Professional Documents
Culture Documents
Tuning Hardness in Calcite by Incorporation of Amino Acids
Tuning Hardness in Calcite by Incorporation of Amino Acids
Structural biominerals are inorganic/organic composites that exhibit remarkable mechanical properties. However, the
structure–property relationships of even the simplest building unit—mineral single crystals containing embedded
macromolecules—remain poorly understood. Here, by means of a model biomineral made from calcite single crystals
containing glycine (0–7 mol%) or aspartic acid (0–4 mol%), we elucidate the origin of the superior hardness of biogenic calcite.
We analysed lattice distortions in these model crystals by using X-ray di�raction and molecular dynamics simulations, and by
means of solid-state nuclear magnetic resonance show that the amino acids are incorporated as individual molecules. We also
demonstrate that nanoindentation hardness increased with amino acid content, reaching values equivalent to their biogenic
counterparts. A dislocation pinning model reveals that the enhanced hardness is determined by the force required to cut
covalent bonds in the molecules.
B
iominerals such as bones, teeth and seashells are characterized retarding dislocation motion13 . Biominerals, notably calcite, often
by properties optimized for their functions. Despite being deform plastically by twinning12 , but because twins grow by motion
formed from brittle minerals and flexible polymers, nature of ‘twinning dislocations’14 , these concepts also broadly apply to
demonstrates that it is possible to generate materials with strengths twin formation. To fully understand the hardening mechanisms
and toughnesses appropriate for structural applications1 . At one of single-crystal biominerals, one must determine the relationship
level, the mechanical properties of these hierarchically structured between the hardness and the concentrations of the different types
materials are modelled as classical composites consisting of a of guest species. For example, the effect of Mg2+ substitutions was
mineral phase embedded in an organic matrix2 . However, the shown to be consistent with solute strengthening15 . Determination
single-crystal mineral building blocks of biominerals are also of the hardening mechanism of occluded organic additives has
composites3 , containing aggregates of biomacromolecules as large as proved far more challenging. Although species ranging from small
20 nm (refs 4,5) and inorganic impurities6,7 . Although it should be molecules16,17 to peptides18 , proteins19,20 , nanoparticles9,10,21,22 and
entirely possible to employ this simple biogenic strategy in materials fibres23,24 have been incorporated in calcite, the effect of these
synthesis8–10 , the strengthening and toughening mechanisms that inclusions on mechanical properties is not yet known. Occlusion
result from these inclusions are still poorly understood11,12 . This of 200-nm latex particles within calcite single crystals was shown
work addresses this challenge by analysing hardening mechanisms to reduce their hardness8 , whereas the incorporation of polymeric
in a simple model biomineral system: calcite single crystals micelles having sizes comparable to those of the protein occlusions
containing known amounts of amino acids. We report the formation in biominerals increased hardness9 . However, owing to an inability
of synthetic calcite crystals with hardnesses equivalent to those of to control the number of micelles occluded, it was not possible to
their biogenic counterparts, and offer a detailed explanation for the quantitatively characterize their hardening effect.
observed hardening. The present study describes a model system—created by
Because plastic deformation in single crystals occurs by the the incorporation of the amino acids aspartic acid (Asp) and
motion of dislocations, hardness is enhanced by features that glycine (Gly) within calcite—in which we can precisely tune the
inhibit dislocation motion. The mechanisms by which guest species compositions of our single-crystal composites over a wide range,
may harden ionic single crystals generally fall into two categories. and thus finally determine the origin of the hardening effects
Second-phase particles directly block dislocation motion, such that of small organic molecules. By application of X-ray diffraction
a dislocation must either cut through (shear) a particle or bypass it and molecular dynamics simulations to characterize the local and
by a diffusive process in order to keep going13 . Solutes (point defects) global distortions of the crystal lattice, solid-state nuclear magnetic
do not directly block dislocation motion, but the stress fields of the resonance (ssNMR) to demonstrate that the amino acids are
dislocations interact with those associated with misfitting solutes, present as individual species within the crystal, nanoindentation
1 School of Chemistry, University of Leeds, Woodhouse Lane, Leeds LS2 9JT, UK. 2 Department of Materials Science and Engineering, 214 Bard Hall, Cornell
University, Ithaca, New York 14853, USA. 3 BioArCh, Departments of Chemistry and Archaeology, University of York, York YO10 5DD, UK. 4 Department of
Materials Science and Engineering, University of She�eld, Mappin Street, She�eld S1 3JD, UK. 5 Department of Chemistry, University of Cambridge,
Lensfield Road, Cambridge CB2 1EW, UK. 6 Diamond Light Source, Harwell Science and Innovation Campus, Didcot OX11 0DE, UK. 7 Department of
Materials Science and Engineering and the Russell Berrie Nanotechnology Institute, Technion Israel Institute of Technology, Haifa 32000, Israel.
8 Kavli Institute at Cornell for Nanoscale Science, 420 Physical Sciences Building, Ithaca, New York 14853, USA. *e-mail: y.y.kim@leeds.ac.uk;
she�ord.baker@cornell.edu; f.meldrum@leeds.ac.uk
Distribution coe�cient
5 0.008
[AA]inc (mol%)
0.004
4 0.000
0.08 0 10 20 30 40 50
3
2.0
Asp/calcite
2 1.5
Gly/calcite
1.0 0.04
1
0.5
0 0.0
0 5 10 15 20 25 0.00
0 100 200 300 400 0 100 200 300 400
Initial [AA]sol at Ca 10 mM (mM) Initial [AA]sol at Ca 10 mM (mM)
c d
2.0
0.16
1.8
1.6
Distribution coe�cient
1.4 0.12
[Asp]inc (mol%)
1.2
1.0 0.08
0.8
0.6
0.04
0.4
0.2
Asp/calcite Asp/calcite
0.0 0.00
0 20 40 60 80 100 0 20 40 60 80 100
Initial [Ca]sol at [Asp]sol 10 mM (mM) Initial [Ca]sol at [Asp]sol 10 mM (mM)
Figure 2 | Occlusion of aspartic acid and glycine in calcite. a, The amount of amino acid occluded within the CaCO3 crystals, [AA]inc , as a function of the
initial concentration of amino acid in solution, [AA]sol , for Asp and Gly at [Ca2+ ]sol = 10 mM. b, The distribution coe�cients for Asp and Gly in calcite as a
function of the initial [AA]sol at [Ca2+ ]= 10 mM. The insets show subsections of the respective graphs. c, The amount of Asp occluded as a function of
[Ca2+ ]sol at an initial [Asp]sol = 10 mM. d, The distribution coe�cient of Asp in calcite at an initial [Asp]sol = 10 mM. The error bars shown are defined as
s.d. for 3–5 measurements, and those for the Gly/calcite system were too small to show.
a b e ε
0.005
0.0006
Lattice distortion ∆a/a
Lattice distortion ∆c/c
Figure 3 | High-resolution PXRD analysis. a,b, Lattice distortions arising from the incorporation of Asp and Gly in calcite along the c-axis (a) and along the
a-axis (b). c–f, XRD peak broadening (FWHM) due to inhomogeneous strains induced by Asp (c) and Gly (d) incorporation. These graphs show that the
lattice distortions continue to increase with the amount of amino acid incorporated, while the inhomogeneous strains approach constant values. This
behaviour is consistent with a model in which the local strain fields (") associated with the individual Asp and Gly molecules (e) start to overlap as the
spacing between the individual molecules become smaller (f).
levelling off above 1 mol% occlusion, whereas for many of the lattice trend was particularly clear for the (006) planes. The consequence
planes [Gly]inc reached a maximum at approximately 2–4 mol%, of this behaviour is that all of the diffraction peaks from the
before decreasing again (Fig. 3c,d, and Supplementary Fig. 5). This crystals occluding 6.9 mol% Gly have comparable broadening levels
g(r)
4
0
c-axis c-axis
3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0
Ca−Ca distance (Å)
c d
8 12,000
8,000
O
2
6,000
0 X X X
4,000
−2
2,000 Asp−2
−4 Gly0
0
−6
−4 −2 0 2 4 6 8 −4 −2 0 2 4 6 8 −4 −2 0 2 4 6 8 0 1 2 3 4 5
∆ab/ab (×10−3) ∆ab/ab (×10−3) ∆ab/ab (×10−3) [AA]inc (mol%)
Figure 5 | Molecular dynamics simulations. a, Schematics showing Asp2 (left) and Gly0 (right) occluded in the calcite crystal. b, Radial distribution
function, g(r), of Ca–Ca distances in pure calcite and calcite with occluded Asp2 and Gly0 . c, Configurational energies of calcite with occluded Asp2 and
Gly0 , calculated on expansion of the calcite crystal along the c-axis and a-axis. The energy minima are found at 1c/c values of 0.002 and 0.003 for the
incorporation of 2.3 mol% Asp2 and 2.8 mol% Gly0 respectively. Red ‘x’ denotes the energy minimum of pure calcite, where the arrow shows the
distortion from the pure calcite sample to the experimental values found at ‘o’. d, The total occlusion energy for Asp2 and Gly0 .
strains in the calcite lattice in the neighbourhood of each AA Crystals with higher mol% of Asp occlusion were not tested owing
molecule. These drop off rapidly with distance from the molecule37 . to the small sizes and irregular morphologies of these samples
Although each occluded Gly is associated with larger strains than (Supplementary Fig. 1). These values are comparable to biogenic
Asp, Gly is also significantly smaller than Asp, such that each calcite (values of 3.4–4.2 GPa have been reported on the {001}
molecule affects a smaller volume of the calcite lattice. Our data face of the mollusc Atrina rigida)11 , and exceed the hardnesses of
suggest that these contrasting effects are of similar magnitude here, other synthetic calcite systems with occluded species reported so far,
giving rise to the similar occlusion energies. including calcite grown with occluded polymer micelles or polymer
nano-objects (Hmax ⇡ 3.0 GPa)9,10 , and with Mg substitutions for
Determination of the mechanical properties Ca (Hmax ⇡ 3.4 GPa)15 . In contrast, the indentation modulus was
Finally, the mechanical properties of the amino acid-containing insensitive to [AA]inc , as expected on the basis of a rule of mixtures
calcite crystals were measured using nanoindentation. Indentations model for a material containing a small volume fraction of a
approximately 200 nm deep and 1.4 µm wide were made using different phase.
a Berkovich (triangular pyramid) indenter, and the load–
displacement data from each indentation were analysed to Origin of hardening e�ect of occluded amino acids
determine hardness, H , and indentation modulus, EIT (ref. 38). Precipitation of CaCO3 in the presence of Asp and Gly provides a
Loads applied were sufficiently small (2.5 mN) such that cracking system in which we can precisely tune the amounts of these small
was suppressed and indentations were formed by plastic molecules within calcite single crystals across a broad composition
deformation (Fig. 6a inset). In performing these experiments, range. This allows us to study in detail the relationship between the
it is noted that twin formation can occur during nanoindentation quantities of molecules occluded, and both the lattice distortions
of single-crystal calcite provided that indentation is executed present, and hardnesses of the composite crystals. Because the
along orientations where twinning is geometrically favourable11,39 , calcite deformed plastically without cracking in these small-scale
or if internal interfaces that can accommodate the enormous experiments (Fig. 6a inset), the origin of the hardening effect can
strains associated with twinning are present12 or created by be determined by considering the slip systems in the material
fracture11,12 . None of our indentations were made in twin- (including both glide and twin systems) and the stresses needed to
compatible orientations, and no evidence of cracking or twin activate them with respect to the amount of amino acid occluded
formation in either the load–displacement data or the surface (see Supplementary Note 2 for full derivations). As hardness
topography was obtained. A more extensive discussion of twinning provides a measure of the resistance to motion of dislocations
is provided in Supplementary Note 5. on these slip planes, the increase in hardness with increasing
The hardness of the calcite crystals increased significantly with AA content demonstrates that the AA molecules provide effective
occluded amino acid content, starting at 2.5 GPa—equivalent to impediments to dislocation motion.
Iceland spar, a pure and highly perfect geologic form of calcite11 — The AA molecules can impede dislocations when either the stress
and reaching values of 4.1 ± 0.3 GPa for occlusion of 2.2 mol% fields around the molecules interact with the stress fields of the
Asp, and 4.1 ± 0.3 GPa for occlusion of 6.9 mol% Gly (Fig. 6a). dislocations (like solutes), and/or when dislocations are blocked by
a
4.5
b c
4.0
T
L′
Hardness (GPa)
3.5
3.0
T
θ
2.5
Asp/calcite F
L
Gly/calcite
2.0 2 µm
Pure calcite
0 1 2 3 4 5 6 7
[AA]inc (mol%)
d 100 e 5.0
Asp/calcite
Gly/calcite
4.5 y = 2.57x + 2.4
R2 = 0.92
10
4.0
Cutting force (nN)
Hardness (GPa)
3.5
1
3.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
Figure 6 | Mechanical properties. a, Hardness versus [AA]inc for calcite occluding Asp and Gly (inset: scanning force image of representative plastic
indentation in calcite). b,c, Schematics of dislocation bowing out between AA molecule pinning points (b) and force balance between dislocation line
tension T and the resisting force F provided by the molecule (c). The ‘cutting force’ Fc is the force needed for the dislocation to cut the molecule. Because
Fc ⌧ T, the e�ective spacing L0 is greater than the actual spacing L. d, Fc calculated using measured hardness and estimated molecular spacings as
compared with expected mechanical bond strengths. e, Hardness versus L 1 = (Cv,AA t)0.5 , where the linear behaviour supports blocking by molecules as
the strengthening mechanism. The error bars shown are defined as s.d. for 28 measurements.
the molecules themselves (like second-phase particles). However, If we assume that T > Fc , and that the hardness of the crystal
while the peak widths (inhomogeneous strains) level off or decrease is just the sum of its intrinsic resistance to dislocation motion—
above [AA]inc ⇡ 1 mol%, hardness continues to increase with [AA]inc that is, the hardness of pure calcite—and the additional hardening
occlusion (compare Fig. 3c,d with Fig. 6a, and see Supplementary due to AA molecules, we can write (see Supplementary Note 2 for
Note 3 for detailed explanations). Furthermore, the hardness is the derivation):
same for both amino acids at the same [AA]inc , despite the fact p
that the distortions are very different. These observations suggest H = H0 + 4.8(Fc /bL) Fc /2T (1)
that direct blocking by the molecules is the dominant hardening
mechanism. To test this idea, we consider a dislocation interacting where Ho is the hardness of pure calcite, b the Burgers vector
with a random array of AA molecules (Fig. 6b). The molecules magnitude, and L the spacing between molecules. Because we know
impede the dislocation’s motion, causing it to bow out40 . Because Ho , can estimate T and b, and have measured H as a function
lattice diffusion is not observed in calcite at room temperature, the of [AA]inc , if we can determine L from [AA]inc , then we can
dislocation can continue to move only by shearing the molecules. determine the actual cutting force for the molecules. To estimate
At each molecule, the dislocation line tension T is balanced by the L, we consider the AA molecules to be uniformly distributed
resisting force of the molecule (Fig. 6c). When a critical force, known throughout the crystal, which is consistent with the ssNMR data,
as the ‘cutting force’ Fc , is reached, the dislocation cuts the molecule and approximate the configuration of the molecules as a square
and moves on. Thus, if blocking by AA molecules dominates the array. This analysis gives L = (Cv,AA t) 0.5 , where Cv,AA is the number
increase in hardness, then Fc corresponds to the force required to of AA molecules/volume and t is the thickness of a thin slice, which
cut a molecule. is set equal to the diameter of the AA molecule (see Supplementary
35. Cho, K. R. et al. Direct observation of mineral-organic composite formation L.A.E., J.D.C., M.E.K. and S.P.B. were supported by the US National Science Foundation
reveals occlusion mechanism. Nature Commun. 7, 10187 (2016). (NSF) via a Materials World Network grant (DMR-1210304). B.P. acknowledges support
36. Bass, J. D. Mineral Physics & Crystallography: A Handbook of Physical Constants from the European Research Council under the European Union’s Seventh Framework
45–63 (American Geophysical Union, 2013). Program (FP/2007–2013)/ERC Grant Agreement no [336077]. B.D. and K.P. were
supported by the Leverhulme Trust and the EU reintegration grant
37. Eshelby, J. D. The determination of the elastic field of an ellipsoidal inclusion,
(PERG07-GA-2010-268429, FP7), project ‘mAARiTIME’. Ashely Coutu and Sheila Taylor
and related problems. Proc. R. Soc. Lond. A 241, 376–396 (1957). are thanked for technical help for RP-HPLC analysis. This work made use of the Cornell
38. Oliver, W. C. & Pharr, G. M. An improved technique for determining hardness Center for Materials Research Shared Facilities, which are supported through the NSF
and elastic-modulus using load and displacement sensing indentation MRSEC Program (DMR-1120296). The authors also thank Matthew Collins for his
experiments. J. Mater. Res. 7, 1564–1583 (1992). intellectual contribution to the development of the study and for helpful discussions
39. Clayton, J. D. & Knap, J. Phase field modeling of twinning in indentation of throughout.
transparent crystals. Modelling Simul. Mater. Sci. Eng. 19, 085005 (2011).
40. Foreman, A. J. E. & Makin, M. J. Dislocation movement through random arrays
of obstacles. Philos. Mag. 14, 911–924 (1966). Author contributions
41. Akbulatov, S., Tian, Y. C. & Boulatov, R. Force-reactivity property of a single Y.-Y.K. led the experimental work, preparing samples and carrying out EM and XRD, and
analysing the data; B.D. and K.P. performed HPLC measurements and analysed the
monomer is sufficient to predict the micromechanical behavior of its polymer.
occlusion data; J.D.C. and M.E.K. carried out nanoindentation measurements, J.D.C.,
J. Am. Chem. Soc. 134, 7620–7623 (2012). S.P.B. and L.A.E. analysed and modelled mechanical data; S.P.B. and L.A.E. co-supervised
42. Diesendruck, C. E. et al. Mechanically triggered heterolytic unzipping of a J.D.C. and M.E.K.; D.S. carried out molecular dynamic simulations work and C.L.F. and
low-ceiling-temperature polymer. Nature Chem. 6, 624–629 (2014). J.H.H. supervised D.S. and analysed simulation data; C.C.T. assisted with the PXRD
43. Grandbois, M., Beyer, M., Rief, M., Clausen-Schaumann, H. & Gaub, H. E. experiments; B.P. assisted with analysis and discussion of the PXRD data; D.G.R. and
How strong is a covalent bond? Science 283, 1727–1730 (1999). M.J.D. performed the NMR studies and analyses; F.C.M. originated and supervised the
44. Spruijt, E., van den Berg, S. A., Stuart, M. A. C. & van der Gucht, J. Direct project. All authors contributed to the preparation of the manuscript.
measurement of the strength of single ionic bonds between hydrated charges.
ACS Nano 6, 5297–5303 (2012).
Additional information
Supplementary information is available in the online version of the paper. Reprints and
Acknowledgements permissions information is available online at www.nature.com/reprints.
This work was supported by an Engineering and Physical Sciences Research Council Correspondence and requests for materials should be addressed to Y.-Y.K., S.P.B. or
(EPSRC) Leadership Fellowship (F.C.M. and Y.-Y.K., EP/H005374/1), by an EPSRC F.C.M.
Materials World Network grant (EP/J018589/1, F.C.M. and Y.-Y.K.) and an EPSRC
Programme Grant (grant EP/I001514/1) which funds the Materials Interface with
Biology (MIB) consortium (F.C.M., J.H.H. and D.S.). We acknowledge Diamond Light Competing financial interests
Source for time on beamline I11 under commissioning time and proposal EE10137. The authors declare no competing financial interests.
Solid-state nuclear magnetic resonance (ssNMR) spectroscopy analysis. All Data availability. The synchrotron powder XRD, molecular dynamic simulations,
ssNMR spectra were recorded on a Bruker Avance I NMR spectrometer with a nanoindentation and RP-HPLC data that support the findings of this study are
9.4 T superconducting magnet, operating at 400 MHz for 1 H and 100 MHz for 13 C. available in the Research Data Leeds Repository with the identifier
Samples were packed into 4-mm zirconia rotors, with Teflon tape being used to fill [http://doi.org/10.5518/46]51 .
out empty space within the rotor. The magic angle spinning (MAS) rate was
10 kHz. The standard cross polarization (CP) sequence in the Bruker pulse References
program library was used; 1 H 90 pulse length 2.5 µs, contact time 2.5 ms, with a 45. Demarchi, B. et al. Intra-crystalline protein diagenesis (IcPD) in Patella
ramped pulse on 1 H and spin lock field 70 kHz. During acquisition, spinal64 vulgata. Part I: isolation and testing of the closed system. Quat. Geochronol. 16,
decoupling at 100 kHz RF field strength was applied on 1 H. The Repetition time 144–157 (2013).
was 2 s between successive acquisitions to allow relaxation. 46. Kaufman, D. S. & Manley, W. F. A new procedure for determining DL amino
Proton driven spin diffusion experiments were conducted using the same initial acid ratios in fossils using reverse phase liquid chromatography. Quat. Sci. Rev.
cross polarization parameters as employed in the 13 C CP experiments. At 10 kHz 17, 987–1000 (1998).
MAS, the magnetization was allowed to evolve at single-quantum coherence during 47. Demarchi, B. et al. New experimental evidence for in-chain amino acid
the incremental delay (t1 ), and returned to zero quantum coherence by a 13 C 90 racemization of serine in a model peptide. Anal. Chem. 85,
pulse of 3.8 µs. 1 H decoupling was switched off during this mixing period of 100 ms 5835–5842 (2013).
48. Szeverenyi, N. M., Sullivan, M. J. & Maciel, G. E. Observation of spin exchange 50. Oliver, W. C. & Pharr, G. M. Measurement of hardness and elastic modulus by
by two-dimensional Fourier-transform C-13 cross polarization-magic-angle instrumented indentation: advances in understanding and refinements to
spinning. J. Magn. Reson. 47, 462–475 (1982). methodology. J. Mater. Res. 19, 3–20 (2004).
49. Thompson, S. P. et al. Beamline I11 at Diamond: a new instrument for high 51. Kim, Y.-Y. et al. Dataset for ‘Tuning hardness in calcite by incorporation of
resolution powder diffraction. Rev. Sci. Instrum. 80, 075107 (2009). amino acids’. Research Data Leeds Repository http://doi.org/10.5518/46 (2016).