gold was performed before examination. 3. Results 3. Results The solid-state synthesis of β-TCP yielded two different products depending on the type of CaHPOThe2H solid-stateO (DCPD) synthesis utilized asof reagent.β-TCP yielded The X-ray two di differentffraction produc patternts of depending the material on obtained the typeusing of 4· 2 freshlyCaHPO synthesized4·2H2O (DCPD) DCPD utilized (β-TCP) as isreagent. compared The withX-ray that diffraction of the product pattern ofof thethe synthesismaterial obtained performed usingusing commercial freshly synthesized DCPD (β -TCP-C)DCPD (β in-TCP) Figure is2 compared. Although with most that of theof the peaks product are the of same the synthesis in the two patterns,performedβ-TCP-C using displayscommercial several DCPD peaks (β-TCP-C) of low in intensity Figure 2. which Although are notmost characteristic of the peaks are of pure the sameβ-TCP, in the two patterns, β-TCP-C displays several peaks of low intensity which are not characteristic of as proved by comparison with calculated pattern (Figure S1). On the other hand, the 31P MAS NMR pure β-TCP, as proved by comparison with calculated pattern (Figure S1). On the other hand, the 31P spectrum of β-TCP-C exhibits a much higher number of narrow and slightly overlapping resonances MAS NMR spectrum of β-TCP-C exhibits a much higher number of narrow and slightly overlapping than that of β-TCP (Figure2). resonances than that of β-TCP (Figure 2).

FigureFigure 2. 2.(A ()A XRD) XRD patterns patterns and and ( B(B))31 31P (242.95 MHz) MAS MAS solid-state solid-state NMR NMR spectra spectra of ofβ-TCPβ-TCP (black), (black), β-TCP-Cβ-TCP-C (blue) (blue) and andβ-TCP β-TCP data fromdata Mellierfrom Mellier et al. [ 34et] (red).al. [34] The (red). reﬂection The markersreflection for markers rhombohedral for β-TCPrhombohedral are reported β-TCP as vertical are reported bars (green). as vertical Red bars arrows (green). highlight Red arrows peaks highlight not belonging peaks tonot rhombohedral belonging β-TCP.to rhombohedral Adapted with β-TCP. permission Adapted from with reference permission [34 ].from Copyright reference (2011) [34]. AmericanCopyright Chemical(2011) American Society. Chemical Society. Similar peculiar XRD pattern and NMR spectrum (also shown in Figure2 for comparison) were previouslySimilar recorded peculiar by XRD Mellier pattern et and al. NMR (2011) spectrum [34], who (also ascribed shown in the Figure NMR 2 for resonances comparison) to atwere least 16previously inequivalent recorded P sites by and Mellier suggested et al. a(2011) lower [34], symmetry who ascribed superstructure the NMR on resonances the basis ofto theat least presence 16 ofinequivalent two weak low P sites angle and reﬂections suggested not a lower allowed symmetry by R3c superstructure symmetry. The on large the basis number of the and presence narrow of line widthtwo ofweak the resonanceslow angle reflections agree with not a highallowed ordering by R3c of symmetry. the vacancies The within large number the β-TCP and structure narrow line rather width of the resonances agree with a high ordering of the vacancies within the β-TCP structure rather than a random distribution, since the latter would result in significantly broad 31P signals.

J. J.Funct. Funct. Biomater. Biomater. 20192019, 10, 10, x, 20FOR PEER REVIEW 5 of5 of 15 15

Mellier et al. (2011) [34] also performed a preliminary assignment of the 31P isotropic chemical shifts 31 bythan DFT a randomcalculations distribution, using the sinceGIPAW the method latter would [35,36]. result in significantly broad P signals. Mellier et 31 al. (2011)However, [34] the also XRD performed pattern arecorded preliminary from assignmentβ-TCP-C, as of well the as Pthat isotropic reported chemical by Mellier shifts et al. by [34] DFT althoughcalculations not usinghighlighted the GIPAW by methodthe Authors, [35,36 ].shows the presence of a number of further weak reflectionsHowever, in addition the XRD to pattern the two recorded weak low from anglβ-TCP-C,e reflections, as well which as that leads reported to suggest by Mellier that et al.these [34 ] materialsalthough contain not highlighted further, not by theidentified, Authors, crystalline shows the phases. presence Thus, of a numberXRD and of NMR further data weak of reflectionsβ-TCP-C seemin addition to lead toto thedifferent two weak conclusions. low angle In reflections, order to bypass which leadsthis apparent to suggest co thatntradiction these materials and highlight contain justfurther, the influence not identified, of ionic crystalline substitution phases. on β-TCP Thus, structure, XRD and we NMR carried data out of βall-TCP-C the syntheses seem to of lead the to differentdifferent samples conclusions. using In DCPD order tosynthesized bypass this in apparent our lab contradictionwhich yields andβ-TCP highlight characterized just the influenceby a XRD of patternionic substitution where all the on peaksβ-TCP correspond structure, we to carriedthose of out the all calculated the syntheses one on of thethe dibasisfferent of rhomboedral samples using structure.DCPD synthesized in our lab which yields β-TCP characterized by a XRD pattern where all the peaks correspondThe powder to those X-ray of pattern the calculated of products one on synthesize the basisd of in rhomboedral the presence structure. of different amounts of Sr2+ 2+ and ZnThe2+ are powder reported X-ray in Figure pattern 3A,B, of products respectively. synthesized In agreement in the presencewith previous of diff dataerent [9], amounts the patterns of Sr 2+ ofand Sr-TCP Zn indicateare reported the presence in Figure of3 A,B,a unique respectively. crystalline In agreementphase up to with a Sr2+ previous content dataof 80 [ at%.9], the patterns of Sr-TCP indicate the presence of a unique crystalline phase up to a Sr2+ content of 80 at%.

FigureFigure 3. 3.XRDXRD patterns patterns of of strontium strontium substituted substituted (A (A) and) and zinc zinc substituted; substituted; (B ()B samples) samples compared compared with with unsubstitutedunsubstituted ββ-TCP.-TCP. The peakspeaks notnot corresponding corresponding to to those those characteristic characteristic of β -TCPof β-TCP are indicated are indicated with *. with *. Moreover, the evident shift of the reflections towards lower angles on increasing the Sr concentrationMoreover, (Figurethe evident3A inset) shift are of in the agreement reflections with towards the substitution lower angles of the on bigger increasing cation (radiusthe Sr concentration0.118 nm) tothe (Figure smaller 3A Ca inset) (0.100 are nm). in agreement At variance, wi theth amountthe substitution of Zn that of can the be bigger hosted cation by the structure(radius 0.118of β -TCPnm) isto noticeablythe smaller lower: Ca (0.100 inset innm). Figure At 3varianceB shows, that the onlyamount thepatterns of Zn that of the can samples be hosted synthesized by the structurein the presence of β-TCP of is 5 noticeably and 10 Zn at%lower: are inset in agreement in Figure with3B shows the presence that only of the a uniquepatterns crystalline of the samples phase, synthesizedwhereas at ain higher the presence Zn concentration of 5 and further10 Zn at% peaks are indicate in agreement the presence with ofthe a presence secondary of phase. a unique crystallineScanning phase, electron whereas microscopy at a higher image Zn of concen pure βtration-TCP shows further the peaks characteristic indicate micrometric the presence particles of a secondarymorphology phase. with rounded edges, which does not seem to be significantly affected by the presence of a lowScanning amount electron of Sr (Figure microscopy4). The productsimage of synthesizedpure β-TCP at shows increasing the Srcharacteristic content, as wellmicrometric as in the particlespresence morphology of Zn, appear with to be rounded constituted edges, of morewhich dense does blocks, not seem which to be however significantly retain theaffected characteristic by the presencemorphology of a low of a amount solid-state of Sr reaction (Figure product. 4). The products synthesized at increasing Sr content, as well as in the presence of Zn, appear to be constituted of more dense blocks, which however retain the characteristic morphology of a solid-state reaction product.

J. Funct. Biomater. 2019, 10, 20 6 of 15 J. Funct. Biomater. 2019, 10, x FOR PEER REVIEW 6 of 15

Figure 4.4.SEM SEM images images of βof-TCP, β-TCP, 20SrTCP, 20SrTCP, 80SrTCP 80SrTCP and 10ZnTCP and 10ZnTCP show the characteristicshow the characteristic morphology ofmorphology a solid-state of reactiona solid-state product. reaction product.

3.1. Structural Analysis The structuralstructural refinementsrefinements were were performed performed by theby Rietveldthe Rietveld method method [37] starting [37] starting from the from atomic the positionatomic position set of β -TCPset of inβ-TCP space in group space R3c group (n. 161)R3c [(n.15 ].161) Since [15]. the Since asymmetric the asymmetric unit contains unit contains five calcium, five threecalcium, phosphorus three phosphorus and ten oxygenand ten atoms,oxygen in atoms, order in to order limit theto limit number the number of free variablesof free variables we used we a rigidused bodya rigid model body withmodel constraints with constraints to maintain to mainta the samein the geometry same geometry of phosphate of phosphate tetrahedra tetrahedra as in pure as βin-TCP. pure Phosphorusβ-TCP. Phosphorus atom positions atom positions were kept were free kept to move free to in move accordance in accordance with the with symmetry the symmetry rules of therules crystal of the system. crystal system. Thermal parametersparameters werewere fixed fixed at at the the values values of βof-TCP β-TCP [15 ][15] because because of their of their high high correlation correlation with occupancywith occupancy factors factors (OF), (OF), but anbut overall an overall thermal thermal parameter parameter factor factor was was allowed allowed to vary.to vary. The The sum sum of theof the OF OF of of calcium calcium and and of of the the substituent substituent ion ion in in each each metal metal (M) (M) site site was was imposedimposed toto unityunity (with(with the exception of site M(4) whichwhich waswas 0.430.43 inin accordanceaccordance withwith ββ-TCP-TCP structurestructure [[15]).15]). No constraintconstraint waswas imposed on the overall metal atoms content and no attempt was made to didifferentiatefferentiate the calcium positions from the substituentsubstituent metal ones. A graphica graphicall plot of the refinementrefinement of sample 20SrTCP is reported in Figure Figure 55 as an example, whereas whereas the the mo mostst relevant relevant structural structural pa parametersrameters are are reported reported in inTables Tables 1,2.1 andFull2 .set Full of setcalculated of calculated vs. measured vs. measured plots are plots reported are reported in Supplementary in Supplementary Material Material (Figure (FigureS2), whereas S2), whereas the structural the structural data of SrTCP data of and SrTCP ZnTCP and samples ZnTCP samplesare deposited are deposited in ICSD files in ICSD (1905508 files (190550810SrTCP, 10SrTCP,1905509 190550910ZnTCP, 10ZnTCP, 1905510 1905510 60SrTCP, 60SrTCP, 1905511 1905511 40SrTCP, 40SrTCP, 1905512 1905512 80SrTCP, 80SrTCP, 1905513 20SrTCP, 19055141905514 5ZnTCP).5ZnTCP).

J. Funct. Biomater. 2019, 10, 20 7 of 15 J. Funct. Biomater. 2019, 10, x FOR PEER REVIEW 7 of 15

Figure 5. Comparison of the observed (dots) and calcul calculatedated (red) patterns of 20SrTCP. At the bottom reﬂectionreflection markersmarkers andand curvecurve didifference.ﬀerence.

Table 1. Lattice parameters of β-TCP in samples synthesized at diﬀerent strontium or zinc content. Table 1. Lattice parameters of β-TCP in samples synthesized at different strontium or zinc content. β-TCP 10SrTCP 20SrTCP 40SrTCP 60SrTCP 80SrTCP 1 5ZnTCP 10ZnTCP β-TCP 10SrTCP 20SrTCP 40SrTCP 60SrTCP 80SrTCP 1 5ZnTCP 10ZnTCP a (Å) 10.4387 (1) 10.4578 (2) 10.4897 (2) 10.5773 (2) 10.6525 (4) 10.6996 (2) 10.3820 (3) 10.3364 (2) a (Å) 10.4387 (1) 10.4578 (2) 10.4897 (2) 10.5773 (2) 10.6525 (4) 10.6996 (2) 10.3820 (3) 10.3364 (2) c (Å) 37.3979 (6) 37.421 (1) 37.599 (1) 38.189(1) 38.708 (2) 39.226 (1) 37.247 (1) 37.1262 (9) c (Å) 37.3979 (6) 37.421 (1) 37.599 (1) 38.189(1) 38.708 (2) 39.226 (1) 37.247 (1) 37.1262 (9) V (Å3) 3529 3543 3583 3700 3804 3889 3477 3435 V (Å3) 3529 3543 3583 3700 3804 3889 3477 3435 1 Parameters based on β’-TCP setting are: a = 10.6996(2) Å, c = 19.613(1) Å, V = 1945 Å3. 1 Parameters based on β’-TCP setting are: a = 10.6996(2) Å, c = 19.613(1) Å, V = 1945 Å3.

Table 2.2. ReﬁnedRefined structural parametersparameters forfor SrTCPSrTCP samples.samples.

M 1M 1 10SrTCP10SrTCP 20SrTCP 20SrTCP 40SrTCP 40SrTCP 60SrTCP 60SrTCP 80SrTCP 80SrTCP 80SrTCP5 5 OF Sr atoms/cellatoms/cell M(1)M(1) 18 180.01 0.01 0 0 0.100.10 1.8 0.340.34 6.16.1 0.62 11.211.2 0.730.73 13.113.1 M(1)M(1) 0.850.8530.6 30.6 M(2)M(2) 18 180.05 0.05 0.90.9 0.090.09 1.6 0.330.33 6.06.0 0.68 12.212.2 0.830.83 15.015.0 M(31)M(31)0.23 0.238.28 8.28 M(3)M(3) 18 180.17 0.17 3.063.06 0.400.40 7.2 0.680.68 12.212.2 0.67 12.112.1 0.800.80 14.414.4 M(32)M(32)0.23 0.238.28 8.28 2 M(4)M(4) 2 6 6 0.32 0.32 1.921.92 0.400.40 2.4 0.410.41 2.42.4 0.41 2.42.4 0.430.43 2.62.6 M(4)M(4) 0.220.222.64 2.64 M(5) 6 0 0 0 0 0.01 0 0.01 0 0.01 0 M(5) 0 0 M(5) 6 0 0 0 0 0.01 0 0.01 0 0.01 0 M(5) 0 0 Sr at./cell 3 5.88 13.0 26.7 37.9 45.1 49.8 3 Sr at./cellSr at% 4 5.889.4 13.020.6 26.7 42.2 37.9 60.0 71.445.1 79.049.8 Sr at%Rwp 4 (%) 9.4 6.3 20.6 6.6 42.2 12.7 60.0 13.8 17.971.4 9.2 79.0 Rwp1 (%)multiplicity of crystal6.3 site; 2 the overall 6.6 content of site12.7 4 is 0.43; 3 total13.8 metal atoms inside17.9 unit cell are 63.2 [15];9.24 from1 multiplicity refinement; of5 crystaldata based site; on 2 theβ’-TCP overall setting content [32]. Sinceof siteβ ’-TCP4 is 0.43; cell 3 is total half ofmetalβ-TCP atoms cell, theinside data unit of atom cell/ cellare have been doubled in order to allow direct comparison with other samples. In β’-TCP M(1) and M(31) + M(32) are 63.2 [15]; 4 from refinement; 5 data based on β’-TCP setting [32]. Since β’-TCP cell is half of β-TCP cell, equivalent in β-TCP to M(1) + M(2) and M(3) respectively. the data of atom/cell have been doubled in order to allow direct comparison with other samples. In β’-TCP M(1) and M(31) + M(32) are equivalent in β-TCP to M(1) + M(2) and M(3) respectively. The variation of the unit cell parameters as a function of Sr content is reported in Figure6A. The trendThe variation is strictly of linear the unit and cell fits parameters well with previouslyas a function reported of Sr content data [is9, 38reported]. The overallin Figure strontium 6A. The contentstrend is resultingstrictly linear from and the refinementsfits well with with previously no imposed reported constraints data are[9,38]. very The close overall to the analyticalstrontium ones.contents However, resulting agreement from the indexesrefinements (Rwp, with Table no2 )im showposed a worseningconstraints asare strontium very close content to the increases,analytical andones. the However, Rwp value agreement obtained indexes for 80SrTCP (Rwp, isTable particularly 2) show high.a worsening In order as to strontium improve thecontent fit, we increases, refined β 80SrTCPand the Rwp using value a somewhat obtained diforfferent 80SrTCP starting is particularly model, namely high. theIn order structure to improve of ’-TCP. the Thisfit, we structure refined R- m β (80SrTCP3 ) was using reported a somewhat to occur different as a thermal starting phase model, transition namely ofthe -TCPstructure promoted of β’-TCP. by strontium This structure [32]. β It(R- is3 topologicallym) was reported similar to occur to -TCP, as a thermal the main phase difference transition being of the β-TCP presence promoted of a symmetry by strontium center [32]. that It c β impliesis topologically a unit cell similar with half to β-TCP,-axis andthe volume.main difference In ’-TCP, being M(1) the and presence M(2) sites, of a as symmetry well as P(2) center and that P(3) tetrahedra,implies a unit are cell equivalent with half to c each-axis other and volume. (M(1) and In P(2)β’-TCP, respectively); M(1) and M(3)M(2) issites, slightly as well displaced as P(2) from and the symmetry element and split into M(31) and M(32), each of which with occupancy of 1 , and M(4) P(3) tetrahedra, are equivalent to each other (M(1) and P(2) respectively); M(3) is slightly4 displaced from the symmetry element and split into M(31) and M(32), each of which with occupancy of ¼, and

J. Funct. Biomater. 2019, 10, 20 8 of 15 J. Funct. Biomater. 2019, 10, x FOR PEER REVIEW 8 of 15

1 hasM(4) occupancy has occupancy of 4 [32 of]. Reﬁnement¼ [32]. Refinement of 80SrTCP of using80SrTCP as a using model as the a structuremodel the of βstructure’-TCP indeed of β’-TCP gives aindeed better gives ﬁtting, a asbetter shown fitting, by the asdata shown reported by the in data Table re2ported and in in Figure Table S3. 2 and In contrast, in Figure samples S3. In atcontrast, lower strontiumsamples at content lower strontium cannot be content described cannot using beβ ’-TCPdescribed structure using sinceβ’-TCP they structure present since some they XRD present peaks whichsome XRD are characteristic peaks which ofareβ -TCPcharacteristic and cannot of β be-TCP indexed and cannot using thebe indexedβ’-TCP unitusing cell the setting, β’-TCP as unit shown cell insetting, Figure as S4. shown in Figure S4.

FigureFigure 6.6.( A(A) Plot) Plot of aof- anda- andc-axis c-axis as a functionas a function of strontium of strontium content; content; (B) Values (B of) normalizedValues of normalized occupancy factorsoccupancy of each factors of the of five each metal of the position five metal as a function position of as Sr a atom function content. of Sr The atom values content. for sample The values 80SrTCP for valuessample are 80SrTCP obtained values from are the obtained refinement from performed the refinement using the performedβ’-TCP structure using the as β the’-TCP starting structure model. as the starting model. The normalized OF of SrTCP samples are reported in Figure6B as a function of strontium content. The choiceThe normalized to plot normalized OF of SrTCP factors samples is obliged are by reported the big diinff erenceFigure in6B the as multiplicity a function ofof M(1),strontium M(2) andcontent. M(3) The (multiplicity choice to =plot18) normalized in comparison factors to that is ofob M(4)liged andby the M(5) big sites difference (multiplicity in the= multiplicity6) that would of maskM(1), theM(2) relative and M(3) cation (multiplicity distribution = if18) plotted in comparison as absolute to values.that of M(4) Results and of M(5) powder sites fitting (multiplicity refinements = 6) indicatethat would that mask at low the degree relative of substitutioncation distribution strontium if plotted displays as the absolute highest values. OF in M(4)Results site, of followed powder byfitting M(3) refinements site; as the indicate overall substitutionthat at low degree degree of increases, substitution the fillingstrontium of sites displays M(1) andthe highest M(2) become OF in important,M(4) site, followed too. It is worthby M(3) mentioning site; as the that overall in the subs wholetitution range degree of investigated increases, compositionthe filling of (strontium sites M(1) upand to M(2) 80 at%), become site M(5) important, is not occupied too. It is by worth Sr atoms. mentioning The ‘filling’ that ofin calcium the whole sites range by strontium of investigated starting fromcomposition M(4) and (strontium M(3) is also up due to 80 to at%), a geometrical site M(5) reason. is not occupiedβ-TCP structure by Sr atoms. can be The described ‘filling’ as of a calcium regular assemblysites by strontium of two kinds starting of columns from M(4) in which and CaM(3) atoms is also and due phosphate to a geometrical tetrahedra reason. are stacked. β-TCP As structure shown incan Figure be described7, the unit as cella regular view down assemblyc-axis of allowstwo kinds to identify of columns A-type in columnswhich Ca filled atoms with and Ca(4), phosphate Ca(5) andtetrahedra P(1) whereas are stacked. B-type As columns shown contain in Figure Ca(1), 7, Ca(2)the unit Ca(3) cell and view P(2) down and P(3).c-axis A-type allows columns to identify are surroundedA-type columns only filled by B-type with ones.Ca(4), At Ca(5) first and strontium P(1) whereas enters intoB-type Ca(4), columns causing contain a significant Ca(1), Ca(2) expansion Ca(3) ofandc-axis. P(2) and Further P(3). strontium A-type columns substitution are surrounded for calcium only is better by B-type accommodated ones. At first in strontium the other enters column into to compensateCa(4), causing the a increment significant along expansion the c-axis. of c-axis. Further strontium substitution for calcium is better accommodatedMoreover, the in the preference other column of strontium to compensate for sites M(4) the andincrement M(3) can along be ascribed the c-axis. to the most favorable distancesMoreover, found inthe these preference sites, which of strontium can accommodate for sites biggerM(4) ionsand thanM(3) calcium.can be ascribed Mean Ca-O to distancesthe most infavorableβ-TCP are:distances Ca(1)-O found 2.505 in Å, these Ca(2)-O sites, 2.487 which Å, can Ca(3)-O accommodate 2.601 Å, Ca(4)-O bigger 2.936ions than Å, Ca(5)-O calcium. 2.263 Mean Å. Ca-OCa-O environmentsdistances in β exhibit-TCP are: the Ca(1)-O longest mean2.505 distanceÅ, Ca(2)-O in site2.487 M(4). Å, Ca(3)-O 2.601 Å, Ca(4)-O 2.936 Å, Ca(5)-OMean 2.263 M(3)-O Å. Ca-O distance environments is second exhibit just to M(4)-Othe longest and mean any further distance increments in site M(4). in Sr content can be easilyMean accommodated M(3)-O distance here, alsois second due to just its higherto M(4)-O storage and abilityany further than increments M(4): the unit in Sr cell content contains can less be thaneasily 3 accommodated M(4) type atoms here, vs. 18 also atoms due ofto theits higher M(3) type. storage On theability other than hand, M(4): location the unit of cell strontium contains in siteless M(5)than 3 is M(4) strongly type unfavorable atoms vs. 18 since atoms the of meanthe M(3) M(5)-O type. distance On the other is the hand, shortest. location Furthermore, of strontium the rangein site ofM(5) M-O is strongly bond values unfavorable in each sitesinc shoulde the mean also M(5)-O play a role:distance M(4) is showsthe shortest. the shorter Furthermore, contact atthe 2.53 range Å, M(5)of M-O at 2.21bond Å values and all in the each other site sitesshould at valuesalso play greater a role: than M(4) 2.30 shows Å [15 the]. These shorter data contact agree at with 2.53 the Å, preferenceM(5) at 2.21 of Å strontium and all the for siteother M(4) sites reported at values previously greater than on β -TCP2.30 Å samples [15]. These containing data agree small with amount the ofpreference strontium of [31 strontium,39], and demonstrate for site M(4) that reported the preference previously is maintained on β-TCP also samples at much containing higher strontium small amount of strontium [31,39], and demonstrate that the preference is maintained also at much higher

J. Funct. Biomater. 2019, 10, 20 9 of 15 J. Funct. Biomater. 2019, 10, x FOR PEER REVIEW 9 of 15 content.strontium In content. fact, the resultsIn fact, obtainedthe results for obtained the sample for 80SrTCPthe sample indicate 80SrTCP the indicate same preference the same both preference when theboth reﬁnement when the isrefinement performed is using performed the structure using the of βstructure-TCP [15 of] and β-TCP that [15] of β and’-TCP that [32 of]. β’-TCP [32].

FigureFigure 7. 7. Left Left, views, views of theof βthe-TCP β-TCP unit cellunit down cell thedownc-axis. theRight c-axis., a viewRight of, thea view atoms of that the characterize atoms that A-characterize and B-type A- ‘columns’. and B-type ‘columns’.

TheThe ZnZn atom, atom, due due to itsto smallerits smaller ionic ionic radius radius (0.075 (0.075 nm) than nm) calcium, than calcium, shows a dishowsfferent a behavior.different Thebehavior. results The of the results structural of the refinements structural refinements (Figure S5, Table(Figure3) showS5, Table that 3) zinc show is distributed that zinc is over distributed all the sitesover in all ZnTCP the sites at a in low ZnTCP substitution at a low degree substitution (5%), but degree M(5) becomes (5%), but the M(5) preferred becomes sitewhen the preferred zinc content site increaseswhen zinc (10%). content The increases octahedral (10%). coordination The octahedral geometry coordination and short distancesgeometry of and M(5) short are particularlydistances of suitableM(5) are for particularly the little zinc suitable ion. Probably for the little due alsozinc toion. the Probably limited substitutiondue also to range,the limited ZnTCP substitution does not presentrange, anyZnTCP discontinuity does not inpresent cell parameters, any discontinuity as previously in cell reported parameters, for MgTCP as previously [40], where reported magnesium for wasMgTCP found [40], to fillwhere both magnesium M(4) and M(5) was sites. found to fill both M(4) and M(5) sites.

Table 3. Refined structural parameters for ZnTCP samples. Table 3. Refined structural parameters for ZnTCP samples. 1 M M 15ZnTCP 5ZnTCP 10ZnTCP 10ZnTCP OF Sr OF Sratoms atoms/cell/cell M(1)M(1) 18 0.0518 0.90.05 0.9 0.030.03 0.5 0.5 M(2)M(2) 18 0.0718 1.30.07 1.3 0.070.07 1.2 1.2 1.3 1.4 M(3)M(3) 18 0.0718 0.07 1.3 0.080.08 1.4 M(4) 2 6 0.05 0.0 0.03 0.0 2 M(5)M(4) 6 0.156 0.90.05 0.0 0.970.03 0.0 5.8 Zn at/cell 3 M(5)- 6 4.40.15 0.9 0.97 8.95.8 Zn at% 4 Zn at/cell- 3 - 7.04.4 8.9 14.1 Rwp (%)Zn at% - 4 - 7.4 7.0 14.1 8.9 1 multiplicity of crystalRwp site; 2 (%)the overall content - of site 4 is 0.43; 3 7.4total metal atoms 8.9 inside unit cell are 63.2 [15]; 4 from refinement. 1 multiplicity of crystal site; 2 the overall content of site 4 is 0.43; 3 total metal atoms inside unit cell are 63.2 [15]; 4 from refinement. 3.2. Solid-State 31P MAS NMR The 31P isotropic chemical shift is very sensitive to local structural modifications, thus providing a reliable probe for this class of samples. For the system under investigation the Zn and Sr substitutions were followed by 31P MAS NMR.

J. Funct. Biomater. 2019, 10, x FOR PEER REVIEW 10 of 15

3.2. Solid-State 31P MAS NMR

J. Funct.The Biomater. 31P isotropic2019, 10, 20chemical shift is very sensitive to local structural modifications, thus providing10 of 15 a reliable probe for this class of samples. For the system under investigation the Zn and Sr substitutions were followed by 31P MAS NMR. The 31P MAS NMR spectrum of pure β-TCP (Figure8 bottom) is characterized by three main The 31P MAS NMR spectrum of pure β-TCP (Figure 8 bottom) is characterized by three main broad resonances at 4.9 (P2), 1.5 (P2) and 0.3 (P3) ppm and three shoulders at 2.6, 0.8 and 0.5 ppm broad resonances at 4.9 (P2), 1.5 (P2) and 0.3 (P3) ppm and three shoulders at 2.6, 0.8 and −−0.5 ppm assigned to P1, P3 and P1, respectively. The spectrum strongly resembles that reported by Obadia et al. assigned to P1, P3 and P1, respectively. The spectrum strongly resembles that reported by Obadia et (2006) [41] for the sample named TCPPrecip obtained by precipitation of a Ca(OH)2 solution by the al. (2006) [41] for the sample named TCPPrecip obtained by precipitation of a Ca(OH)2 solution by addition of aqueous H3PO4, followed by calcination. the addition of aqueous H3PO4, followed by calcination. Even for a very low substitution of Ca with Zn or Sr in the β-TCP structure, the 31P resonances Even for a very low substitution of Ca with Zn or Sr in the β-TCP structure, the 31P resonances change in terms of chemical shift and line width (Figures8 and9). In particular, the broadening change in terms of chemical shift and line width (Figures 8 and 9). In particular, the broadening observed in the substituted samples suggests a random substitution of the foreign ions for calcium. observed in the substituted samples suggests a random substitution of the foreign ions for calcium. As discussed in Section 3.1, XRD data indicate that Zn substitution takes place preferentially As discussed in Section 3.1, XRD data indicate that Zn substitution takes place preferentially on on the M(5) site, which is surrounded by the P2 and P3 phosphorus sites (Figure1). On the other the M(5) site, which is surrounded by the P2 and P3 phosphorus sites (Figure 1). On the other hand, hand, Sr exhibits a clear preference for an M(4) site, which is connected to P1 and P2 (Figure1). Sr exhibits a clear preference for an M(4) site, which is connected to P1 and P2 (Figure 1). In In particular, at low concentration strontium occupies preferentially M(4) and M(3) sites, whereas at particular, at low concentration strontium occupies preferentially M(4) and M(3) sites, whereas at higher concentration it also substitutes calcium in M(1) and M(2), but not in M(5). higher concentration it also substitutes calcium in M(1) and M(2), but not in M(5). The comparison among the 31P MAS NMR spectrum of β-TCP and those of the ZnTCP and SrTCP The comparison among the 31P MAS NMR spectrum of β-TCP and those of the ZnTCP and samples conﬁrms XRD results. SrTCP samples confirms XRD results. In particular, ZnTCP spectra (Figure8) show a low-frequency shift (about 1 ppm) and a drastic In particular, ZnTCP spectra (Figure 8) show a low-frequency shift (about 1 ppm) and a drastic narrowing (FWHM from ~530 to ~195 Hz) of the P2 signal at 4.9 ppm. The P3 signal, at 0.3 ppm, narrowing (FWHM from ~530 to ~195 Hz) of the P2 signal at 4.9 ppm. The P3 signal, at 0.3 ppm, though being less aﬀected, moves toward higher frequencies to 0.6 ppm. though being less affected, moves toward higher frequencies to 0.6 ppm.

Figure 8. 31P (242.95 MHz) MAS spectra of β-TCP (with relevant assignment), 5ZnTCP and 10ZnTCP, Figure 8. 31P (242.95 MHz) MAS spectra of β-TCP (with relevant assignment), 5ZnTCP and 10ZnTCP, acquired at room temperature with a spinning speed of 20 kHz. acquired at room temperature with a spinning speed of 20 kHz.

In the case ofof thethe SrSr substitutionsubstitution (Figure (Figure9 ),9), a morea more complex complex peak peak evolution evolution is observed:is observed: for for low low Sr Srcontents contents (up (up to to 20%) 20%) the the spectrum spectrum gets gets simpler simpler and and is is characterized characterized byby fourfour relativelyrelatively broad peaks centered at 5.2 (P2), 1.9 (P2), 0.7 (P1(P1 andand P3)P3) andand −0.8 (P1 and P3) ppm.ppm. At higher Sr contents, the − spectrum changes significantly signiﬁcantly with with a severe broa broadeningdening in agreement with the XRD analysis (see above) which suggests the Sr substitutionsubstitution on all calcium sites but M(5). These data concerning zinc and strontium strontium substi substitutedtuted TCP, TCP, in the whole are in good agreement with those reported by Grigg et al.al. (2014) for β-TCP-TCP doped doped with with aluminum, aluminum, gallium gallium and and sodium, sodium, which showed that that ions bigger than calcium (N (Na)a) preferentially enter site (4), and those with shorter radius (Al, Ga) substitu substitutete calcium at site (5) [[42].42].

J. Funct. Biomater. 2019, 10, 20 11 of 15 J. Funct. Biomater. 2019, 10, x FOR PEER REVIEW 11 of 15

Figure 9. 31P (242.95 MHz) MAS spectra of β-TCP (with relevant assignment), 10SrTCP, 20SrTCP Figure 9. 31P (242.95 MHz) MAS spectra of β-TCP (with relevant assignment), 10SrTCP, 20SrTCP (with (with relevant assignment), 40SrTCP, 60SrTCP and 80SrTCP acquired at room temperature with a relevant assignment), 40SrTCP, 60SrTCP and 80SrTCP acquired at room temperature with a spinning spinning speed of 20 kHz. speed of 20 kHz.

3.3. ATR-FTIR Spectroscopy The infrared absorption spectrumspectrum ofof ββ-TCP shows a number of bands due to the vibration of the phosphate groups.groups. InfraredInfrared andand Raman Raman studies studies identiﬁed identified the the bands bands due due to to the the symmetric symmetric stretching stretchingν1 1 −1 1 −1 (940–980ν1 (940–980 cm −cm), the), the triple-degenerate triple-degenerate asymmetric asymmetric stretching stretchingν3 (1000–1100 ν3 (1000–1100 cm− cm), and), and the doublethe double and 1 −1 1 −1 triple-degenerateand triple-degenerate bending bendingν2 (400–500 ν2 (400–500 cm− ) and cmν4) (550–600and ν4 cm(550–600− )[43 –cm45].) The [43–45]. infrared The absorption infrared spectrumabsorption of spectrum pure β-TCP of pure is compared β-TCP is with compared those of with samples those at of increasing samples Srat contentincreasing in Figure Sr content 10A. in FigureThe 10A. comparison shows that Sr substitution provokes an increasing degeneracy and a general shift of the bands towards lower wavenumbers. This is particularly evident for the ν1 band which 1 1 shifts from 943 cm− in the spectrum of β-TCP to 936 cm− in the sample at the highest Sr content 1 (Table S1). In the Raman spectrum of β-TCP this band was ascribed to P(1), the band at about 968 cm− was assigned to P(3), whereas the intermediate band, not appreciable in the ATR-FTIR spectrum, was assigned to P(2) [46,47]. P(1) and P(2) are close to Ca(4); in particular P(1) and P(2) tetrahedra share a common face and an edge, respectively, with Ca(4) site (Figure1), whereas P(3) is not in the close 1 vicinity of Ca(4). It follows that the shift of the 943 cm− band at lower wavenumbers on increasing Sr substitution is consistent with its preferential occupancy of Ca(4) site. Incorporation of zinc into β-TCP structure induces a signiﬁcantly increased degeneracy of the infrared absorption bands (Figure 10B), in agreement with previously reported data [9,26], which suggests an increasing structural disorder.

Figure 10. ATR-FTIR spectra of β-TCP at different Sr (A) and Zn (B) content. The arrow indicates the ν1 band at 943 cm−1 in the spectrum of β-TCP.

J. Funct. Biomater. 2019, 10, x FOR PEER REVIEW 11 of 15

Figure 9. 31P (242.95 MHz) MAS spectra of β-TCP (with relevant assignment), 10SrTCP, 20SrTCP (with relevant assignment), 40SrTCP, 60SrTCP and 80SrTCP acquired at room temperature with a spinning speed of 20 kHz.

3.3. ATR-FTIR Spectroscopy The infrared absorption spectrum of β-TCP shows a number of bands due to the vibration of the phosphate groups. Infrared and Raman studies identified the bands due to the symmetric stretching ν1 (940–980 cm−1), the triple-degenerate asymmetric stretching ν3 (1000–1100 cm−1), and the double and triple-degenerate bending ν2 (400–500 cm−1) and ν4 (550–600 cm−1) [43–45]. The infrared absorption spectrum of pure β-TCP is compared with those of samples at increasing Sr content in FigureJ. Funct. 10A. Biomater. 2019, 10, 20 12 of 15

Figure 10. ATR-FTIR spectra of β-TCP at different Sr (A) and Zn (B) content. The arrow indicates the Figure 10. ATR-FTIR spectra of β-TCP at diﬀerent Sr (A) and Zn (B) content. The arrow indicates the −1 ν1 band at 943 cm in1 the spectrum of β-TCP. ν1 band at 943 cm− in the spectrum of β-TCP.

4. Conclusions B-TCP samples were prepared at increasing contents of zinc or strontium. XRD data show that Zn-TCP can be prepared as a single phase up to a zinc content of about 10 at%, whereas a much wider range of substitution, up to about 80 at%, has been obtained for strontium. The results of Rietveld refinements indicate that Zn occupies preferentially the octahedral Ca(5) site, in agreement with the results reported by Kannan et al. (2009) [26] and by Kawabata et al. (2011) [27], whereas it does not show any preference for the Ca(4) site as previously suggested [29]. Zinc substitution provokes a reduction of the cell parameters, a shift of the solid-state 31P NMR resonances of the two phosphate close to Ca(5), namely P(2) and P(3), and a general disorder of β-TCP structure as shown by the broadening of the ATR-FTIR bands. In contrast, the relatively small Ca(5) site is not appreciated by strontium, which does not occupy it even at the highest concentrations: after filling the Ca(4) site, strontium occupies Ca(1), Ca(2) and Ca(3), but not Ca(5). The clear preference of strontium for the Ca(4) site, not only at a small content as previously reported [31,32,39] but also at a relatively high content, is further confirmed by the significant shift of the infrared symmetric stretching band at 943 1 cm− due to P(1), that is the phosphate more involved into Ca(4) coordination. The preference is also maintained for a strontium content of about 80 at%, even if the high amount of strontium provokes a slight modification the β-TCP structure into the more symmetric β’-TCP. These results provide further, more detailed, information about the influence played by ionic substitution on β-TCP structure, which is of particular interest in the case of strontium and zinc in view of their important biological roles.

Supplementary Materials: The following are available online at http://www.mdpi.com/2079-4983/10/2/20/s1, Figure S1: XRD patterns of β-TCP-C, β-TCP and calculated pattern; Figure S2: observed and calculated XRD patterns of SrTCP samples; Figure S3: observed and calculated XRD patterns of 80SrTCP, refinement performed using β’-TCP setting; Figure S4: comparison between refinements carried out in β’-TCP and β-TCP settings for sample 60SrTCP; Figure S5: observed and calculated XRD patterns of ZnTCP samples; Table S1: ATR-FTIR bands wavenumbers. Author Contributions: Contributions: conceptualization, E.B. and A.B.; data curation, E.B., M.G., M.R.C.; formal analysis, E.B., M.G., C.N.; investigation, E.B., M.G., M.R.C. and K.R.; methodology, M.G., C.N.; validation, E.B., M.G. and K.R.; supervision R.G., A.B.; writing—original draft, E.B., M.G., R.G. and A.B. Funding: This research received no external funding Acknowledgments: The authors are grateful to the support of the University of Bologna. Conflicts of Interest: The authors declare no conflict of interest. J. Funct. Biomater. 2019, 10, 20 13 of 15

References

1. Elliott, J.C. Structure and Chemistry of the Apatites and Other Calcium Orthophosphates; Elsevier: Amsterdam, The Netherlands, 1994. 2. Bigi, A.; Cojazzi, G.; Panzavolta, S.; Ripamonti, A.; Roveri, N.; Romanello, M.; Noris Suarez, K.; Moro, L. Chemical and structural characterization of the mineral phase from cortical and trabecular bone. J. Inorg. Biochem. 1997, 68, 45–51. [CrossRef]

3. Dorozhkin, S.V. Multiphasic calcium orthophosphate (CaPO4) bioceramics and their biomedical applications. Ceram. Int. 2016, 42, 6529–6554. [CrossRef] 4. Boanini, E.; Gazzano, M.; Bigi, A. Ionic substitutions in calcium phosphates synthesized at low temperature. Acta Biomater. 2010, 6, 1882–1894. [CrossRef] 5. Shepherd, J.H.; Shepherd, D.V.; Best, S.M. Substituted hydroxyapatites for bone repair. J. Mater. Sci. Mater. Med. 2012, 23, 2335–2347. [CrossRef] 6. Bigi, A.; Boanini, E.; Gazzano, M. Ion substitution in biological and synthetic apatites. In Biomineralization and Biomaterials, Fundamentals and Applications, 1st ed.; Aparicio, C., Ginebra, M.P., Eds.; Woodhead Publishing (Elsevier): Sawston, UK, 2015; pp. 235–266. ISBN 9781782423386. 7. Šupová, M. Substituted hydroxyapatites for biomedical applications: A review. Ceram. Int. 2015, 41, 9203–9231. [CrossRef] 8. Bigi, A.; Boanini, E. Functionalized biomimetic calcium phosphates for bone tissue repair. J. Appl. Biomater. Funct. Mater. 2017, 15, e313–e325. [CrossRef] 9. Bigi, A.; Foresti, E.; Gandolfi, M.; Gazzano, M.; Roveri, N. Isomorphous substitutions in β-tricalcium phosphate: The different effects of zinc and strontium. J. Inorg. Biochem. 1997, 66, 259–265. [CrossRef] 10. Mayer, I.; Cuisinier, F.J.G.; Gdalya, S.; Popov, I. TEM study of the morphology of Mn2+-doped calcium hydroxyapatite and β-tricalcium phosphate. J. Inorg. Biochem. 2008, 102, 311–317. [CrossRef] 11. Kannan, S.; Lemos, I.A.F.; Rocha, J.H.G.; Ferreira, J.M.F. Synthesis and characterization of magnesium substituted biphasic mixtures of controlled hydroxyapatite/β-tricalcium phosphate ratios. J. Solid State Chem. 2005, 178, 3190–3196. [CrossRef]

12. Li, X.; Ito, A.; Sogo, Y.; Wang, X.; LeGeros, R.Z. Solubility of Mg-containing β-tricalcium phosphate at 25◦C. Acta Biomater. 2009, 5, 508–517. [CrossRef] 13. Frasnelli, M.; Sglavo, V.M. Eﬀect of Mg2+ doping on beta–alpha phase transition in tricalcium phosphate (TCP) bioceramics. Acta Biomater. 2016, 33, 283–289. [CrossRef][PubMed] 14. Zhang, M.; Wu, C.; Li, H.; Yuen, J.; Chang, J.; Xiao, Y. Preparation, characterization and in vitro angiogenic capacity of cobalt substituted β-tricalcium phosphate ceramics. J. Mater. Chem. 2012, 22, 21686–21694. [CrossRef] 15. Yashima, M.; Sakai, A.; Kamiyama, T.; Hoshikawa, A. Crystal structure analysis of β-tricalcium phosphate

Ca3(PO4)2 by neutron powder diffraction. J. Solid State Chem. 2003, 175, 272–277. [CrossRef] 16. Bonnelye, E.; Chabadel, A.; Saltel, F.; Jurdic, P. Dual effect of strontium ranelate: Stimulation of osteoblast differentiation and inhibition of osteoclast formation and resorption in vitro. Bone 2008, 42, 129–138. [CrossRef][PubMed] 17. Salamanna, F.; Giavaresi, G.; Parrilli, A.; Torricelli, P.; Boanini, E.; Bigi, A.; Fini, M. Antiresorptive properties of strontium substituted and alendronate functionalized hydroxyapatite nanocrystals in an ovariectomized rat spinal arthrodesis model. Mater. Sci. Eng. C-Mater. Biol. Appl. 2019, 95, 355–362. [CrossRef] 18. Shepherd, D.V.; Kauppinen, K.; Brooks, R.A.; Best, S.M. An in vitro study into the effect of zinc substituted hydroxyapatite on osteoclast number and activity. J. Biomed. Mater. Res. Part A 2014, 102, 4136–4141. [CrossRef] 19. Sutha, S.; Karunakaran, G.; Rajendran, V. Enhancement of antimicrobial and long-term biostability of the zinc-incorporated hydroxyapatite coated 316 L stainless steel implant for biomedical application. Ceram. Int. 2013, 39, 5205–5212. [CrossRef] 20. Yamada, Y.; Ito, A.; Kojima, H.; Sakane, M.; Miyakawa, S.; Uemura, T.; LeGeros, R.Z. Inhibitory effect of Zn2+ in zinc-containing β-tricalcium phosphate on resorbing activity of mature osteoclasts. J. Biomed. Mater. Res. Part A 2008, 84, 344–352. [CrossRef] 21. Roy, M.; Bose, S. Osteoclastogenesis and osteoclastic resorption of tricalcium phosphate: Effect of strontium and magnesium doping. J. Biomed. Mater. Res. Part A 2012, 100, 2450–2461. [CrossRef] J. Funct. Biomater. 2019, 10, 20 14 of 15

22. Roy, M.; Fielding, G.A.; Bandyopadhyay, A.; Bose, S. Effects of zinc and strontium substitution in tricalcium phosphate on osteoclast differentiation and resorption. Biomater. Sci. 2013, 1, 74–82. [CrossRef] 23. Chou, J.; Hao, J.; Hatoyama, H.; Ben-Nissan, B.; Milthorpe, B.; Otsuka, M. The therapeutic effect on bone mineral formation from biomimetic zinc containing tricalcium phosphate (ZnTCP) in zinc-deficient osteoporotic mice. PLoS ONE 2013, 8, e71821. [CrossRef] 24. Boanini, E.; Torricelli, P.; Sima, F.; Axente, E.; Fini, M.; Mihailescu, I.N.; Bigi, A. Gradient coatings of strontium hydroxyapatite/zinc β-tricalcium phosphate as a tool to modulate osteoblast/osteoclast response. J. Inorg. Biochem. 2018, 183, 1–8. [CrossRef] 25. Salamanna, F.; Giavaresi, G.; Contartese, D.; Bigi, A.; Boanini, E.; Parrilli, A.; Lolli, R.; Gasbarrini, A.; Barbanti Brodano, G.; Fini, M. Effect of strontium substituted β-TCP associated to mesenchymal stem cells from bone marrow and adipose tissue on spinal fusion in healthy and ovariectomized rat. J. Cell. Physiol. 2019, in press. [CrossRef] 26. Kannan, S.; Goetz-Neunhoeffer, F.; Neubauer, J.; Ferreira, J.M.F. Synthesis and structure refinement of zinc-doped β-tricalcium phosphate powders. J. Am. Ceram. Soc. 2009, 92, 1592–1595. [CrossRef] 27. Kawabata, K.; Yamamoto, T.; Kitada, A. Substitution mechanism of Zn ions in β-tricalcium phosphate. Phys. B 2011, 406, 890–894. [CrossRef] 28. Matsunaga, K.; Kubota, T.; Toyoura, K.; Nakamura, A. First-principles calculations of divalent substitution of Ca(2+) in tricalcium phosphates. Acta Biomater. 2015, 23, 329–337. [CrossRef] 29. Gomes, S.; Nedelec, J.M.; Jallot, E.; Sheptyakov, D.; Renaudin, G. Unexpected mechanism of Zn2+ insertion in calcium phosphate bioceramics. Chem. Mater. 2011, 23, 3072–3085. [CrossRef] 30. Kannan, S.; Pina, S.; Ferreira, J.M.F. Formation of strontium-stabilized β-tricalcium phosphate from calcium-deficient apatite. J. Am. Ceram. Soc. 2006, 89, 3277–3280. [CrossRef] 31. Nandha Kumar, P.; Boovarasan, M.; Singh, R.K.; Kannan, S. Synthesis, structural analysis and fabrication of

coatings of the Cu2+ and Sr2+ co-substitutions in β-Ca3(PO4)2. RSC Adv. 2013, 3, 22469–22479. [CrossRef] 32. Belik, A.A.; Izumi, F.; Stefanovich, S.Y.; Malakho, A.P.; Lazoryak, B.I.; Leonidov, I.A.; Leonidova, O.N.; Davydov, S.A. Polar and centrosymmetric phases in solid solutions Ca Srx(PO ) (0 x 16/7). Chem. Mater. 3-x 4 2 ≤ ≤ 2002, 14, 3197–3205. [CrossRef] 33. Lutterotti, L. Total pattern fitting for the combined size-strain-stress-texture determination in thin film diffraction. Nucl. Instrum. Meth. B 2010, 268, 334–340. [CrossRef] 34. Mellier, C.; Fayon, F.; Schnitzler, V.; Deniard, P.; Allix, M.; Quillard, S.; Massiot, D.; Bouler, J.M.; Bujoli, B.; Janvier, P. Characterization and properties of novel gallium-doped calcium phosphate ceramics. Inorg. Chem. 2011, 50, 8252–8260. [CrossRef] 35. Segall, M.D.; Lindan, P.J.D.; Probert, M.J.; Pickard, C.J.; Hasnip, P.J.; Clark, S.J.; Payne, M.C. First-principles simulation: Ideas, illustrations and the CASTEP code. J. Phys.-Condensed Matter 2002, 14, 2717–2744. [CrossRef] 36. Clark, S.J.; Segall, M.D.; Pickard, C.J.; Hasnip, P.J.; Probert, M.J.; Refson, K.; Payne, M.C. First principles methods using CASTEP. Z. Kristallogr. 2005, 220, 567–570. [CrossRef] 37. Young, R.A. The Rietveld Method; Oxford University Press: Oxford, UK, 1993.

38. Nord, A.G. Incorporation of divalent metals in whitlockite-related β-Ca3(PO4)2. Neues Jahrbuch fuer Mineralogie Monatshefte 1983, 11, 489–497. 39. Kannan, S.; Goetz-Neunhoeﬀer, F.; Neubauer, J.; Pina, S.; Torres, P.M.C.; Ferreira, J.M.F. Synthesis and structural characterization of strontium and magnesium co-substituted β-tricalcium phosphate. Acta Biomater. 2010, 6, 571–576. [CrossRef] 40. Bigi, A.; Falini, G.; Foresti, E.; Ripamonti, A.; Gazzano, M.; Roveri, N. Rietveld structure reﬁnement of synthetic magnesium substituted β-tricalcium phosphate. Z. Kristallogr. 1996, 211, 13–16. [CrossRef] 41. Obadia, L.; Deniard, P.; Alonso, B.; Rouillon, T.; Jobic, S.; Guicheux, J.; Julien, M.; Massiot, D.; Bujoli, B.; Bouler, M. Characterization and properties of novel gallium-doped calcium phosphate ceramics. Chem. Mater. 2006, 18, 1425–1433. [CrossRef] 42. Grigg, A.T.; Mee, M.; Mallinson, P.M.; Fong, S.K.; Gan, Z.; Dupree, R.; Holland, D. Cation substitution in β-tricalcium phosphate investigated using multi-nuclear, solid-state NMR. J. Solid. State Chem. 2014, 212, 227–236. [CrossRef] 43. Jillavenkatesa, A.; Condrate, R.A. The infrared and Raman spectra of β- and α-tricalcium phosphate

(Ca3(PO4)2). Spectrosc. Lett. 1998, 31, 1619–1634. [CrossRef] J. Funct. Biomater. 2019, 10, 20 15 of 15

44. De Aza, P.N.; Santos, C.; Pazo, A.; De Aza, S.; Cusco, R.; Artus, L. Vibrational properties of calcium phosphate compounds. 1. Raman spectrum of β-tricalcium phosphate. Chem. Mater. 1997, 9, 912–915. [CrossRef] 45. De Aza, P.N.; Guitian, F.; Santos, C.; De Aza, S.; Cusco, R.; Artus, L. Vibrational properties of calcium phosphate compounds. 2. Comparison between hydroxyapatite and β-tricalcium phosphate. Chem. Mater. 1997, 9, 916–922. [CrossRef] 46. Jakeman, R.J.B.; Cheetham, A.K.; Clayden, N.J.; Dobson, C.M. A magic angle spinning NMR study of the

phase diagram Ca3-xZnx(PO4)2. J. Solid State Chem. 1989, 78, 23–34. [CrossRef] 47. Quillard, S.; Paris, M.; Deniard, P.; Gildenhaar, R.; Berger, G.; Obadia, L.; Bouler, J.M. Structural and spectroscopic characterization of a series of potassium and/or sodium-substituted β-tricalcium phosphate. Acta Biomater. 2011, 7, 1844–1852. [CrossRef]

© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).