S

Sa

b

c

a

A

R

R

2

A

K

Z

S

X

R

S

O

1

TYws

G

0d

d e n t a l m a t e r i a l s 2 6 ( 2 0 1 0 ) 295–305

avai lab le at www.sc iencedi rec t .com

journa l homepage: www. int l .e lsev ierhea l th .com/ journa ls /dema

urface characterization of zirconia dental implants

. Zinelisa, A. Thomasb, K. Syresb, N. Silikasc, G. Eliadesa,∗

Department of Biomaterials, University of Athens, GreeceSchool of Physics and Astronomy, Photon Science Institute, University of Manchester, UKBiomaterials Unit, University of Manchester, UK

r t i c l e i n f o

rticle history:

eceived 27 May 2009

eceived in revised form

7 September 2009

ccepted 17 November 2009

eywords:

irconia dental implants

urface analysis

PS

aman microanalysis

EM/EDX

ptical profilometry

a b s t r a c t

Objectives. The aim of the study was to characterize the chemical composition, microstruc-

ture and roughness of two commercially available zirconia dental implants (WhiteSky and

Zit-Z).

Methods. The chemical composition of the cervical collar and threaded root parts of the

implants (n = 2) were studied by XPS and HV-EDX. LV-SEM was used for morphological

assessment, Raman microanalysis for microstructural characterization and optical pro-

filometry for surface roughness measurements. XRD, HV-EDX and Raman microanalysis

of bulk regions (longitudinal sections) were used as reference.

Results. XPS showed the presence of C, O, Zr and Y (collar) plus Al (root) at implant surfaces.

More C (10–26 at%) and a lower Al/Zr ratio were found in WhiteSky (1.05 vs 1.26 in Zit-Z). Zr,

Y and Al were detected in single, fully oxidized states. The same elements, plus Hf, were

identified by HV-EDX at bulk and surface regions, with a Al/Zr ratio higher in WhiteSky

(0.17 vs 0.09 in Zit-Z). Na, K and Cl contaminants were traced at implant root parts by both

methods. XRD analysis of cross-sectioned specimens revealed the presence of monoclinic

and tetragonal zirconia along with cubic yttria phases. Raman microanalysis showed that

the monoclinic zirconia volume fraction was higher at root surfaces than the collar. No

monoclinic phase was found at bulk regions. Significantly higher Sa and Sq values were

recorded in WhiteSky than Zit-Z, whereas Zit-Z showed higher Rt value.

Significance. The differences found between the implants in the extent of carbon contami-

nation, residual alumina content, tetragonal to monoclinic ZrO2 phase transformation and

3D-roughness parameters may contribute to a substantial differentiation in the cellular and

tissue response.

emy

aesthetic tooth-like colour have made 3Y-TZP a viable alterna-

© 2009 Acad

. Introduction

etragonal stabilized ZrO , with the addition of 3 mol%

2

2O3 (3Y-TZP) has long been considered as a strong, tough,ear-resistant and osseoconductive ceramic suitable for

tress-bearing implantable applications [1,2]. The less inflam-

∗ Corresponding author at: Department of Biomaterials, University ofreece. Tel.: +30 210 7461101; fax: +30 210 7461306.

E-mail address: geliad@dent.uoa.gr (G. Eliades).109-5641/$ – see front matter © 2009 Academy of Dental Materials. Puoi:10.1016/j.dental.2009.11.079

of Dental Materials. Published by Elsevier Ltd. All rights reserved.

matory response and better stabilization of soft-tissues incontact with zirconia [3,4], the lower plaque retention capac-ity and higher affinity to ostoeblasts [5–7] along with the more

Athens, School of Dentistry, 2 Thivon Str-Goudi, 115 27 Athens,

tive to titanium implants.Already, several zirconia implants have been introduced

with proven efficacy in animal studies [8,9]. However,

blished by Elsevier Ltd. All rights reserved.

l s 2

296 d e n t a l m a t e r i a

long-term human trials to establish their clinical success arestill missing [10].

From the material standpoint, concerns have beenexpressed on the low temperature degradation of the tetrag-onal to monoclinic ZrO2 phase, that has been associated within-service failures of orthopaedic implants [11,12]. Treatmentsto avoid or reverse such transformations have been advocatedduring manufacturing or sterilization [13]. Moreover, it hasbeen shown that the osseointegration capacity of machinedZrO2 surfaces is substantially increased after modificationby Al2O3 sandblasting [14]. Although such techniques havealready been adopted in some products, there is lack of infor-mation on the surface chemistry, structure and morphologyof currently commercially available ZrO2 dental implantsthat would facilitate understanding of the implant–boneinteractions and the bone augmentation mechanismsinvolved.

The aim of the present study was to investigate the surfacechemistry, morphology and structure of two commerciallyavailable zirconia implants. The testing hypothesis was thatsignificant differences exist in these properties between thetwo implants.

2. Materials and methods

The products tested were WhiteSky (Ø: 3.5 mm, l: 12 mm,lot 59235, Bredent Medical, Senden, Germany) and Zit-Z (Ø:3.5 mm, l: 13 mm, lot 49744, Ziterion, Uffenheim, Germany).Five specimens from each product were subjected to the fol-lowing testing procedures:

2.1. X-ray photoelectron spectroscopy (XPS)

The samples as received (n = 2) were placed in the ultra-highvacuum chamber of an X-ray photoelectron spectrome-ter (Scienta 300 ESCA system, NCESS, Cheshire, UK). Tworegions, one at the cervical collar and the other at thethreaded root portion, were located on each implant andanalyzed under the following conditions: Al Ka (1486.7 eV)monochromated rotating anode, 14 kV accelerating voltage,0.2 A current emission, 2.8 kW maximum power, 10−9 mbarpressure, ∼90◦ electron take-off angle, 0.30 eV energy reso-lution and 6 mm × 0.5 mm sampling area. An electron floodgun operated at 4 eV was used for charge compensation.Energy calibration was performed based on Ag3d5/2 peak stan-dard. Survey scans (150 eV pass energy) were taken fromeach region to identify the elements present on the surface.Then, high resolution narrow scans (150 eV pass energy) wererecorded over the predominant peaks and the elemental bind-ing states were determined. All spectra were aligned on thebinding energy scale to the C 1s peak arising from adven-titious hydrocarbons (–CH2–, 285 eV BE). The core level datawere analyzed using the CASA XPS software package. A Shirleybackground was subtracted from the data and 80:20 Gaus-sian:Lorenzian peaks fitted to give binding energy positions.

The fitted peak areas in conjunction with relative inten-sity factors allowed the elemental ratios at the surface tobe quantified. The depth of analysis was estimated as to3 nm.

6 ( 2 0 1 0 ) 295–305

2.2. Low-vacuum scanning electron microscopy andhigh vacuum X-ray energy dispersive microanalysis(LV-SEM/HV-EDX)

The same specimens analyzed by XPS were subjected toLV-SEM imaging. Secondary electron images were acquiredemploying a large field detector (LFD) attached to a SEMunit (Quanta 200, FEI, Hillsboro, OR, USA) operated at 30 kVaccelerating voltage, 90 �A beam current, 1 Torr pressure(without electron conductive coating to avoid masking offmicrostructural features) at 40×, 300×, 600× and 2400×magnifications.

The elemental composition of cervical collar, threadedroot and bulk regions (the latter prepared for XRD analy-sis as described below) of each implant were determinedby HV-EDX analysis (this technique provides better accuracyin quantitative determinations than LV-EDX). All the spec-imens were coated with a thin layer of conductive carbonin a sputter-coating unit (SCD 004 Sputter-Coater with OCD30 attachment, Bal-Tec, Vaduz, Liechtenstein). EDX analy-sis was performed using a liquid N2-cooled Si(Li) detectorwith super ultra-thin Be window (Sapphire SUTW+ CDU,EDAX Int, Mahwah, NJ, USA) attached to the SEM unit underthe following conditions: 30 kV energy range, 10−6 Torr pres-sure, 110 �A beam current, 128 eV resolution, 250 s acquisitiontime, 210 �m × 210 �m sampling window and 28–34% detectordead time. The depth of analysis was estimated as to 1 �m.The quantitative analysis was performed in non-standardmode using ZAF and coating corrections employing Gene-sis v. 5.2 software (EDAX, Int). Elemental mapping of regionsof interest was based on compositional backscattered elec-tron images, obtained with a solid-state detector (SSD) atthe same conditions as above, but at 15 kV acceleratingvoltage.

2.3. Raman microanalysis

The cervical collar, threaded root parts bulk regions (pre-pared for XRD analysis as described below) of the implants(n = 2) were analyzed by Raman microscopy to identify andmap the distribution of the tetragonal (145 and 262 cm−1

Raman shift) and monoclinic (180 and 190 cm−1 Ramanshift) ZrO2 phases at the surface region [15]. A Ramanmicroscope was used (LabRAM Aramis, Horiba Jobin-Yvon,Villeneuve d’Ascq, France) operated under the followingconditions: Ar laser (532 nm), 10 mW power at sample,50× LWD objective, 1800 grit/mm grating, 1000 �m confo-cal hole (defocused mode), 100 �m slit, 10 s acquisition.Three regions were randomly located at each implant sur-face location (collar/root) with the optical system of themicroscope (10× optical objective) and analyzed. For map-ping of the monoclinic ZrO2 phase, 20 �m × 35 �m areaswere selected and scanned at 5 �m steps and 5 s acquisitiontime.

From the net peak height intensities of the monoclinic andtetragonal ZrO phases, the percentage volume of the mono-

2

clinic phase (Vm%) was calculated according to the equation:Vm% = {I180 + I190/0.97(I145 + I262) + I180 + I190}× 100, where I thenet peak height intensities at the corresponding Raman shifts[15].

2 6

2

Ioeawfdbtf4pd

2

Ttabear(

Fs

d e n t a l m a t e r i a l s

.4. X-ray diffraction (XRD)

n order to examine the presence of the tetragonal and mon-clinic ZrO2 phases in the entire implant, one specimen fromach implant was embedded in epoxy resin and sectioned atlongitudinal direction using a microtome under continuousater cooling. The specimens were ground to a smooth sur-

ace using SiC paper up to 1200 grit size, polished with a 3 �miamond paste and ultrasonically cleaned in a distilled waterath for 3 min. The sections were studied in an X-ray diffrac-ometer (D8 Focus, Bruker AXS, Karlsruhe, Germany) under theollowing conditions: CuKa anode, 40 kV accelerating voltage,0 mA beam current, 20–90◦ 2� range, 0.02◦ step, 2 s exposureer step. The identification of phases was based on the ICSDatabase [16].

.5. Optical profilometry

he 3D-surface roughness parameters of the threaded parts ofhe implants (n = 2) were evaluated by optical profilometry. Therea between two successive implant threads was analyzedy an optical profiler (NT 1100, Veeco, S. Barbara, CA, USA)

quipped with Michelson/Mirau interferometric objectivest 10–100× magnifications. Quantification of the 3D-surfaceoughness parameters Sa (arithmetic mean deviation), Sqroot mean square deviation), and Rt (maximum peak to val-

ig. 1 – XPS spectra of implant surfaces.(a) Survey spectrum of Wpectra of Zr 3d (c) and Y 3d peaks (d).

( 2 0 1 0 ) 295–305 297

ley height) was performed by Veeco-Vision software at 40×magnification (160 �m × 120 �m).

2.6. Statistical analysis

The results of elemental at% obtained from XPS and EDX weresubjected to a two-way ANOVA (implant type and region asindependent variables, and at% per element as dependentvariable), whereas a two-way ANOVA was used for Vm% com-parisons (implant type and region as independent variables).A Tukey’s test was used for pairwise multiple comparisons.Finally, the differences in the roughness parameters betweenthe implants were evaluated by a t-test. In all comparisonsa 95% confidence level was used (˛: 0.05). Statistical analysiswas performed by SigmaStat software (Jandel, St. Raphael, CA,USA).

3. Results

3.1. XPS

Fig. 1(a and b) shows representative survey XPS spectra of theimplants taken from cervical collar and threaded root regions.The elements identified were C, O, Al, Zr and Y. The results ofthe XPS elemental atomic percentage (at%) are listed in Table 1.

hiteSky; (b) Survey spectrum of Zit-Z; High resolution

298 d e n t a l m a t e r i a l s 2 6 ( 2 0 1 0 ) 295–305

Fig. 2 – Low-vacuum secondary electron images of the implant surfaces. WhiteSky: (a) Low magnification view (40×, bar:1 mm); (b) cervical collar-threaded root transitional zone (300×, bar: 100 �m); (c) threaded root (600×, bar: 50 �m); (d)threaded root at high magnification (2400×, bar: 20 �m). Zit-Z: (e) Low magnification view (40×, bar: 1 mm); (f) cervicalcollar-threaded root transitional zone (300×, bar: 100 �m); (g) threaded root (600×, bar: 50 �m); (h) threaded root at highmagnification (2400×, bar: 20 �m).

d e n t a l m a t e r i a l s 2 6 ( 2 0 1 0 ) 295–305 299

Table 1 – XPS elemental composition of the Zirconia implant surfaces (mean and sd in parentheses)*.

Element (%) WhiteSky Zit-Z

Collar Threaded root Collar Threaded root

C (1s) 76.00a,b,c (3.18) 53.80a (1.88) 49.30b (1.62) 43.30c (1.94)O (1s) 17.70d,e,f (1.44) 32.50d,g,h (0.84) 38.90e,g (1.78) 41.20f,h (1.02)Zr (3d) 5.90i (0.12) 6.50j (0.44) 11.22i,j,k (0.56) 6.60k (0.86)

l 0.08) l

0.80)

erenc

TAdtc

FtiA

Y (3d) 0.40 (0.06) 0.40 (Al (2s) – 6.80 (

∗ Same letters indicate mean values with statistically significant diff

he implant surfaces have similar qualitative composition.

l was identified only at threaded root regions. WhiteSkyemonstrated more C contamination and less O content athe cervical collar. The highest Zr content was found at Zit-Zollar. The XPS Y/Zr atomic ratios, as derived from the data

ig. 3 – High-vacuum EDX spectra of the implants tested along whreaded root regions. (a) Spectra of cervical collar for WhiteSky;maged onto the corresponding compositional backscattered elecl–O–C for WhiteSky (e) and of Al–Na–K–C for Zit-Z (f) imaged on

0.58 (0.1) 0.70 (0.08)– 8.20 (0.64)

es per element between implant groups (p < 0.05).

listed in Table 1, ranged from 0.05 to 0.07 (p > 0.05), with the

exception of Zit-Z collar, where the highest value of 0.15 wasrecorded (p < 0.05). The Al/Zr atomic ratios for WhiteSky andZit-Z were 1.05 and 1.27 respectively (p < 0.05). The Zr 3d andY 3d peaks showed only one doublet arising from 3d5/2 and

ith the elemental distributions of the dispersed phases at(b) Spectra of threaded root for Zit-Z; Al X-ray maps (green)tron images for WhiteSky (c) and Zit-Z (d); X-ray maps ofto the corresponding backscattered electron images.

300 d e n t a l m a t e r i a l s 2 6 ( 2 0 1 0 ) 295–305

Table 2 – EDX elemental composition of the Zirconia implants (mean and sd in parentheses)*.

Element (%) WhiteSky Zit-Z

Bulk Collar Threaded Root Bulk Collar Threaded Root

O (K) 64.55 (0.85) 68.10 (0.83) 72.28a (1.58) 64.90 (0.66) 61.99 (1.28) 54.78a (4.27)Zr (L) 31.50 (0.80) 28.12 (0.87) 21.51b (1.58) 30.97 (0.83) 33.57 (1.22) 37.45b (3.70)Y (L) 3.49 (0.24) 3.29 (0.32) 2.37c (0.14) 3.40 (0.17) 3.88 (0.38) 3.38c (0.09)

d (0.03 d

(0.06

erenc

Hf (L) 0.46 (0.14) 0.48 (0.04) 0.25Al (K) – – 3.58

∗ Same letters indicate mean values with statistically significant diff

3d3/2 spin orbit splitting, indicating that Zr and Y exist in oneoxidation state in the implants (Fig. 1c and d). The bindingenergy of the Zr 3d5/2 peak at 182.4 eV corresponds to ZrO2,whereas that of theY 3d5/2 peak at 156.8 eV to Y2O3. For Al, thebinding energy of the Al 2p peak at 74.68 eV corresponds toAl2O3. Traces of Na were also identified (Na 1s at 1023 eV andNa Auger at 499 eV) in both implants.

3.2. LV-SEM/HV-EDX

Fig. 2(a and e) illustrates low-vacuum secondary electronimages of the threaded roots of the implant surfaces at lowmagnification, which revealed distinct morphological differ-ences. WhiteSky demonstrated a double thread self-cuttingdesign with a longitudinal retentive groove, whereas Zit-Z acylindrical screw design. Fig. 2(b and f) exhibits the transitionalzone between the cervical collar and the threaded root part ofthe implants. The cervical collars showed horizontal machin-ing serrations, more pronounced in Zit-Z. The transition ofthe collar to the rough threaded root part was more clearlydefined in WhiteSky. In Zit-Z the machining tracks were vis-ible even in the roughened threaded root region (Fig. 2c andg). Both implants showed a grainy texture at the root regions.However, the size of the grainy domains was much smaller inZit-Z (Fig. 2d and h).

The HV-EDX analysis showed that the implants were com-posed of O, Hf, Y, and Zr. On the threaded root regions, Al wasidentified as well (Fig. 3a and b). The quantitative results aresummarized in Table 2. A significant interaction was foundamong the independent variables. Comparison between theimplants per same region showed that Zit-Z exhibited thelowest O and the highest Zr content at the threaded root part.

The EDX Y/Zr atomic ratios as derived from the results pre-sented in Table 2 ranged from 0.10 to 0.12 (p > 0.05), whereasthe Al/Zr atomic ratios were 0.17 (WhiteSky) and 0.09 (Zit-Z,p < 0.05). Elemental mapping showed that the Al distributionfollowed the dark regions (low atomic number regions) of thebackscattered images (Fig. 3c and d). At some of these regionsC, Na, K and Cl contaminants were mapped (Fig. 3e and f).

3.3. Raman microanalysis

Raman spectra taken from cervical collar and the treaded rootparts are illustrated in Fig. 4(a–d). The low intensity peaks at

180 and 190 cm−1 Raman shift correspond to the monoclinicZrO2 phase and the high intensity peaks at 145 and 262 cm−1

Raman shift to the tetragonal ZrO2 phase. The threaded partsexhibited more intense monoclinic peaks in both implants.

) 0.47 (0.03) 0.55 (0.06) 1.08 (0.38)) – – 3.30 (0.12)

es per element between implant groups (p < 0.05).

The results of the percentage volume of the monoclinic phase(Vm%) are presented in Table 3. No interaction was foundbetween the independent variables. The threaded root partsdemonstrated the highest Vm% (p = 0.05). No monoclinic phasewas detected at bulk regions. Raman mapping showed that atcervical collars the monoclinic phase distribution was relatedto deep machining grooves, whereas a more uniform distribu-tion was found at the treaded parts (Fig. 4e and f).

3.4. X-ray diffraction (XRD)

Representative XRD graphs are given in Fig. 5(a and b). Themonoclinic ZrO2 phase was clearly identified (2�: 28.2◦) alongwith the tetragonal ZrO2 (2�: 30.2◦) and cubic Y2O3 (2�: 29.1◦)phases. The presence of the latter phase was verified by thesecond peak as well (2�: 48.5◦). No cubic zirconia phase wasidentified.

3.5. Optical profilometry

3D-topometric images of the threaded root regions are givenin Fig. 6(a–c).

WhiteSky demonstrated a rougher surface texture com-pared with Zit-Z. The machining tracks were apparent in theroot part of Zit-Z. Fig. 6(c) illustrates a top-view image ofthe surface between two successive threads after automaticplane subtraction; the latter was used for calculation of thesurface roughness parameters. The surface roughness param-eters calculated at 40× are summarized in Table 4. WhiteSkydemonstrated significantly higher roughness parameters thanZit-Z (p < 0.05).

4. Discussion

The results of the present study confirmed the testing hypoth-esis, that significant differences exist between the implants inthe properties tested.

HV-EDX analysis of bulk sectioned specimens confirmedthat in both implants yttria-stabilized zirconia was used asraw material with a composition complying with the lim-its described for biomedical applications (4.5–5.5% Y2O3, <5%HfO2, ZrO2 + Y2O3 + HfO2 ∼99%) [17,18]. According to the man-ufacturers, sintering of the implants was performed by hot

isostatic pressure (HIP). XRD revealed that a fraction of themonoclinic ZrO2 phase existed in the implants along with thepredominating tetragonal ZrO2 phase. However, since XRD asapplied, lacks of lateral resolution, it is unclear whether the

d e n t a l m a t e r i a l s 2 6 ( 2 0 1 0 ) 295–305 301

Fig. 4 – Raman microanalysis of the implant surfaces. Spectra taken at 3 different locations at cervical collar (a: WhiteSky, c:Zit-Z) and threaded root locations (b: WhiteSky, d: Zit-Z). Mapping of the monoclinic ZrO2 phase distribution at cervicalcollar (e) and threaded root regions (f) for WhiteSky implant.

302 d e n t a l m a t e r i a l s 2 6 ( 2 0 1 0 ) 295–305

Fig. 5 – XRD graphs of WhiteSky (a) and Zit-Z (b) implants.

Table 3 – Percentage of the monoclinic ZrO2 volume(Vm%, mean and sd in parentheses)*.

Vm% WhiteSky Zit-Z

Collar 1.97b (1.82) 3.00c (0.50)Threaded root 4.10a (3.61) 6.27c (0.81)Bulk 0a,b 0c

from raw materials (bulk region). The more uniform patternand the higher extent of the Vm% found on the threaded sur-faces implies that sandblasting induced more stresses that

Table 4 – The 3D-roughness parameters of the threadedroot implant regions (mean and sd in parentheses)*.

Parameters (�m) WhiteSky Zit-Z

Sa 1.31 (0.06) 0.66 (0.05)

t: tetragonal ZrO2, m: monoclinic ZrO2 and c: cubic Y2O3

phases.

monoclinic phase detected corresponded to surface or bulkregions. Moreover, it is not known if the sectioning proce-dure created stresses that contributed to monoclinic phaseformation [19]. The same observations may apply for thecubic Y2O3 phase. Raman microspectroscopy showed lack ofmonoclinic zirconia phase at bulk regions. Although Ramanmicroanalysis does not provide information for the entireimplant section, the randomly analyzed regions may be con-sidered representative of the bulk implant microstructure. Theabsence of monoclinic phase at bulk regions implies that theraw materials used were free of monoclinic surface and thatthe sectioning technique employed did not induce this phase.

The surface chemistry of the implants showed substantialdifferences from bulk. XPS and EDX documented the presenceof Al as Al2O3 on the threaded root part. As per manufac-turers’ information these surfaces were alumina sandblasted,but no further details are given for the procedure. Sandblast-ing with Al2O3, usually at 5 atm with 250 �m size particles,has been described as a method to increase implant macro-roughness, that when combined with acid-etching (to increasemicro-roughness), leads to enhanced bone apposition andincreased torque removal strength [14]. A 3–4 at% Al was foundat the uppermost ∼1 �m surface layer of the root parts of theimplants by EDX, while a 6–8 at% Al by XPS. The highest valuesof the latter should be assigned to the smaller probing depth

of XPS (∼3 nm) that is much more sensitive to surface treat-ments. The Al/Zr atomic ratios of the implants after XPS were1.05 for WhiteSky and 1.26 for Zit-Z. The corresponding ratios,though, after EDX were 0.17 for WhiteSky and 0.09 for Zit-Z.

∗ Same letters indicate mean values with statistically significantdifferences (p < 0.05).

These values may imply that the residual Al2O3 particles inZit-Z are aggregated in a thinner superficial layer. The highestY/Zr XPS atomic ratio found at Zit-Z threaded root part mayimply regional yttria segregation. Although the role of resid-ual Al2O3 on implant surfaces is still a matter of controversy,animal studies have shown that these particles do not affectthe osseointegration pattern [20]. Nevertheless, since the lowcontent of residual Al2O3 particles on implant root surfacesdemonstrates a randomly dispersed pattern, it is quite difficultto conclude on a positive or negative effect from low magni-fication histomorphometric images in the absence of Al2O3

particles from the field of view.No Hf was detected by XPS, implying that Hf compounds

do not occupy surface sites. Zr and Y were found as ZrO2

and Y2O3 in one, fully oxidized state that complies with ISOand ASTM standards for ceramic surgical implants basedon Y-TZP [17,18]. WhiteSky showed more carbon contamina-tion, possibly related to the cleaning procedures after surfacetreatments. It is not known if the increased level of carboncontamination on zirconia surfaces affects the protein andcell adsorption phenomena, as documented for Ti surfaces[21]. The presence of Na, K and Cl contaminants, occasionallydetected by XPS and EDX on implant surfaces, may be residuesof proprietary cleaning treatments performed after sandblast-ing, for which no information is given. The biological role ofthese contaminants is undefined [22].

The Raman analysis of the collar and treaded implantsurfaces showed the presence of the weak monoclinic ZrO2

phase in both implants. The low laser energy used on tar-get (10 mW) and the short spectra acquisition time (10 s max)did not affect the phase composition, due to overheating, asconfirmed by pilot measurements. Since the implants wereanalyzed as received, it can be concluded that the mono-clinic phase was developed either during implant machiningor/and sandblasting, considering the absence of this phase

Sq 1.58 (0.03) 0.85 (0.04)Rt 10.95 (2.10) 25.72 (5.42)

∗ All mean values showed statistically significant differencesbetween implants (p < 0.05).

d e n t a l m a t e r i a l s 2 6 ( 2 0 1 0 ) 295–305 303

Fig. 6 – 3D-optical profilometric images of threaded root regions of WhiteSky (a, 100X) and Zit-Z (b, 100×). For calculation oft e su

diisdatsfidoimc

pcitt

ce

he surface roughness parameters, the top-view images wer

estabilized tetragonal phase more aggressively than machin-ng. The tetragonal to monoclinic ZrO2 phase transformations followed by a 4–6% increase in volume that generatestress fields, which are associated with the development ofefects, like surface and subsurface cracking, grain pull-outnd damage, that may affect load bearing interfaces likehe bone–implant interface [23]. The Vm% found on implanturfaces was quite low (0–7.2%). However, studies on trans-ormation kinetic parameters have shown that the growth ofnitial fraction of monoclinic phase further affects the surfaceegradation mechanism, despite that formation of new mon-clinic nuclei is considered as the predominant effect [24]. It

s quite interesting that no requirement on the pre-existingonoclinic phase content is listed in the relevant ISO specifi-

ation [17].The Raman analysis as performed in the present study,

robed only the surface layer, extending to a depth of a few zir-onia grain diameters. However, the importance of this zones high, since a low percentage of monoclinic phase extendingo a depth of some microns has been found to severely reduce

he strength of Y-TZP [25].

Further studies employing confocal Raman microanalysisould provide important information regarding the in-depthxtent of the tetragonal to monoclinic ZrO2 phase trans-

bjected to automatic plane subtraction (c: WhiteSky).

formation and the distribution of the stresses developed[23].

Dental implants are not similar to orthopaedic implants,where many problems have been encountered from the lowtemperature degradation and transformation of the tetrago-nal to monoclinic ZrO2 phase, that resulted in grain pull-outand catastrophic 3-body abrasive wear of the zirconia articu-lating surfaces (hip and acetabular cap) [12,26]. Nevertheless,osseointegration with a superficial zirconia zone rich in mono-clinic phase could create subsurface cracking under functionalloading due to the great interfacial elastic modulus mismatchwith bone. Additionally, in-service development of the mono-clinic zone, starting from the surface and progressing into thematerial bulk, might further contribute to bulk and interfacialproblems. The exact mechanism of tetragonal to monoclinicZrO2 transformation in aqueous medium is unknown. Threemechanisms have been proposed so far [11]: First is the cor-rosion mechanism, where the transformation proceeds fromthe hydrolytic degradation of the Zr–O–Zr bonds at the cracktips leading to formation of Zr–OH, that relieve the strains

stabilizing the tetragonal phase, while at the same time OH-

groups diffuse into the lattice to occupy oxygen vacancies. Itshould be mentioned that the addition of Y3+ in Zr4+ sites ofY-TZP creates oxygen vacancies due to establishment of Zr–O

l s 2

r

304 d e n t a l m a t e r i a

charge density differences [27]. Second, the destabilizationmechanism, concerns interaction of OH− groups with Y2O3

and formation of hydroxylated yttria compounds (i.e. YO(OH),Y(OH)3) that destabilize the tetragonal phase. Finally, thestress induced transformation mechanism applies for casesof water vapour stress development (i.e. during autoclave ster-ilization) that leads to Y(OH)3 formation and subsequently tomonoclinic phase formation. As most zirconia dental implantsare sterilized by �-rays, the first and second mechanismscould potentially apply. Retrieval analysis of zirconia dentalimplants would greatly assist in understanding the possibleimplication of the monoclinic phase to the failure mecha-nisms.

Of particular interest for osseointegration is the surfacetopography of the implants, especially within the areas con-fined by single threads that are representative of the implants’microtopography. Measurements at this region are quite chal-lenging since the exaggerated form of the threaded structureprovides physical limitations to several stylus-type instru-ments (i.e. stylus profilometers, AFM scanners, etc.) [28]. Theoptical profiler used in the present study, operating in non-contact mode, overwhelmed these limitations. Although noneof the roughness parameters has been directly related tobone response, height descriptors (like Ra, Rq, Rz, Rt, etc.)are more frequently used to characterize implant surfaces.In addition, 3D-roughness parameters are considered moreimportant than the 2D-analogues, since the surface textureanisotropy that modifies cell response is taken into accountin the former, the latter providing information only for in-line measurements [28]. The Sa and Sq values of WhiteSkywere within the range of the values previously reported fora sandblasted and acid-etched cpTi implant [14], whereasthe corresponding values of Zit-Z were almost half of theabove. The differences in the surface roughness between theimplants tested may be attributed to different sandblastingand etching conditions after machining. The lower Sa andSq values of Zit-Z comply with the SEM and 3D-topometricimages, where the machining tracks were detectable even onthe sandblasted surfaces. Rt, being a single value per surfacescan, does not represent a statistical finding of the surface butrather an indication of the maximum vertical surface variance.Increased surface roughness of sandblasted and acid-etchedzirconia implants has been associated with increased removaltorque strength and bone stability, despite that no differencesin osseointegration have been recorded in histomorphometricstudies [14]. Although no comparative studies are available forthese implants, it has long been recognized that roughness is acritical factor for a strong bone–implant augmentation [29,30].

The results of the present study show that signifi-cant differences exist in the surface chemistry, structureand roughness of the implants tested, although both werebased on the same composition and received similar sur-face treatments. The differences found in the extent ofcarbon contamination, residual alumina content, tetragonalto monoclinic ZrO2 phase transformation and 3D-roughnessparameters may contribute to a substantial differentiation in

the cellular and tissue response. Nevertheless, the significanceof these properties in the success of Y-TZP dental implants canbe verified only by controlled long-term clinical studies thatcurrently are not available [10].

6 ( 2 0 1 0 ) 295–305

Acknowledgements

The study was supported by the ELKE 70/4/5768 research fundfrom the National and Kapodistrian University of Athens. Thesupport provided by Dr. E. Lancelot (Raman microanalysis), Dr.H. Stadler (optical profilometry) and Dr. D. S.-L. Law (ESCAfacility at NCESS) is gratefully acknowledged. The implantsamples were generously provided by the manufacturers. Theauthors would like to acknowledge Daresbury NCESS facilityand EPSRC grant ref EP/E025722/1.

e f e r e n c e s

[1] Piconi C, Maccauro G, Muratori F, Branch del Prever E.Alumina and zirconia ceramics in joint replacements. J ApplBiomater Biomech 2003;1:19–32.

[2] Hisbergues M, Vendeville S, Vendeville P. Zirconia:established facts and perspectives for biometerial in dentalimplantology. J Biomed Mater Res Part B: Appl Biomater2009;88B:519–29.

[3] Degidi M, Artese L, Scarano A, Perrotti V, Gehrke P, PiattelliA. Inflammatory infiltrate, microvessel density, nitric oxidesynthase expression, and proliferative activity inperi-implant soft tissues around titanium oxide healingcaps. J Periodontol 2006;77:73–80.

[4] Rimondini L, Cerroni L, Carrasi A, Torricelli P. Bacterialcolonization of zirconia ceramic surfaces: an in vitro and invivo study. Int J Oral Maxillofac Implants 2002;17:793–8.

[5] Bianchi AE, Bosetti M, Dolci G, Sberna MT, Sanfilippo F,Cannas M. In vitro and in vivo follow-up of titaniumtransmucosal implants with zirconia collar. J Appl BiomaterBiomech 2004;2:143–50.

[6] Scarano A, Piatelli M, Caputi S, Favero GA, Piattelli A.Bacterial adhesion on commercially pure titanium andzirconium oxide disks: an in vivo human study. J Periodontol2004;75:292–6.

[7] Depprich R, Ommerborn M, Zipprich H, Naujoks C,Handschel J, Wiesmann H-P, et al. Behavior of osteoblasticcells cultured on titanium and structured zirconia surfaces.Head Face Med 2008;4:29, doi:10.1186/1746-160X-4-29.

[8] Scarano A, Di Carlo F, Quaranta M, Piattelli A. Bone responseto zirconia ceramic implants: an experimental study inrabbits. J Oral Implantol 2003;29:8–12.

[9] Kohal RJ, Weng D, Bachle M, Strub JR. Loaded custom-madezirconia and titanium implants show similarosseointegration: an animal experiment. J Periodontol2004;75:1262–8.

[10] Wenz HJ, Bartsch J, Wolfart S, Kern M. Osseointegration andclinical success of zirconia dental implants: a systematicreview. Int J Prosthodont 2008;21:27–36.

[11] Lawson S. Environmental degradation of zirconia ceramics. JEuro Ceram Soc 1995;15:485–502.

[12] Chevalier J. What future of zirconia as a biomaterial?Biomaterials 2006;27:535–43.

[13] Chevalier J, Gremillard L, Deville S. Low-temperaturedegradation of zirconia and implications for biomedicalimplants. Ann Rev Mater Res 2007;37:1–32.

[14] Gahlert M, Gudehus T, Eichhorn S, Steinhauser S, Kniha H,Erhardt W. Biomechanical and histomorphometric

comparison between zirconia implants with varying surfacetextures and a titanium implant in the maxilla of miniaturepigs. Clin Oral Impl Res 2007;18:662–8.

[15] Clarke DR, Adar F. Measurement of the crystallographicallytransformed zone produced by fracture in ceramics

2 6

Mater Res 1985;19:875–901.

d e n t a l m a t e r i a l s

containing tetragonal zirconia. J Am Ceram Soc1982;65:284–8.

[16] http://icsdweb.fiz-karlsruhe.de/.[17] ISO 13356. Implants for surgery-ceramic materials based on

yttria stabilized tetragonal zirconia (Y-TZP). Geneva:International Standardization Organization; 1997.

[18] ASTM F1873-98: High-purity dense yttria-tetragonalzirconium oxide polycrystal (Y-TZP) for surgical implantapplications. American Society for Testing and Materials1998:13.01;908 [withdrawn].

[19] Deville S, Chevalier J, Gremillard L. Influence of surfacefinish and residual stresses on the aging sensitivity ofbiomedical grade zirconia. Biomaterials 2006;27:2186–92.

[20] Piattelli A, Degidi M, Paolantonio M, Mangano C, Scarano A.Residual aluminum oxide on the surface of titaniumimplants has no effect on osseointegration. Biomaterials2003;24:4081–9.

[21] Aita H, Hori N, Takeuchi M, Suzuki T, Yamada M, Anpo M, etal. The effect of ultraviolet functionalization of titanium

on integration with bone. Biomaterials 2009;30:1015–25.

[22] Massaro C, Rotolo P, De Riccardis F, Milella E, Napoli A,Wieland M, et al. Comparative investigation of the surfaceproperties of commercial titanium dental implants. Part I:chemical composition. J Mater Sci Mater Med 2002;13:538–48.

( 2 0 1 0 ) 295–305 305

[23] Pezzotti G, Yamada K, Sakakura S, Pitto RP. Ramanspectroscopic analysis of advanced ceramic composite forhip prosthesis. J Am Ceram Soc 2008;91:1199–206.

[24] Chevalier J, Cales B, Drouin JM. Low-temperature aging ofY-TZP ceramics. J Am Ceram Soc 1999;82:2150–4.

[25] Thompson I, Rawlings RD. Mechanical behaviour of zirconiaand zirconia-toughened alumina in a simulated bodyenvironment. Biomaterials 1990;11:505–8.

[26] Fernandez-Fairen M, Sala P, Gil FJ. Failures of yttria stabilizedtetragonal zirconia: 52 retrieved ceramic femoral heads oftotal hip prostheses. Biomed Mater Eng 2006;16:415–22.

[27] Munoz MC, Gallego S, Beltrán JI, Cerdá J. Adhesion atmetal–ZrO2 interfaces. Surf Sci Rep 2006;61:303–44.

[28] Macdonald W, Campbell P, Fisher J, Wennerberg A. Variationin surface texture measurements. J Biomed Mater Res Part B:Appl Biomater 2004;70B:262–9.

[29] Thomas KA, Cook S. An evaluation of variables influencingimplant fixation by direct implant apposition. J Biomed

[30] Shalabi MM, GortemakerA, Van’t Hof MA, Jansen JA,Creugers NH. Implant surface roughness and bone healing:a systematic review. J Dent Res 2006;85:496–500.