lable at ScienceDirect

Biomaterials 30 (2009) 3028–3034

Contents lists avai

Biomaterials

journal homepage: www.elsevier .com/locate/biomater ia ls

Nucleation and growth of aragonite crystals at the growth front of nacresin pearl oyster, Pinctada fucata

Kazuko Saruwatari a,*, Tomoyuki Matsui a, Hiroki Mukai a, Hiromichi Nagasawa b, Toshihiro Kogure a

a Department of Earth and Planetary Science, Graduate School of Science, The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-0033, Japanb Department of Applied Biological Chemistry, Graduate School of Agricultural and Life Sciences, the University of Tokyo, Yayoi 1-1-1, Bunkyo-ku, Tokyo 113-8657, Japan

a r t i c l e i n f o

Article history:Received 20 January 2009Accepted 7 March 2009Available online 28 March 2009

Keywords:BiomineralNacreAragoniteGeometrical selectionMineral bridge

* Corresponding author.E-mail address: kazuko@eps.s.u-tokyo.ac.jp (K. Sar

0142-9612/$ – see front matter � 2009 Elsevier Ltd.doi:10.1016/j.biomaterials.2009.03.011

a b s t r a c t

The growth front of nacreous layer, which lies just above the outer prismatic layer, is one of the crucialareas to comprehend the formation of nacreous aragonite. The crystallographic properties of aragonitecrystals at the growth front in pearl oyster, Pinctada fucata, were investigated using scanning electronmicroscopy with electron back-scattered diffraction, and transmission electron microscopy with focusedion beam sample preparation technique. Nano-sized aragonite crystals nucleate with random crystal-lographic orientation inside the dimples on the surface of the organic matrix that covers the outerprismatic columns. The dimples are filled with horn-like aragonite crystals, which enlarge from thebottom to the upper surface to form hemispheric domes. The domes grow concentrically and coalescetogether to become the initial nacreous layer. The c-axes of aragonite at the top surface of the domes arepreferentially oriented perpendicular to the surface. The horn-like aragonite and its crystallographicorientation are probably attained by geometrical selection with the fastest growth rate of aragonite alongthe c-axis, until organic sheets are continuously formed and interrupt the crystal growth of aragonite.The further crystal growth along the shell thickness is attained via mineral bridges through discontinuityor holes in the organic sheets. These results indicate that the crystal growth of aragonite at the growthfront results from not only biotic process but also inorganic ones such as geometrical selection andmineral bridges.

� 2009 Elsevier Ltd. All rights reserved.

1. Introduction

Nacre in mollusk shell is an elegant natural organization withthe pearly optics and marvelous mechanical stiffness, which orig-inate from regular alternation of organic matrix sheets and platyaragonite crystals. It was proposed that the organic matrix sheet isan assembly of b-chitin frameworks, silk-like protein and acidicmacromolecules, e.g., [1–3]. The platy aragonite crystals arecomposed of monocrystalline aragonite tablets and their c-axes areperpendicular to the shell surface according to TEM observationsand XRD measurements, e.g., [4,5]. However, the formationmechanism of the composite nacre is still under debate includingthe problem of the polymorph switching between calcite prism andaragonite nacres in one extrapallial space. One explanation is basedon a hetero-epitaxial nucleation model of aragonite tablets on theorganic chitin sheet proposed by Weiner and Traub [6]. In thismodel, specific organic macromolecules bind to organized b-chitin

uwatari).

All rights reserved.

fibrils, and the binding points set off amorphous calcium carbonate(ACC) and the following developments of aragonite tablets [7].However, this model is not perfect to explain hierarchical particleand laminated structures in an aragonite tablet [8,9], as Marin et al.[10] pointed out them. The particle structures may be related withmany pores in organic sheets observed by Shaffer et al. [11], whoproposed mineral bridge model that the aragonite crystals grewthrough these pores. The direct evidence of mineral bridges wasmainly presented for gastropods, but not for bivalves [9,12,13].

As one clue to understand the nacre formation mechanismincluding the polymorph switching, several researches have paidattention to the interface between prismatic and nacreous layers[14,15]. According to their studies, the initial nacres were poly-crystalline and/or fibrous. Checa et al. [16] investigated the crys-tallographic properties using electron back-scattered diffraction(EBSD) in a scanning electron microscope (SEM), showing that thec-axes of early aragonite tablets were oriented perpendicular to theshell surface and the a- and b-axes were random. With accumu-lating the nacreous layers, b-axes were progressively orientedtowards the growth direction due to the competition between theadjacent tablets within the lamellae. Since their SEM-EBSD results

Fig. 1. (a) Internal surface view of the shell of P. fucata. (b) Internal surface view at the boundary between prismatic and nacreous layers. (c) Enlarged view in the white box in (b).The arrowheads and white arrows indicate the remained top surface of prismatic columns and dimples on the organic matrix, respectively. (d) Oblique perspective SEM view atthe growth front of nacres. The large white arrow indicates the growth direction. The arrowheads and small white arrows indicate the same ones in (c). (e) Enlarged dimplestructures on the organic matrix. (f) A further grown dimple showing concentric structures inside the dimple. (g) A fully grown dome. (h) A surface structure of the initial nacreouslayer, showing the development of irregular shaped tablets of around 2 mm. (i) Step-like structure at the orderly laminated nacreous layers with round tablets. The white arrowindicates the growth direction. (j) Schematic figure of the cross-section around the growth front of the nacreous layer on the prismatic layer.

K. Saruwatari et al. / Biomaterials 30 (2009) 3028–3034 3029

10 µ m

a

c

<100> <010> <001>

b

WD 17

Fig. 2. (a) Sloping SEM photo of the coalesced domes for the EBSD analysis. (b) AnEBSD pattern from the surface of the dome in (a). (c) Stereographic projection resultedfrom the EBSD analyses in (a). The origin in the stereo net corresponds to the normal ofthe shell surface, indicating that their c-axes are almost perpendicular to the shellsurface.

K. Saruwatari et al. / Biomaterials 30 (2009) 3028–30343030

are only information obtained from the surface of the initialnacreous layer, for more comprehensive understanding, we per-formed nanoscopic and cross-sectional investigation of the initialnacre in pearl oyster, Pinctada fucata, using transmission electronmicroscopy (TEM) with focused ion beam (FIB) sample preparationtechnique, as well as SEM-EBSD analyses.

Fig. 3. (a) SEM image of a dimple structure on the organic matrix, inside of which finecalcium carbonate materials are precipitated. The broken-line indicates the part of thecross-sectional specimen for TEM observation, prepared using FIB technique. (b,c) TEMdark field images of fine CaCO3 crystals in the dimple. Note that these crystals areformed intricately with the organic matrix. (d) Nano-diffraction pattern from one ofthe fine crystals, showing that the mineral phase is aragonite. Examination of theseveral other crystals indicated that there is no preferred crystal orientation for theseinitial aragonites.

2. Sample and methods

The shell samples of P. fucata were donated by Mie PrefectureFisheries Research Institute, Mie, Japan, and washed by pure waterbefore drying. SEM and EBSD studies were carried out using HitachiS-4500 SEM with a cold-FE gun and a ThermoNoran PhaseID EBSDsystem. The growth features of the nacreous layer were observed atthe acceleration voltage of 2–5 kV, while EBSD patterns wereacquired at 20 kV. The samples were coated with Pt–Pd for high-resolution SEM observation and with carbon for EBSD analyses. Theacquired EBSD patterns were analyzed using a program developedby Kogure [17]. TEM observations were performed using a JEOL JEM-2010 TEM operated at 200 kV, and the TEM specimens for the initialnacre were prepared using Hitachi FB-2100 FIB with the micro-sampling system. In order to prevent the damage by Ga ion beam onthe sample, the samples were coated thickly with carbon and/or Pt–Pd before the FIB procedure, and FIB operations were performed

Fig. 4. (a) TEM bright field image of a flat dome, showing horn-like shapes of aragonite crystals in the dome. The circles of b and c correspond to the diffraction pattern (b) and (c),respectively. The c-axes are generally perpendicular to the surface.

K. Saruwatari et al. / Biomaterials 30 (2009) 3028–3034 3031

using Ga ion beam of 30 kV. As the final process, the specimens weretrimmed at lower voltage of 10 kV to 100–200 nm in thickness.

3. Results

3.1. SEM-EBSD analyses

Fig. 1(b) and (c) shows the growth front of nacres on the pris-matic layer. The organic matrix between the prismatic columnsgradually spreads over the column remaining the capital andcompletely covers the columns in the final stage (Fig. 1c,d). Smalldimples are formed on the organic matrix between the prismaticcolumns (Fig. 1c,d) and sparse precipitations of small materials areobserved on the inside (Fig. 1e). The dimples are completely filledwith the precipitations to form hemispheric dome (Fig. 1f,g) andthey grow up to 30–40 mm until they coalesce together to becomethe initial nacreous layer (Fig. 1d). The domes become flat duringthe growth probably owing to encapsulation by the organic matterwhich must be removed during sample preparation. Small irregularto round shaped tablets in size about 3–5 mm appear on the surfaceof the initial nacreous layer (Fig. 1h). At the inside of 100–150 mmaway from the growth front of nacreous layers, the tablets aresystematically aligned and the step-like structures start (Fig. 1i).

Surfaces of the domes were investigated by EBSD analyses(Fig. 2). All acquired EBSD patterns (Fig. 2b) were explained by thecrystallographic parameters of aragonite. Fig. 2c is the stereo-graphic projection of the crystallographic axes obtained from thepoints on one of coalesced domes (Fig. 2a) and the pole of thestereograph corresponds to the normal of the shell surface. The c-axes were approximately normal to the surface, although theydeviated up to w25�. On the other hand, the a- and b-axes wererandomly oriented. These results indicate that the top surfaces ofthe domes were almost uni-axially oriented polycrystalline.

3.2. FIB-TEM observation

Three types of cross-sections were fabricated by FIB froma dimple, a dome and the closest ordered nacreous layers to thegrowth front. The fabricated dimple where the initial inorganicmaterials began precipitating is shown in Fig. 3a. Dark field imagesin TEM indicate that the inorganic materials are fine crystalsbecause of their bright contrast (Fig. 3b,c). Nano-diffractionpatterns from the crystallites showed that the mineral phase wasaragonite (Fig. 3d). Further diffraction analyses indicated that thecrystallites did not have specific crystal orientation. In particular, c-axes were not always oriented perpendicular to the shell surface. It

is also noted that these crystallites are not always on the surface ofthe organic matrix, but often buried in the matrix.

The cross-sectional TEM specimen from a dome is shown inFig. 4a. The contrasts due to the different crystallographic orien-tations indicate that the dome was polycrystalline with no inter-lamellar organic sheet. The contrasts show a horn-like shape fromthe bottom to the surface. This is generally consistent with the SEMobservation of a chemically etched cross section of the shell byDauphin et al. [15], who described the structure as ‘‘fibrousaragonite’’. Diffraction patterns near the surface were obtainedfrom the selected area around 100 nm (Fig. 4b,c). The c-axes werealmost perpendicular to the dome surface and a- and b-axes wereoriented differently. This result is consistent with the above EBSDanalyses.

The closest ordered nacreous layers to the growth front, whichlocated w150 mm far inside from the growth front, were fabricatedby FIB micro-sampling technique. About 8–10 interlamellar organicsheets were observed in this cross section (Fig. 5a). The lowerorganic sheets were discontinuous laterally as the distinct darkdiffraction contrast from aragonite crystals, whose c-axes weregenerally perpendicular to the interlaminated layers, werecontinued from the lower horn-like crystals to the upper laminatedones. This indicates that the oriented c-axes perpendicular to theshell surface in the plate-stacked nacreous layer inherit those in thedome through these continuous mineral parts, which are consid-ered to be mineral bridges. The width of the mineral bridges variedfrom several hundreds to several tens nanometers across the crosssection (Fig. 5a,d,e), though it is very difficult to estimate the realsize of the mineral bridges from the two dimensional TEM obser-vations. When the organic sheets become continuous and thick,these ‘‘mineral bridges’’ tended to be narrower (Fig. 5e) and manyconvex structures were observed on the surface of aragonite layers.

4. Discussion

The above results have shown that inorganic constitute of thenacreous layer initiates as precipitates inside the small dimplesof the organic matrix between prismatic columns. Theseprecipitates were aragonite crystallites with random orientationand any clue for the existence of amorphous precursors whichWeiss et al. [18] proposed was not found. Interestingly, thesecrystallites are not formed on the surface of the organic matrixbut tangled with the matrix. These results imply that nucleationof calcium carbonate crystals are not so simple as explained byorganic template or hetero-epitaxy at planar organic–inorganicinterface. Instead, the observed structure reminds us ionotropical

Fig. 5. (a) TEM bright field image of the cross section of the laminated nacreous layers. The asterisk (*) indicates the initial nacreous layer. (b,c) The diffraction patterns from thecircled regions in (a). Both diffraction patterns show that the aragonite c-axes are almost perpendicular to the nacreous layers. (d) Enlarged TEM image of the white dashed box d in(a), presenting a mineral bridge with w100 nm in width (the white arrow). (e) Enlarged TEM image between two nacreous layers with same crystallographic orientations, which islocated at the white dashed box e in (a). The crystallographic orientation is shown in the inset SAED pattern. Upper and lower nacreous layers show convex surface structure (theblack arrows). (f) HRTEM image at the part surrounded by the white box in (e), indicating that the lattice fringe is continued through the bridge.

K. Saruwatari et al. / Biomaterials 30 (2009) 3028–30343032

nucleation of the crystal [19,20]. Ionotropy is the accumulationmechanism of the ions required for inorganic crystallization ina confined space of organic matrix whose surface has highdensity of calcium biding sites, like carboxyl groups [19,20].

According to Dauphin et al. [15], the organic matrix between theprismatic and nacreous layers was compositionally similar tothe organic wall between prismatic layers, indicating that thearagonite nucleation occurred on the similar organic matrix that

a

b

Fig. 6. (a) Schematic drawing of the growth process of the initial nacreous tablet.Aragonite crystals nucleated inside of the dimple structure and the complete initialnacre was grown to the tablet shape with geometrical selection following theconcentric growth. (b) A model of the nacre growth deriving the crystallographicorientation from the initial tablet to the laminated nacreous layers via the initialnacreous layer. Thick discontinuous lines indicate interlamellar organic sheets. Somecrystallographic orientations were weeded out and some were inherited, inducing theless various crystallographic orientations for the upper laminated layers than thevarious ones for the lower domes.

K. Saruwatari et al. / Biomaterials 30 (2009) 3028–3034 3033

is related for the calcite prismatic formation. Thus, the ionotropicnucleation may be occurred by the assist of the acidic organicmacromolecules such as nacrein e.g., [21,22] bound at thedimples on the organic matrix, though we do not know how thedimples are developed on the organic matrix.

The aragonite precipitates concentrically grow to the largerdomes. Each dome is composed of polycrystalline aragonite whichshowed a horn-like form enlarging from the bottom to the surfaceand whose c-axes were perpendicular to the surface. This indicatesthat the nucleated aragonite polycrystalline with random crystal-lographic orientations develop into larger crystals with alignedcrystallographic orientations. This crystal growth is reasonablyexplained by a geometrical selection theory (Fig. 6a). In general,growth rates of a crystal are different along various crystallographicorientations. The crystals favorably oriented with maximumgrowth rate can grow largely, whereas the crystals unfavorablyoriented cannot grow by the interruption of the neighboringcrystals. As the result, preference of crystallographic orientationsoccurs during crystal growth. This phenomenon is called geomet-rical selection [23] and observed in several biominerals such as

prismatic layers [24] and eggshells [25]. In the case of aragonite, thefastest crystal growth is known to occur in the c-axis direction.The observed horn-like crystals oriented towards the c-axis in thepolycrystalline dome conform to this model.

At the further growing stage when the domes begin connectingwith the adjacent ones, organic sheets probably begin beingconstantly produced or propagating from the inward to the growthfront to regulate the laminated layer structures. The schematicmicrostructure of nacres at the growth front is presented in Fig. 6b.As shown in Fig. 5a, the organic sheets above the initial nacre arediscontinuous and the horn-like crystals in the initial nacrecontinue upward passing through discontinuous parts of theorganic sheets, which are restated as mineral bridges. Concerningthe upper laminated layers, the size of mineral bridges likely variesdepending on the thickness and the continuity of the organic sheet.For instance, mineral bridges became very narrow in thick organicsheets and convex structures were prominent on the surface of thearagonite plates (Fig. 5f). The narrow mineral bridges and theconvex structures possibly result from the minuscule pores andthe roughness of organic sheets in bivalves, respectively, asobserved by Carwright and Checa [26]. As they pointed out that thesize of mineral bridge was important to control the crystallographicorientations, it is a question whether all size of these mineralbridges contribute the formation of the new nacreous layersinheriting the crystallographic orientations. The narrower mineralbridges might happen to be connected through the pores in theorganic sheets after the growth.

Among the several growth mechanisms of nacre proposed so far,the mineral bridge model [11] is most suitable to explain thepresent microstructures as discussed above. Neither templatemodel [6,27] nor compartment model [28] are fit for the presentresults because the organic sheets were not developed in the firstnacreous layer and discontinuous above the first one. However,since we only investigated near the growth front of nacre, it is stillan open question whether the mineral bridge model works overallthe nacreous formation. The growth mechanism might vary withthe maturation of nacre. In order to reach a conclusion, furtherresearches and observations for the whole nacreous layers areneeded.

5. Conclusions

At the growth front in pearl oyster, P. fucata, nano-sizedaragonite crystals nucleated with random crystallographic orien-tation inside the dimples on the surface of the organic matrix thatcovers the outer prismatic columns. As aragonite crystals grewfilling the dimples and coalescing to form the initial nacreouslayer, they became horn-like shapes and their c-axes were pref-erentially oriented perpendicular to the surface. At the earlierlaminated nacreous layers just above the initial nacre, aragonitecrystals were continued from the horn-like shape to the severalupper laminated layers through discontinuity or holes in theorganic sheets. These results indicate that the crystal growth ofaragonite at the growth front is controlled by the inorganicprocesses such as geometrical selection and mineral bridges inaddition to the organic ones.

Acknowledgments

This work was supported by a Grant-in-Aid for scientificresearch (No. 17GS0311) from the Ministry Education, Culture,Sports, Science and Technology of Japan.

K. Saruwatari et al. / Biomaterials 30 (2009) 3028–30343034

Appendix

Figures with essential colour discrimination. Certain figures inthis article, in particular Figs. 1, 2 and 6, are difficult to interpret inblack and white. The full colour images can be found in the on-lineversion, at doi:10.1016/j.biomaterials.2009.03.011.

References

[1] Levi-Kalisman Y, Falini G, Addadi L, Weiner S. Structure of the nacreousorganic matrix of a bivalve mollusk shell examined in the hydrated state usingcryo-TEM. J Struct Biol 2001;135:8–17.

[2] Nudelman F, Gotiv BA, Addadi L, Weiner S. Mollusk shell formation: mappingthe distribution of organic matrix components underlying a single aragonitetablet in nacre. J Struct Biol 2006;153:176–87.

[3] Bezares J, Asaro RJ, Hawley M. Macromolecular structure of the organicframework of nacre in Halitis rufescens: implications for the growth andmechanical behavior. J Strcut Biol 2008;163:61–75.

[4] Wada K. Nucleation and growth of aragonite crystals in the nacre of somebivalve mollusks. Biomineralization Res Rep 1972;X:141–59.

[5] Chateigner D, Hedegaard C, Wenk HR. Mollusc shell microstructures andcrystallographic textures. J Struct Geol 2000;22:1723–35.

[6] Weiner S, Traub W. Macromolecules in mollusk shells and their functions inbiomineralization. Phil Trans R Soc London Ser B 1984;304:421–38.

[7] Addadi L, Joester D, Nudelman F, Weiner S. Mollusk shell formation: a sourceof new concepts for understanding biomineralization processes. Chem Eur J2006;12:980–7.

[8] Oaki Y, Imai H. The hierarchical architecture of nacre and its mimetic material.Angew Chem Int Ed 2005;44:6571–5.

[9] Rousseau M, Lopez E, Stempfle P, Brendle M, Franke L, Guette A, et al. Multi-scale structure of sheet nacre. Biomaterials 2005;26:6254–62.

[10] Marin F, Luquet G, Marie B, Medakovic D. Molluscan shell proteins: primarystructure, origin, and evolution. Curr top dev biol 2008;80:209–76.

[11] Shaffer TE, Inoescu-Zanetti C, Proksch R, Fritz M, Walters DA, Almqvist N, et al.Does abalone nacre form by heteroepitaxial nucleation or by growth throughmineral bridge? Chem Mat 1997;9:1731–40.

[12] Song F, Zhang XH, Bai YL. Microstructure and characteristics in the organicmatrix layers of nacre. J Mater Res 2002;17(7):1567–70.

[13] Lin AYM, Chen PY, Meyers MA. The growth of nacre in the abalone shell. ActaBiomat 2008;4:131–8.

[14] Checa GA, Rodriguez-Navarro AB. Self-organization of nacre in the shells ofPterioda (Bivalvia; Mollusca). Biomaterials 2005;26:1071–9.

[15] Dauphin Y, Ball AD, Cotte M, Cuif JP, Meibom A, Salome M, et al. Structure andcomposition of the nacre-prisms transition in the shell of Pinctada margarir-ifera (Mollusca, Bivalvia). Anal Bioanal Chem 2008;390:1659–69.

[16] Checa AG, Okamoto T, Rodriguez-Navarro AB. Organization pattern of nacre inPteriidae (Bivalvia: Mollusca) explained by crystal competition. Proc R Soc B2006;273:1329–37.

[17] Kogure T. A program to assist Kikuchi pattern analysis. J Crystal Soc Japan2003;45:391–5. Available from: http://www.gbs.eps.s.u-tokyo.ac.jp/kogure/edana, (in Japanese with English abstract).

[18] Weiss IM, Tuross N, Addadi L, Weiner S. Mollusc larval shell formation:Amorphous calcium carbonate is a precursor phase for aragonite. J Exp Zool2002;293:478–91.

[19] Greenfield EM, Wilson DC, Crenshaw MA. Ionotropic nucleation of calciumcarbonate by molluscan matrix. Amer Zool 1984;24:925–32.

[20] Mann S. Biomineralization, principles and concepts in bioinorganic materialschemistry. Oxford: Oxford University Press; 2001.

[21] Takeuchi T, Endo K. Biphasic and dually coordinated expression of the genesencoding major shell matrix proteins in the pearl oyster Pinctada fucata.Marine Biotech 2006;8:52–61.

[22] Gong NP, Shangguan JL, Liu XJ, Yan ZG, Ma ZJ, Xie LP, et al. Immunolocalizationof matrix proteins in nacre lamellae and their in vivo effects on aragonitictablet growth. J Struct Biol 2008;164:33–40.

[23] Grigor’ev DP. Ontogeny of minerals. Jerusalem: Israel Program for ScientificTranslations; 1965.

[24] Ubukata T. Architectural constraints on the morphogenesis of prismaticstructure in bivalvia. Palaeontology 1994;34:241–61.

[25] Garcıa-Ruiz JM, Rodriguez-Navarro A, Kalin O. Textural analysis of eggshells.Mater Sci Eng 1995;C3:95–100.

[26] Carwright JHE, Checa AG. The dynamics of nacre self-assembly. J R SocInterface 2007;4:491–504.

[27] Erben HK, Watabe N. Crystal formation and growth in bivalve nacre. Nature1974;248:128–30.

[28] Bevelander G, Nakahara H. An electron microscope study of the formation of thenacreous layer in the shell of certain bivalve molluscs. Calc Tiss Res 1969;3:84–92.