ARTICLESPUBLISHED ONLINE: 30MARCH 2014 | DOI: 10.1038/NMAT3920

Pervasive nanoscale deformation twinning as acatalyst for e�cient energy dissipation in abioceramic armourLing Li and Christine Ortiz*

Hierarchical composite materials design in biological exoskeletons achieves penetration resistance through a variety ofenergy-dissipating mechanisms while simultaneously balancing the need for damage localization to avoid compromisingthe mechanical integrity of the entire structure and to maintain multi-hit capability. Here, we show that the shell of thebivalve Placuna placenta (∼99wt% calcite), which possesses the unique optical property of ∼80% total transmission ofvisible light, simultaneously achieves penetration resistance and deformation localization via increasing energy dissipationdensity (0.290 ±0.072 nJ µm−3) by approximately an order of magnitude relative to single-crystal geological calcite(0.034 ±0.013 nJ µm−3). P. placenta, which is composed of a layered assembly of elongated diamond-shaped calcite crystals,undergoes pervasive nanoscale deformation twinning (width ∼50nm) surrounding the penetration zone, which catalyses aseries of additional inelastic energy dissipating mechanisms such as interfacial and intracrystalline nanocracking, viscoplasticstretching of interfacial organic material, and nanograin formation and reorientation.

B iological exoskeletons achieve mechanical robustnessand penetration resistance through exquisite and diversestructural designs that incorporate, for example, hierarchy,

heterogeneity, multilayering, anisotropy, functionally gradedinterfaces, and sacrificial bonding1–9. Although the energydissipation mechanisms arising from these structural features arebeginning to be elucidated by numerous studies, only recentlyhas attention begun to focus on the simultaneous need for spatiallocalization of damage10. Generally, biological structural materialsexhibit increased penetration resistance and energy dissipation withincreased volume of deformation1,2,11,12. Conversely, deformationlocalization is essential to avoid degradation of the structuraland mechanical integrity of the entire structure, to maintainmulti-hit capability10. Particularly, in transparent armour systems,deformation over large areas increases light scattering and reducesvisibility13. Biological exoskeletons offer important potentialinsights and designs for how to achieve multi-hit capability viaincreased energy dissipation density and damage localization—akey challenge in engineered armour materials, which are oftenmade of ceramics and prone to radial cracking and catastrophicfracture13. In this paper, we investigate this topic using thefascinating nanocomposite bioceramic armour of the bivalvePlacuna placenta as a model system.

Recently, we elucidated the structure of the bivalve P. placenta14.The shell of P. placenta (mineral content: 98.9± 0.1 wt%) possessesthe unique optical property of ∼80% total transmission of visiblelight, which is enabled by its single foliated nano/microstructurecomposed of elongated diamond-shaped calcitic laths (thickness∼300 nm) arranged in a tiled assembly with ∼2 nm organicinterfaces (Fig. 1a), small intracrystalline organic inclusions(∼25 nm), a low overall shell thickness (0.5–1mm) and curvature,and lack of an external organic periostracum14. Similar to aragoniticnacre, each calcite lath diffracts as a single crystal, although

crystallographic misorientations are present among adjacentbuilding blocks14. The c-axes of these calcitic laths are tilted by24.4◦± 3.5◦ relative to their surface normal, which makes thesurfaces of the laths close to the {108} planes of calcite (the anglebetween the {108} and {001} lattice planes is 26.3◦ in calcite,Fig. 1b; refs 14,15). Given the presumed necessity of maintainingstructural integrity and robustness against predatory attacks, herewe explore the detailed mechanical behaviour of this system. Thegoals of this study were: to quantify the penetration resistance ofthe shell of P. placenta relative to its mineral constituent calciteusing instrumented nanoindentation; to identify the detailednanoscale deformation mechanisms in relation to the shell’s uniquecrystallographic and micro-/nanostructural characteristics, ascompared to single-crystal calcite, through electron microscopy;and to explore how P. placenta balances spatially driven energydissipating deformation mechanisms with preservation of thestructural, mechanical and optical integrity of the overall system(providing multi-hit capability and visibility) via increased energydissipation efficiency (per unit volume of material). Design ofsuch penetration-resistant materials, which have amplified energydissipation efficiency resulting in more localized deformation, hasgreat potential for the development of bio-inspired engineeringstructures with improved multi-hit capability and preserved opticaltransparency under impact, which is desirable for a variety of bothcommercial and military applications13,16.

Quantification of the mechanical properties of freshly cleavedP. placenta shells in comparison to single-crystal geological calcitewas carried out via instrumented nanoindentation with the loadingaxis perpendicular to their {108} planes (Methods). Using anobtuse Berkovich diamond tip, it was determined that P. placentaretains the modulus (Eo−p) of calcite (P. placenta: 71.1 ± 4.2GPa,calcite: 73.4 ± 1.7 GPa), whereas its hardness (Ho−p) is increasedby ∼50% (P. placenta: 3.5 ± 0.3GPa, calcite: 2.3 ± 0.1GPa),

Department of Materials Science and Engineering, Massachusetts Institute of Technology, Massachusetts 02139, USA. *e-mail: cortiz@mit.edu

NATUREMATERIALS | VOL 13 | MAY 2014 | www.nature.com/naturematerials 501

© 2014 Macmillan Publishers Limited. All rights reserved

ARTICLES NATUREMATERIALS DOI: 10.1038/NMAT3920

aWidth: ~6 μm

Thickness: ~300 nm Thickness of

intercrystalline organicinterface: ~1−2 nm

Diameter ofintracrystalline organic

inclusions: ~25 nm

Tip angle: ~10°

Length: ~140 μm

{108} planes

Foliated microstructure

Individual building block

Intra- and intercrystallineorganic materials

L

Load

ing

Load

ing

0.0 0.5 1.0 1.5 2.0 2.50

2

4

6

8

10

Load

(mN

)

Depth (μm)

Fi

di

P. placenta

Δd

Calcite

c

b

0 20 40 60 800.0

0.2

0.4

0.6

0.8

1.0

Rela

tive

freq

uenc

y

Tilting angle (°)

P. placentaCalcite

{001}∧{104}: 44.6°

{001}∧{108}: 26.3°

d

e

Ri RoCP. placenta

L

2 μm

L

Calcite10 mN

2 μm

f EO−PHO−Phmaxhdef

RoC/Ro

Fidi

ΔdEdissVdefediss

Relative values1 2 8 9 100

Berkovich

Conospherical

P. placenta

Δ

Avg. s.d.

Calcite

10 mN

Figure 1 | Microstructural/crystallographic features and mechanical behaviour of biogenic calcite in Placuna placenta in comparison to single-crystalgeological calcite. a, Schematic diagram (not to scale) of the foliated microstructure in P. placenta. ‘L’ refers the longitudinal direction of the laths. b, Tiltingangles of c axes of the calcitic laths with respect to the surface normal in the shell of P. placenta (black, as reported in ref. 14). A single-crystal calcite samplewas sectioned and polished along one of the {108} planes (red). Standard interplanar angles between {001} and {104}, and {001} and {108} planes incalcite are also indicated. c, Loading portions of multiple individual load–depth curves (conospherical diamond tip, semi-angle= 30◦, tip radius=∼1 µm,maximum load= 10 mN). Fi and di are the load and depth corresponding to the initial fracture event detected, and 1d represents ‘pop-in’ depth. d,e, SEMimages of indentation residues of P. placenta (d) and calcite (e). Three-dimensional parameters are defined: Ri, the radius of the inner indentation crater;Ro, the radius of the entire fracture pattern by fitting it with the smallest circle; C, the distance between the centres of the two fitted circles. f, Comparisonof the mechanical parameters of P. placenta and calcite. The values corresponding to calcite are normalized to 1. EO−P, 1= 73.42± 1.74 GPa;HO−P, 1= 2.34±0.10 GPa; hmax, 1= 1.53±0.13 µm; hdef, 1= 6.96± 1.29 µm; Ro, 1= 5.94±0.88 µm; C/Ro, 1= 0.54±0.06; Fi, 1= 5.34± 1.35 mN;di, 1= 675.2± 167.0 nm; 1d, 1= 129.0± 192.1 nm; 1Ediss, 1= 8.77± 1.71 nJ; Vdef, 1= 257.2±94.2 µm3; ediss, 1= 0.034± 0.013 nJ µm−3. Avg. s.d.,average standard deviation; hmax, maximum depth.

demonstrating its enhanced resistance to plastic deformation(Supplementary Fig. 1; ref. 17). Fracture and cracking was inducedby indentation with a sharp axisymmetric conospherical diamondindenter, which also provides a better approximation of thegeometry of predatory loading18 and avoids potential complicationsas a consequence of the anisotropic mechanical response of single-crystal calcite19. Load–depth curves of P. placenta and abioticcalcite using the conospherical indenter reveal displacement burstswith almost no increase in load (‘pop-in’ events) associated withdiscrete fracture events, although noticeable differences are present(Fig. 1c). Compared with calcite, the fractures of P. placenta areinitiated at lower loads, Fi, (P. placenta: 2.10± 0.58mN, calcite:5.34± 1.35mN), and depths, di, (P. placenta: 256.5 ± 72.6 nm,calcite: 675.2 ± 167.0 nm). However, the average ‘pop-in’ depth,1d , for P. placenta is much smaller than that for calcite (P.placenta: 38.7 ± 35.1 nm, calcite: 129.0 ± 192.1 nm), indicatingthat P. placenta fractures in a more graceful way. By integratingthe area under the load–depth hysteresis, the single-crystalcalcite was found to dissipate slightly higher total deformationenergy, 1Ediss, relative to P. placenta during one indentation cycle(P. placenta: 6.04± 0.48 nJ, calcite: 8.77± 1.71 nJ; refs 20,21).

Representative scanning electron microscopy (SEM) images ofthe indentation residues of the two samples are shown in Fig. 1d,e(for P. placenta and single-crystal calcite, respectively). The single-crystal calcite shows an anisotropic distribution of large radialcracks and fractured regions, whereas the P. placenta shell exhibits a

more isotropic, localized deformation with multiple small deflectedcracks and nanosized fractured pieces. Three parameters weredefined to quantitatively compare their fracture patterns: Ri, theradius of the inner indentation crater that was directly in contactwith the tip during the test; Ro, the radius of the entire fracturepattern by fitting it with a smallest-possible circle; C , the distancebetween the centres of the inner and outer circles (Fig. 1d). Despiteexhibiting similar Ri values (P. placenta: 1.11 ± 0.06 µm, calcite:1.36± 0.06 µm), P. placenta shows a much smaller overall fracturesize, Ro, (P. placenta: 3.37 ± 0.37 µm, calcite: 5.94 ± 0.88 µm),indicating its more localized deformation behaviour. Also, thefracture patterns of P. placenta are more isotropic, indicated byits low C/Ro ratio (P. placenta: 0.24 ± 0.10, calcite: 0.54 ± 0.06).For an ideally isotropic material, the centres of the inner andouter circles should coincide; C , and hence C/Ro, approach to zero.Supplementary Fig. 2 schematically compares the averaged sizeand orientation of the fracture patterns of the two samples on thesame scale.

Through cross-sectional transmission electron microscopy(TEM) analysis of the indentation zone (as discussed in detail later),it is possible to directly determine the depth to which the materialhas undergone permanent deformation (hdef, SupplementaryFig. 3). Hence, the total volume that was permanently deformed,Vdef, can be estimated using Vdef = πRo

2hdef/3, approximatingthe deformed volume as a conical shape. The energy dissipationdensity, ediss, (that is, energy dissipation per unit volume) can then

502 NATUREMATERIALS | VOL 13 | MAY 2014 | www.nature.com/naturematerials

© 2014 Macmillan Publishers Limited. All rights reserved

NATUREMATERIALS DOI: 10.1038/NMAT3920 ARTICLES

1 μm

200 nmL

100 nm

(018)

Zone axis: [881]

500 nm

Pt

2 μm

Indentationdirection

a1

a2

a3

c

{108}

{108}

{018} {118}

L

a = 4.988 Å c/8 = 2.133 Å

54.8°

23.2°

10 μm

10 μm 500 nm

2 nm

(128)

(110)

)

(018)(128)

(110)

Matrix

Matrix

Twin

Matrix

Twin

Twin

TwinL

L

L

TB: (018)

(012)

(006)Matrix

Twin

(014)

Zone axis: [100]

g

Organic interface

Organicinterface

a

e

f

b c d

g h

i

j

{018}

Twin (014)

(014)

(012)

Matrix

Zone axis: [100]

{118}

{018}{108}

⊗L ⊗L

⊗L⊗L

Zone axis: [881]

(012)

Figure 2 | Nanoscale deformation twinning in P. placenta. a, TEM image of the entire cross-section of the indentation zone (conospherical tip;semi-angle= 30◦; tip radius=∼1 µm; maximum load= 10 mN). The yellow dashed line marks the boundary between the plastically deformed regionclose to the indentation tip and surrounding undeformed regions. White arrows indicate the location of deformation twins. Inset: Top-view SEM image ofthe original indentation residue. The yellow solid line indicates the location and orientation of the TEM sample prepared by FIB. Pt, protective platinumlayer. b, TEM image showing deformation twinning bands with parallel boundaries running across the laths. White arrows indicate the interfacial openingsassociated with the twinning bands. c, Corresponding SAED patterns in the matrix (top) and twinned (bottom) regions with zone axis= [8̄81̄]. ‘Matrix’ inb,c refers to untwinned regions of the calcitic laths that maintain the original crystallographic orientation19,20. d, HRTEM image of the twinning boundary(TB) of {018} . e,f, Top-view (e) and cross-section-view (f) SEM images of microindentation residue (conospherical tip; semi-angle= 30◦;tip radius= 2 µm; maximum load= 500 mN). g, TEM image of multiple deformation twinning bands within the deformed zone shown in f. h,i, SEM imagesof deformation twinnings in multiple orientations induced by manually compressing the shell with a mortar and pestle. j, Schematic model of the threecrystallographically equivalent {018} twinning systems in calcite. ‘L’ refers to the longitudinal direction of the laths. In all figures the symbol⊗ indicates theorientation of L is into the page.

be estimated for P. placenta as 0.290 ± 0.072 nJ µm−3, which isapproximately an order of magnitude higher than that of calcite(0.034± 0.013 nJ µm−3, Supplementary Fig. 3). This findingindicates that P. placenta, although it incorporates only ∼1wt% oforganic materials, is much more efficient in dissipating energy onpenetration than its primary mineral constituent, calcite. Figure 1fcompares the key mechanical parameters of P. placenta and abioticcalcite by setting the values of calcite as the reference; it is observedthat P. placenta is mechanically superior in almost every aspect.

To identify the underlying deformation mechanisms, cross-sectional TEM imaging of indentation residues of P. placentawas carried out (Fig. 2a, Supplementary Fig. 4 and Methods).The permanently deformed region surrounding the indentationsite can be clearly distinguished from the undamaged region (asmarked by the dashed line) by image contrast and the disruptionof the organic interfaces between adjacent laths (Fig. 2a). Thispermanently deformed region also follows the V-shaped profileof the conospherical indenter, with a semi-angle of ∼60◦, andextends beyond the maximum indentation depth by ∼0.4 µm(less than two mineral layers) without any vertical cracks. Closeto the boundary of the inelastic deformation zone we observedplanar defects with two closely spaced parallel flat boundaries(Fig. 2a,b). These defects were identified as deformation twins byelectron microscopic imaging and diffraction, as described below

in detail. The twin bands are ∼50 nm in width and propagatethrough the entire thickness of each mineral lath at an angle of∼20◦–30◦, and are arrested by the organic interfaces (Fig. 2b).The twin bands in the same mineral lath are typically parallelto each other as a result of their single crystal nature, whereasthose from different laths are usually not parallel as a result ofcrystallographic misorientation. Furthermore, local enlargement oropening of the organic interfaces without catastrophic delaminationis usually observed at the terminations of twin bands as a result ofthe displacement mismatch between adjacent laths (Fig. 2b, whitearrows). Selected area electron diffraction (SAED) patterns obtainedfrom the untwinned (‘matrix’) and twinned regions clearly reveale-type deformation twinning (twin boundaries: {018}) associatedwith the rhombohedral crystal structure (space group, R3̄c) ofcalcite (Fig. 2c; refs 22,23). Such twin bands are not observedin undeformed regions. Figure 2d shows a high-resolution TEM(HRTEM) image of a twinning boundary corresponding to the[1̄00] zone axis and further illustrates themirror symmetry betweenthe twinned and untwinned regions at atomic resolution. With theindentation loads increased to 500mN, and even 2.5 N, P. placenta isstill able to confine the large inelastic deformation to a small volumewithout any radial cracks (Fig. 2e,f and Supplementary Fig. 5). TEManalysis of the deformed region again reveals a large population of{018} deformation twin bands that were present in the majority of

NATUREMATERIALS | VOL 13 | MAY 2014 | www.nature.com/naturematerials 503

© 2014 Macmillan Publishers Limited. All rights reserved

ARTICLES NATUREMATERIALS DOI: 10.1038/NMAT3920

⊗L

⊗L

⊗L

Pt

200 nm

Inde

ntat

ion

dire

ctio

ndb

c

200 nm

100 nm

L 2 μm

h

a b

c

⊗L

Grain 1

Grain 3

Grain 2

Am

orph

ous

Amorphous

(102)

(104)

(024)

2 nm

e g

h

0 120Height (nm)

200 nm

f

⊗L 100 nm

d

Zone axis: [551]

(110)

(105)

Figure 3 | Nanoscopic inelastic deformation in individual calcitic layers of P. placenta. a, Bright-field TEM image of the permanently deformed region closeto the indentation crater. Inset: Top-view SEM image of the original indentation residue. The yellow line indicates the location and orientation of the TEMsample. Pt, protective platinum layer. b,c, SAED patterns acquired in the deformed (b) and surrounding undeformed (maintaining original single crystalstructure, c) regions indicated by the circles in a. d, Dark-field TEM image of the deformed region (corresponding to the region in the white box in a) withthe corresponding selected di�raction spots indicated in b with the red circle. e, HRTEM image of misoriented calcite nanograins in the permanentlydeformed region. f, Tapping mode AFM height image of an indent corner (Berkovich tip) showing the flattening of nanoscopic asperities within anindentation crater. g,h, Bright-field TEM images showing crack deflection within individual laths (white arrows).

calcite crystals (Fig. 2g) andwhich are believed to play a primary rolein mitigating severe local deformation, especially in the region closeto the tip of the indenter. The twinning-induced interface openingcauses significant light scattering, thus reducing light transmissionin the deformation region (Supplementary Fig. 6). Manual grindingof the shell using a mortar and pestle can be also used to inducepervasive deformation twinning bands (Fig. 2h,i), which suggeststhat deformation twinning is not only induced by the high localstresses generated by an indenter tip24. Three equivalent twinningsystems with twinning boundaries of (11̄8), (018) and (1̄08) wereobserved, intersecting with the longitudinal direction of the laths at0◦, 54.8◦, and 120◦, respectively, as illustrated in the crystallographicconfiguration in Fig. 2j. The orientations of the prepared TEMsamples (perpendicular to the longitudinal direction of the laths)are close to the (1̄00) plane, and the twinning boundaries of (018)and (11̄8) create angles of 23.2◦ with the lath-organic interface,consistent with experimental observations (Fig. 2b,g).

In addition to deformation twinning, a series of other nanoscaleinelastic deformation mechanisms at the building block levelwere observed (Fig. 3). The bright-field TEM image of Fig. 3ashows that the deformed zone (containing area ‘b’) has afractionated brightness, in stark contrast to the more homogeneousfeatureless brightness of the undeformed region (containing area ‘c’),suggesting the formation of nanosized grains. SAED patterns reveal

that the deformed and undeformed regions exhibit polycrystalline-like (Fig. 3b) and single-crystal-like (Fig. 3c) patterns, respectively.Thus, deformation induces the formation of nanograins withcrystallographic misorientations. The dark-field TEM image ofFig. 3d, obtained using selected diffraction spots from Fig. 3b(indicated by the red circle), selectively lightens up some ofthe deformation-induced nanograins (∼50 nm in diameter). TheHRTEM image of Fig. 3e further reveals misoriented nanoscopicgrains, all of which were indexed to calcite, within the plasticallydeformed zone. Amorphous regions were also observed alongthe boundaries of grains, or occasionally entirely encapsulatedwithin grains, suggesting that individual calcite laths are capableof undergoing ductile deformation. This capability is furtherdemonstrated by the significant flattening of nanoscopic asperitieswithin the residual indentation area (Fig. 3f). Furthermore,crack deflections within individual building blocks were observed(Fig. 3g,h and Supplementary Fig. 7). These experimental resultsindicate that the calcite laths, despite their single-crystal-likenature, do not fracture catastrophically, but rather dissipateconsiderable energy through a series of nanoscopic inelasticdeformation processes8,25.

It has long been known that calcite deforms plastically viadeformation twinning and slip, primarily at elevated temperaturesand confined hydrostatic pressures, whereas at room temperature

504 NATUREMATERIALS | VOL 13 | MAY 2014 | www.nature.com/naturematerials

© 2014 Macmillan Publishers Limited. All rights reserved

NATUREMATERIALS DOI: 10.1038/NMAT3920 ARTICLES

• Deformation twinning• Interface opening

• Interface opening• Dislocation restriction• Micro-/nano-cracking

• Original structure • Crack growth/deflection• Fragmentation

• Grain rotation/reorientation• Amorphization

Inelastic deformation

Vertical crack

Large fracturedtwinning blocks

1 µm 1 µm

Maximum load = 10 mN

Stage 1 Stage 2 Stage 3 Stage 4

a

c

b

Lateral cracks

Maximum load = 10 mN

Figure 4 | Nanoscale deformation mechanisms in P. placenta and single-crystal calcite under indentation. a,b, Schematic of the deformation zone close tothe indenter for P. placenta (a) and single-crystal calcite (b; conospherical tip; semi-angle= 30◦; tip radius=∼1 µm; maximum load= 10 mN). Thediagrams are drawn to scale based on microscopic dimensional measurement. Dislocation arrays in calcite are not shown in b. c, TEM images illustratingthe progression of nanoscale deformation mechanisms of P. Placenta (listed below the images) with decreasing distance from the indenter. All scale bars,100 nm.

and atmospheric pressure, calcite readily fractures on compression,tension, and particularly indentation22,23. Bright-field andcorresponding dark-field TEM imaging of the indentationzone of calcite samples, which had undergone the same indentationloading conditions as P. placenta, reveal large fractured pieces (size>2 µm) and cracks (Supplementary Fig. 8), consistent with top-viewSEM observations (Fig. 1e). At the bottom of indentation craters, aprimary vertical crack is produced downwards (∼1–2 µmdeep) andmultiple lateral cracks are formed. As the load increases, the verticalcracks advance further downwards and the lateral cracks propagatenearly parallel to the surface to form large chipped pieces, similar toother brittle ceramic materials26. In contrast to P. placenta, extendeddislocation arrays were also developed underneath indentationcraters (Supplementary Fig. 9). These events result in the largepenetration depth (hdef), ‘pop-in’ depth (1d), fracture size (Ro),and volume of permanent deformation (Vdef) observed (Fig. 1f).TEM samples prepared along the longitudinal direction reveal avery limited number of e-type deformation twins, with only (1̄08)twinning boundaries under this loading condition (SupplementaryFig. 9). The width of the twin bands (∼200 nm) is usually muchlarger than those in P. placenta, inducing large cracks because oflarge local displacement incompatibilities. Owing to the absenceof effective crack arrestors and deflectors, such as the organicinterfaces in P. placenta, these cracks propagate to the surface,generating large chipped blocks.

A summary of the entire progress of deformation mechanismsin P. placenta compared to those of single-crystal calcite is providedin Fig. 4. On indentation, the calcitic laths in P. placenta shells firstundergo e-type twinning in all three crystallographically equivalentorientations [(018), (11̄8), and (1̄08)], whereas only one orientation(1̄08) is primarily activated in single-crystal calcite (Fig. 4a,b andSupplementary Fig. 9). This is possibly facilitated by the nanoscopic

thickness of the calcitic laths in the shell. Such grain-size effectsfor deformation twinning have been observed in some metallicnanocrystalline materials, although further reducing the grain sizebelow a certain critical size may also lead to increased resistanceto deformation twinning27,28. Whether the thickness of the calciticlaths (∼300 nm) is optimal to facilitate deformation twinningremains a question; further investigations, especially theoreticalmodelling, may provide more insights29. The nanoasperities onthe surface of laths might also assist the initiation of multipledeformation twinning bands with nanoscale spacing as a resultof roughness-induced stress concentrations during loading. Theas-formed twin bands make inclination angles with the surfacenormal and longitudinal direction of the laths of 26◦/55◦, 26◦/0◦and 52◦/120◦ for (018), (11̄8) and (1̄08) twin bands, respectively(Fig. 2j), which allows the biogenic calcite to mitigate stressconcentrations more effectively than abiotic calcite, leading toa more isotropic deformation response (Fig. 4c, stage 1). Thevariation of the thickness and crystal orientation of the calciticlaths allows a sequential activation of deformation twinning bandsdepending on the specific loading conditions, which is expectedto result in a work hardening effect30. Unlike the growth twins,these deformation-induced twins only activate when and wherethey are needed, that is, ahead of the impact region on penetration.Moreover, deformation twinning in P. placenta is activated onquasi-static deformation conditions, in contrast to the recentlyobserved deformation twinning in aragonitic nacre under highstrain rate deformations (∼103 s−1; ref. 9). The twin boundariessurrounding the penetration region act as effective barriers fordislocation motion and catastrophic crack propagation, leading toenhanced penetration resistance (Fig. 4c, stage 2; ref. 31). Whereasthe twinning bandsmainly act as barriers in the horizontal direction(Supplementary Fig. 10), the organic interfaces play a similar role

NATUREMATERIALS | VOL 13 | MAY 2014 | www.nature.com/naturematerials 505

© 2014 Macmillan Publishers Limited. All rights reserved

ARTICLES NATUREMATERIALS DOI: 10.1038/NMAT3920

in the vertical direction, resisting dislocation motion and crackpropagation from one layer to another (Supplementary Fig. 11).

The nanoscale twinning bands and organic interfaces increaseenergy dissipation density by acting as catalysts to promote othernanoscopic deformation mechanisms, including interfacial andintracrystalline nanocracking, organic viscoplastic stretching, aswell as nanograin formation and reorientation. The relatively smallamount of organicmaterial (∼1wt%) present in the shell is expectedto contribute to energy dissipation in a variety of ways. Thenanoscopic openings associated with the ends of twin bands resultin fibrillar viscoplastic stretching of intercrystalline organicmaterialbridging neighbouring calcite laths (Supplementary Fig. 12), leadingto increased energy dissipation (Fig. 4c, stage 2), which hasbeen previously observed in nacre32. It is also hypothesized thatthe nanoscopic intracrystalline organic inclusions, in additionto twinning boundaries, contribute to deflecting and arrestingintracrystalline cracks, which leads to fragmentation of the regularsingle-crystalline laths into irregular micro-/nanosized pieces(Fig. 4c, stage 3; ref. 31). Interactions between the accumulateddislocations and twinning boundaries might also contribute tothis fragmentation process30,31. Further compressive deformationleads to formation and reorientation of nanosized grains (incontrast to the observation of inherent nanograined structure innacre25), and the formation of amorphous regions. At this stage,the deformed material is transformed to an isotropic homogeneousnanocrystalline material, as the original laths and organic interfacesare destroyed (Fig. 4c, stage 4).

It is noted that the microscopic structural features responsiblefor pervasive nanoscale deformation twinning in P. placenta alsoexist in other calcitic bioceramic structural materials, includingother bivalve shells15, brachiopod shells33, and sea urchin spinesand teeth34,35. Therefore, it is possible that this phenomenon mayalso play a similar role in these systems. Also, some aragonite-based mollusk shells possess basic building blocks densely packedwith nanoscale {110} growth twins36,37. These pre-existing twinningboundaries may play a similar role in constraining the dislocationmotions and deflecting the propagation of micro-/nanocracks31.However, the displacement incompatibilities resulting from theformation of deformation twinning bands will not be prevalentin aragonite structures with growth twins; therefore, they are notexpected to effectively catalyse other nanoscopic energy dissipationprocesses in a fashion similar to P. placenta.

Energy dissipation enhancement resulting from spreading thedeformation zone over a large area/volume has been observedin a variety of biological structural material systems throughdifferent mechanisms, such as the tablet interlocking of nacre11,crack deflection and bridging of crossed-lamellar structure1,2, andnanoscale heterogeneity of bone12. As well as maximizing the totalenergy dissipation on deformation, maintaining the structural andmechanical integrity of the entire structure is also critical to theperformance of biological exoskeletons during predatory attacks.Damage localization in the protective scales of a ‘living fossil’Polypterus senegalus is achieved throughpreferential circumferentialcracking rather than more damaging radial cracking, which resultsfrom the complex quad-layered design of the scales10. In thiswork, we demonstrated that the bioceramic armour of P. placenta,despite its very high mineral content and relatively monolithiccomposition and structure, is able to achieve a great balancebetween energy dissipation and damage localization by efficientenergy dissipation in a confined volume via pervasive nanoscaledeformation twinning surrounding the deformation zone. A seriesof additional nanoscopic deformation mechanisms, includinginterface opening and viscoplastic stretching of organic materials,crack deflection within individual building blocks, fragmentation,nanograin formation and reorientation, and amorphization, worksynergistically to increase the energy dissipation density by almost

an order of magnitude relative to pure calcite. The nanoscalestructural and crystallographic architecture in this biologicalnanocomposite determine the activation and control of theseefficient energy dissipation mechanisms. Apart from deformationlocalization, which maintains the structural and mechanicalintegrity of the entire system (multi-hit capability), the opticalproperties (∼80% total transmission of visible light) are alsopreserved at distances away from the penetration site. The findingsin this work may provide design principles for engineeringlightweight structural materials with efficient energy dissipation, inparticular transparent armour, through control and design of thematerial systems at the nanometre scale.

MethodsSamples. Edge-trimmed and intact P. placenta shells were purchased fromSeashell World and Conchology, respectively. Samples of single crystal geologicalcalcite (origin, Mexico) were obtained from Pisces Trading Company, LLC.

Nanoindentation. P. placenta shells were cleaved using a razor blade and testedimmediately. Single-crystal calcite samples were sectioned using a diamond saw(Buehler, Isomet 5000 Lake Bluff) along one of their {108} planes (Fig. 1b), andpolished on a polishing wheel (South Bay Technology, Model 920) withaluminium oxide pads stepwise (15, 6, 3, and 1 µm), and finally with 50 nm silicananoparticles on a microcloth (South Bay Technology). Nanoindentationexperiments were conducted in ambient conditions using a TriboIndenter(Hysitron). Load-controlled nanoindentation was performed using Berkovich(trigonal pyramid, semi-angle = 65.3◦) and conospherical (tip radius = ∼1 µm,semi-angle = 30◦) diamond probe tips. The piezoelectric transducer was firstallowed to equilibrate for 105 s (the last 45 s with digital feedback) and another40 s for calculating drift automatically before each indent. Typical load functionsincluded loading (10 s), holding (20 s) and unloading (10 s). Maximum loadsvaried from 1 to 10mN. The Oliver–Pharr (O–P) methodology was used toquantify material properties—that is, indentation modulus (EO–P) and hardness(HO–P; ref. 17). The probe tip area function A(hc), which is the projected area ofthe indentation tip as a function of the contact depth hc, and frame compliancewere calibrated before each set of experiments using a fused quartz sample.Indentation experiments with maximum loads higher than 10mN were carriedout using a Micro Materials microindenter (Wrexham).

Electron microscopy. Samples were coated with ultra-thin carbon to reducecharging effects prior to SEM imaging. Samples were imaged using a HeliosNanolab 600 Dual Beam electron microscope (FEI, OR) at acceleration voltagesof 2 and 5 kV and a working distance of ∼4mm. Cross-sectional samples andTEM samples were prepared using ion beam milling with the same instrument. Adetailed TEM sample preparation procedure is as follows: a platinum protectivelayer (∼0.5 µm) was first laid down on top of the desired region; anotherplatinum protective layer (∼1.5 µm) was further deposited on top of the regionwhere the TEM slab was to be milled out; two trenches, one on each side of theplatinum protective stripe, were milled by FIB, leaving the slab of specimen(thickness: ∼1.5 µm); the slab was then cut through by FIB and transferred to acopper TEM grid by an Omniprobe and welded securely with platinumdeposition; the lift-out lamellar of specimen was sequentially thinned by FIB at30, 16, 5, and 2 kV ion beam voltages. Final cleaning at 2 kV and 28 pA isimportant to obtain a clean surface and minimize damage. TEM imaging withtypical bright-field, dark-field, and SAED techniques was carried out using aJEOL 2011 operated at 120 kV. The image magnification and camera constantswere calibrated using a standard sample (MAG*I*CAL, Electron MicroscopySciences, PA). HRTEM imaging was carried out using a field emission JEOL2010F at 200 kV. A gold foil standard (Lattice plane resolution test-646, Pelco)was first used to align the instrument before P. placenta and calcite samples. Tominimize electron beam damage, the correct zone axis was first found using areasaway (∼2–3 µm in distance) from the twin boundary, but on the same lath, usingKikuchi patterns, taking advantage of the fact that each lath in P. placentadiffracts as a single crystal. Once the correct zone axis was found and imagingconditions were optimized, the twin boundary was brought into the field of viewwith the beam spread. The beam was then focused and an image was takenimmediately (typical exposure, ∼0.5 s). Usually only two images could be taken atone boundary before the region was damaged.

Atomic force microscopy. Tapping mode AFM (TMAFM) imaging in ambientconditions was carried out using a Digital Instruments Multimode SPM IIIA(Veeco) with an AS-130 ‘JV’ scanner. TMAFM imaging was conducted withNANOSENSORS Si TMAFM cantilevers (PPP-NCHR-10). Typical scan speedswere 1–5 µms−1; other parameters were optimized on tuning.

506 NATUREMATERIALS | VOL 13 | MAY 2014 | www.nature.com/naturematerials

© 2014 Macmillan Publishers Limited. All rights reserved

NATUREMATERIALS DOI: 10.1038/NMAT3920 ARTICLESElectron backscattered diffraction. The pre-cut square-shaped specimens fromP. placenta shells were cleaved, fixed onto steel plates, and subsequently coatedwith ultra-thin carbon. EBSD analysis was carried out using a FEI HeliosFIB/SEM system equipped with an HKL Technology ‘Channel 5’ EBSD system.The sample was mounted on a 70◦ pre-tilted EBSD stage and the workingdistance was 6mm. SEM images and EBSD patterns were generated using anaccelerating voltage of 20 kV and a beam current of 2.7 nA. The typical scanningstep size was 1 µm and EBSD patterns with a mean angular uncertainty of 1◦ andabove were discarded. Scans with at least 80% of points indexed were furtheranalysed using HKL software.

Received 4 October 2013; accepted 20 February 2014;published online 30 March 2014

References1. Kamat, S., Su, X., Ballarini, R. & Heuer, A. H. Structural basis for the fracture

toughness of the shell of the conch Strombus gigas. Nature 405,1036–1040 (2000).

2. Kamat, S., Kessler, H., Ballarini, R., Nassirou, M. & Heuer, A. H. Fracturemechanisms of the Strombus gigas conch shell: II-micromechanics analyses ofmultiple cracking and large-scale crack bridging. Acta Mater. 52,2395–2406 (2004).

3. Aizenberg, J. et al. Skeleton of Euplectella sp: Structural hierarchy from thenanoscale to the macroscale. Science 309, 275–278 (2005).

4. Weaver, J. C. et al. The stomatopod dactyl club: A formidable damage-tolerantbiological hammer. Science 336, 1275–1280 (2012).

5. Ortiz, C. & Boyce, M. C. Biological structural materials. Science 319,1053–1054 ( 2008).

6. Dunlop, J. W. C. & Fratzl, P. Biological composites. Annu. Re. Mater. Res. 40,10.1–10.24 (2010).

7. Yang, W. et al. Nature flexible dermal armor. Adv. Mater. 25, 31–48 (2013).8. Gao, H., Ji, B., Jäger, I. L., Arzt, E. & Fratzl, P. Materials become insensitive to

flaws at nanoscale: Lessons from nature. Proc. Natl Acad. Sci. USA 100,5597–5600 (2003).

9. Huang, Z. et al. Uncovering high-strain rate protection mechanism in nacre.Sci. Rep. 1, 148 (2011).

10. Bruet, B. J. F., Song, J., Boyce, M. C. & Ortiz, C. Materials design principles ofancient fish armour. Nature Mater. 7, 748–756 (2008).

11. Barthelat, F., Tang, H., Zavattieri, P. D., Li, C-M. & Espinosa, H. D. On themechanics of mother-of-pearl: A key feature in the material hierarchicalstructure. J. Mech. Phys. Solids 55, 306–337 (2007).

12. Tai, K., Dao, M., Suresh, S., Palazoglu, A. & Ortiz, C. Nanoscale heterogeneitypromotes energy dissipation in bone. Nature Mater. 6, 454–462 (2007).

13. Sands, J. M., Patel, P. J., Dehmer, P. G., Hsieh, A. J. & Boyce, M. C. Transparentarmour materials.Military Tech. 12, 82–91 (2008).

14. Li, L. & Ortiz, C. Biological design for simultaneous optical transparency andmechanical robustness in the shell of Placuna placenta. Adv. Mater. 25,2344–2350 (2013).

15. Checa, A. G., Esteban-Delgado, F. J. & Rodríguez-Navarro, A. B.Crystallographic structure of the foliated calcite of bivalves. J. Struct. Biol. 157,393–402 (2007).

16. Arciszewski, T. & Cornell, J. Intelligent Computing in Engineering andArchitecture 32–53 (Springer, 2006).

17. Oliver, W. C. & Pharr, G. M. An improved technique for determining hardnessand elastic modulus using load and displacement sensing indentationexperiments. J. Mater. Res. 7, 1564–1583 (1992).

18. Song, J., Ortiz, C. & Boyce, M. C. Threat-protection mechanics of an armoredfish. J. Mech. Behav. Biomed. Mater. 4, 699–712 (2011).

19. Kunitake, M. E., Mangano, L. M., Peloquin, J. M., Baker, S. P. & Estroff, L. A.Evaluation of strengthening mechanisms in calcite single crystals from molluskshells. Acta Biomater. 9, 5353–5359 (2013).

20. Bruet, B. J. F. et al. Nanoscale morphology and indentation of individual nacretablets from the gastropod mollusc Trochus niloticus. J. Mater. Res. 7,2400–2419 (2005).

21. He, L. H. & Swain, M. V. Energy absorption characterization of human enamelusing nanoindentation. J. Biomed. Mater. Res. A 81, 484–492 (2007).

22. Turner, F. J., Griggs, D. T. & Heard, H. Experimental deformation of calcitecrystals. Bull. Geo. Soc. Amer. 65, 883–934 (1954).

23. Barber, D. J. & Wenk, H. R. Deformation twinning in calcite, dolomite,and other rhombohedral carbonates. Phys. Chem. Minerals 5,141–165 (1979).

24. Chen, M. et al. Deformation twinning in nanocrystalline aluminum. Science300, 1275–1277 (2003).

25. Li, X., Xu, Z-H. &Wang, R. In-situ observation of nanograin rotation anddeformation in nacre. Nano Lett. 6, 2301–2304 (2006).

26. Oyen, M. L. & Cook, R. F. A practical guide for analysis of nanoindentationdata. J. Mech. Behav. Biomed. Mater. 2, 396–407 (2009).

27. Yu, Q. et al. Strong crystal size effect on deformation twinning. Nature 463,335–338 (2010).

28. Zhu, Y. T., Liao, X. Z. & Wu, X. L. Deformation twinning in nanocrystallinematerials. Prog. in Mater. Sci. 57, 1–62 (2012).

29. Zlufarska, I., Nakano, A. & Vashishta, P. A crossover in the mechanicalresponse of nanocrystalline ceramics. Science 309, 911–914 (2005).

30. Li, Y. S., Tao, N. R. & Lu, K. Microstructural evolution and nanostructureformation in copper during dynamic plastic deformation at cryogenictemperatures. Acta Mater. 56, 230–241 (2008).

31. Lu, K., Lu, L. & Suresh, S. Strengthening materials by engineering coherentinternal boundaries at the nanoscale. Science 324, 349–352 (2009).

32. Smith, B. L. et al.Molecular mechanistic origin of the toughness of naturaladhesives, fibres and composites. Nature 399, 761–763 (1999).

33. Goetz, A. J. et al. Interdigitating biocalcite dendrites form a 3-D jigsawstructure in brachiopod shells. Acta Biomater. 7, 2237–2243 (2011).

34. O’Neill, P. L. Polycrystalline echinoderm calcite and its fracture mechanics.Science 213, 646–648 (1981).

35. Wang, R. Z., Addadi, L. & Weiner, S. Design strategies of sea urchin teeth:Structure, composition and micromechanical relations to function. Phil. Trans.R. Soc. Lond. B 352, 469–480 (1997).

36. Suzuki, M., Kim, H., Mukai, H., Nagasawa, H. & Kogure, T. Quantitative XRDanalysis of {110} twin density in biotic aragonites. J. Struct. Biol. 180,458–468 (2012).

37. Younis, S., Kauffmann, Y., Pokroy, B. & Zolotoyabko, E. Atomic structureand ultrastructure of theMurex troscheli shell. J. Struct. Biol. 180,539–545 (2012).

AcknowledgementsWe gratefully acknowledge the support of the National Science Foundation through theMIT Center for Materials Science and Engineering (DMR-0819762), the US ArmyResearch Office through the MIT Institute for Soldier Nanotechnologies (ContractW911NF-07-D-0004), the National Security Science and Engineering Faculty FellowshipProgram (N00244-09-1-0064), and the Office of Assistant Secretary of Defense forResearch and Engineering. The authors would like to thank J. Li, Y. Zhu, M. Dao and A.Schwartzman for fruitful general discussions of the manuscript. The authors would liketo thank J. C. Weaver for his assistance in scanning electron microscopy imaging. Theauthors would also like to thank L. Han, A. Schwartzman, S. Chen, Y. Zhang and M. J.Connors for their technical assistance.

Author contributionsL.L. and C.O. designed the research, analysed the data and wrote the manuscript. L.L.conducted the experiments

Additional informationSupplementary information is available in the online version of the paper. Reprints andpermissions information is available online at www.nature.com/reprints.Correspondence and requests for materials should be addressed to C.O.

Competing financial interestsThe authors declare no competing financial interests.

NATUREMATERIALS | VOL 13 | MAY 2014 | www.nature.com/naturematerials 507

© 2014 Macmillan Publishers Limited. All rights reserved