Origin of the variability of the mechanical properties of silk fibers: 4. Order/crystallinity along silkworm and spider fibers

Page 1

Research Article

Received: 12 September 2011 Revised: 15 October 2011 Accepted: 28 October 2011 Published online in Wiley Online Library: 2 March 2012

(wileyonlinelibrary.com) DOI 10.1002/jrs.3122

1042

Origin of the variability of the mechanicalproperties of silk fibres: 3. Order andmacromolecule orientation in Bombyx moribave, hand-stretched strings and Nephilamadagascarensis spider fibresPhilippe Colomban,a* Hung Manh Dinh,a,c Aurélie Tourniéa andVincent Jauzeina,b

The comparison of the lowwavenumber of polarized Raman spectra (50–300cm–1) from Bombyx mori (fresh cocoons fibres, hand-stretched ‘Crins de Florence’ strings from the gland content, dried gland, regenerated silk films) and Nephila madagascarensis silks

reveals the high polarisation of fibre modes and the absence of polarisation for dried gland and regenerated silk films. This isconsistent with X-ray diffraction measurements. The orientation of the fibroin/spidroin chains is due to the stretching duringproduction, as for advanced synthetic fibres. The bandwidth of the ‘ordered chains’ signature is almost the same for the differentfibres. However, the degree of polarisation seems to be higher in the case of spider fibre. The huge bandwidth of lowwavenumbercomponents of regenerated films indicates high disorder. Measurements along the fibre point out conformation changes with aperiodicity (~20mm) related to the silkworm head motion during the fabrication of the cocoon. Copyright © 2012 John Wiley &Sons, Ltd.

Supporting information may be found in the online version of this article.

Keywords: fibres; silk; Bombyx mori; spider; mechanics

* Correspondence to: Philippe Colomban, Laboratoire de Dynamique, Interac-tions et Réactivité, UMR 7075 CNRS – Université Pierre et Marie Curie-Paris 6,4 Place Jussieu, c49, 75252 Paris Cedex 05, France.E-mail: [email protected]

a Laboratoire de Dynamique, Interactions et Réactivité, UMR 7075 CNRS –

Université Pierre et Marie Curie-Paris 6, 4 Place Jussieu, c49, 75252 Paris Cedex05, France

b Centre des Matériaux, Ecole des Mines de Paris, 91003 Evry Cedex, France

c Faculty of Physics, Hanoi National University of Education, 136 Xuan Thuy, CauGiay, Hanoi, Vietnam

Introduction

Silkworm silk has been a commodity for over 3000 years. Its use isexplained by an exceptional combination of high tensile strengthand thermal stability. Variability of mechanical properties of naturalfibres limits their use, even though the inevitable inflation of petro-leum-based synthetic products creates an incentive to again usenatural or biosourced products. Therefore, alongside the efforts in-volved in the control of production of natural and regenerated silk,a better understanding of the variability and its origins is required.Its biodegradability and biocompatibility offer new opportunitiesfor clinical applications (temporary immobilization, drug delivery,artificial tissues, etc.)[1–5] and motivate studies on regenerated silk.In previous papers[6–10] we have pointed out the interest of consid-ering the entire Raman signature and not only the 700–1800 cm–1

range as usually presented in the literature[11–18] to obtain a betterdescription of the material structure. We also discussed the rela-tionship between the general tensile behaviour and the silk historyand demonstrated the modifications resulting from water absorp-tion and drying.[9,10]

In this paper, we will focus our attention on the low wavenum-ber Raman spectra. The collective character of the low wavenum-ber modes makes them very dependent on the medium and longrange order and hence they are very efficient for comparing theorganisation of poorly crystallised materials. Although the lowwavenumber region is intensively studied to clarify structuremodifications of inorganic crystals and materials, this spectral

J. Raman Spectrosc. 2012, 43, 1042–1048

domain is poorly considered for polymeric materials. Becausethe micro/nanostructure of advanced synthetic fibres is the resultof a combination of stress and thermal treatments,[19–21] we willcompare natural silk fibres, produced by the creature as it sticksthe tip of the silk while stretching it with its head motion duringcocoon formation, with hand-stretched fibres from the fresh gutcontent. The influence of the fibre diameter will be addressed.

Material and methods

Materials

The studied materials have been extensively described in previ-ous papers.[6,7,9,10,22,23] A series of domestic Bombyx mori (Bm) silk

Copyright © 2012 John Wiley & Sons, Ltd.

Page 2

Origin of the variability of the mechanical properties of silk fibres

fibres were extracted from fresh cocoons produced at the UnitéNationale Séricicole (USN, La Mulatière, France) - Institut Nationalde la Recherche Agronomique (INRA). Nephila madagascarensis(Nm) spider dragline fibres extracted from a ~20-cm-long bundle(Fig. 1(a)) date from the 1990s when a silk farm pilot was createdin Madagascar under USN-INRA supervision. Note that the spiderfibre is free from any sericine coating although a sericine sheathcoats the silkworm bave, which is actually made of two adjacentfibres. The extracted bave/(degummed or not) fibres were firstcut into ~100-mm-long samples and then cut again in threeparts, then mounted and stuck in a paper frame as previously de-scribed.[6–8,13,24,25]

Silk coarse fibres, traditionally called ‘Crins de Florence’ or ‘silk-worm gut strings’ (Fig. 1(b)) have been prepared by hand-stretch-ing the gland content (Fig. 1(c)) of B. mori silkworm just dissectedfrom creatures ready to spin its silk cocoon.[22] These strings weretraditionally used for fishing or making surgical sutures.

The fibre diameters were measured by optical microscopy. Thespider fibre diameter ranged between 5 to 12mm (median value8mm)whereas the equivalent diameters of degummed B. mori fibreranged between 7 to 16mm (median value 14mm, Fig. 1 (d)). The di-ameter of hand-stretched Crins de Florence strings was very variable,from ~100 to 600mm, i.e. similar to the diameter of fibre precursors,the first step in the production of advanced synthetic fibres. Thediameter of dried glands ranges between 0.5 and 2.5mm (Fig. 1(e)).

Films (Fig. 1(f)) were prepared by slow evaporation of a fibroinsolution. The solution was obtained by dissolving cocoons in a LiBraqueous solution followed by dialysis and washing to removeadded ions. Film thickness ranged between 5 and 30mm.[22,26]

Methods

The mechanical tensile properties of the samples were obtainedusing a computer controlled DISCAPELEC Universal Fibre Testeras previously described.[9]

Figure 1. Photomicrographs of (a) Nephila madascarensis bundle, (b) Crins de(f) regenerated dried film pieces. The section of a silk yarn (d) shows the variab

J. Raman Spectrosc. 2012, 43, 1042–1048 Copyright © 2012 Joh

Raman spectra were recorded using a Sentera Bruker Optics in-strument (Ettlingen, Germany) using a 785 nm laser source. Theillumination powers ranged between 25 mW (spider goldcoloured fibre) and 100 mW (B. mori materials). A �40 OLYMPUSobjective, optimised for the infrared range (numerical aperture(NA): 0.65, long working distance (LWD): 17mm) was used to an-alyse the Nephila fibre and a �50 NIKON LWD objective (NA: 0.76,LWD: 17mm). The analysed volumes are respectively ~50�15for �40 and 30�10 mm3 for �50 objective, i.e. the eventual con-tribution of Bm sericine residues were negligible and the possibledifferences between the fibre core and skin mixed. Typicalrecording times of a multiwindow 30–3500 cm–1 spectrumranged with tenths of accumulations between 20 and 60min.

The band component analysis was performed usingORIGIN 5.0 soft-ware (OriginLab, USA). As previously described,[6–9,20,21,24,25,27–29]

the following assumptions weremade: a Lorentz shape was chosenfor the narrow and/or symmetric peak characteristic for orderedmaterial.[30] The Rayleigh wing is also described with a Lorentzianpeak. Broad components characteristic for amorphous material orvibrators with a large static or dynamic disorder were fitted usinga Gaussian curve.

Principal component analysis (PCA) has been processed usingMATLAB software (The MathWorks, Inc. Natick, USA). Spectra werestandardised by subtraction of mean value of variables (wavenum-bers) and divided by the standard deviation before calculation.

X-ray diffraction measurements were performed using theDebye–Scherrer (DS) Chamber film and the wide angle X-ray scat-tering techniques using the Co Ka radiation on silk yarn bundlesand dried glands. The DS films were then scanned to determineBragg ‘peaks’ and a ring intensity and position.[23]

Thermogravimetric analysis (TGA) was performed under hel-ium atmosphere using a SETSYS EVOLUTION thermobalance(SETARAM, Calluire, France). The sample were placed in a platinumcrucible and then heated from ~30 to 200 �C with a 20 �C/minheating rate.

Florence string, (c) dissected Bombyx mori gland, (e) dried central gland, andility of the B. mori fibre section; note the sericine matrix between the fibres.

n Wiley & Sons, Ltd. wileyonlinelibrary.com/journal/jrs

1043

Page 3

Ph. Colomban et al.

1044

Results and discussion

Fibre anisotropy

Figure 2 compares the Raman spectra recorded for two orientationsof the fibre, parallel and perpendicular, with respect to the directionof the laser electric field. The spectra were normalised with respectto the nC–Hmassif, peaking at ~2935 cm–1. The very noisy characterof the spectra above 3000 cm–1 resulted from the fast decrease ofthe charge-coupled device detector efficiency when using the785nm excitation. Consequently, the relative intensity of the bandsbelow and above 2000 cm–1 highly depended on the instrumentand laser excitation wavelength. However, with the 785nm laserexcitation, the Rayleigh wing was minimised allowing better obser-vation of the low wavenumber components. Previous studies ofthis region[6,7] with 514.5 and 532nm laser light failed to clearly dis-tinguish the lowwavenumber components from the Rayleigh wing.The very amorphous character of silk gave rise to a strong Rayleighwing that limited access to the very low wavenumber profile.The strong anisotropy of the fibres was obvious with the high

intensity of the ~85 (collective chain modes), 1080 cm–1 (nC–C),1450 cm–1 (dCH2), 1665 cm

–1 (nsC =O, Amide I) and 3285 cm–1

(nN–H) with parallel fibre orientation and of the 1229 (AmideIII), 1404 (nCOO) and 2875–2935–2990 cm–1 (nC–H ) triplet bandsfor the perpendicular one (see Refs[6,11–17] for the assignments).The most important polarised features are the huge ~85 cm–1

a)

0 500 1000 1500 2000 2500 3000 3500

νC-CνC-N νC-H

νC=O ν N-H

Crin de Florence

fibre

dried gland

Ram

an In

tens

ity

Wavenumber / cm-1

b)

50 100 150 200 250

85

90

95

100

8 %

5 %

0.5 - 2 %dried Films : from cocoon

from gland

Crin (700µm)6.5%

dried gland

TG

/ %

Temperature / °C

Figure 2. (a) Polarised Raman spectra of a dried gland (sericine sheathremoved), a degummed fibre and a hand-stretched Crins de Florence.Arrow: note the high intensity of the ~85 cm–1 peak, see Fig. 5; (b) TGAplots of dried films, Crins and gland; note the variability of film weight lossas a function of the preparation route.

wileyonlinelibrary.com/journal/jrs Copyright © 2012 John

band (parallel) characteristic of collective chain motions and the2935 cm–1 triplet (perpendicular) characteristic of (C–Hn)m groups.This is consistent with the opposite orientation of the side graftedamino acid groups and of the chain backbone, respectively mostlyoriented perpendicularly and parallel to the fibre axis.

On the contrary, the Raman signature of the dried gland and ofthe film surface did not show any difference as a function of theorientation of the gland/film versus the electric vector of the laserbeam. We did not succeed in making measurements on the filmedge so no conclusion could be drawn on film anisotropy.

TGA plots show two water losses, the first below 110 �Cassigned to surface water and the second at 170 �C assigned towater molecules bound to silk macromolecules. The highest lossis measured for 70 �C dried gland (~8%). Film water loss rangesbetween 5 and 0.5%, as a function of the preparation route(raw material, dialyse time, thickness). The loss measured on Crinsranges between 6 and 7%, as a function of the string diameter.

Measurements were also performed along the diameter(3–4mm) from the skin to the centre of pure colourless fibroingland (the sericine yellow sheath removed before the drying).The differences were small and concerned the variation of therelative intensity of the nC–H triplet versus the other bands thatindicated some radial anisotropy. Note the absence of the collec-tive chain peak at ~85 cm–1. A sericine trace can be characterisedby a very narrow and strong 2880 cm–1 peak.

The comparison of the X-ray diffraction DS patterns is given inFig. 3. The pattern of the dried gland shows a strong broad ring at~0.45nm and additional small intensity rings characteristic of anamorphous material. Some condensation of the ring intensity intovery broad Bragg spots is observed for the fibre pattern (Fig. 3(b)).The structure of silk (II form, density ~1.4g�cm–3) fibre is composedof different orthorhombic (P22121) ormonoclinic (P21) unit-cells witha (corresponding to interslab distance) = 0.895 to 0.938, b (interchaindistance) = 0.944 and c (along the fibre axis) = 0.698nm.[31–33] TheBragg bumps have been indexed accordingly. The very large exten-sion and the very small number of the Bragg spots arise from thevery poor crystallinity of the material: the DS coherence length isclose to 1.5 nm for most spots, the larger values being measuredfor 121/102 (~3nm) and 002 (~7nm) spots. These values are ordersof magnitude lower than in synthetic polyamide fibres[34] but verysimilar to those measured for many fibrous polymers, e.g. polyani-line[35–37] in which the coherence length strongly decreases whenthe macromolecular chain length increases. Electron microscopydetected the presence of small ‘crystallites’ with 2�6 or even10�66 nm dimensions.[38,39] larger sizes (16�23 or even 2�160nm) were claimed on the basis of atomic forcemicroscopymeasure-ments after specific preparation of the samples.[38,39] Obviously, such

Figure 3. X-ray pattern recorded for a B. mori (a) dried gland and (b) a fibre.

Wiley & Sons, Ltd. J. Raman Spectrosc. 2012, 43, 1042–1048

Page 4

40

Parallel Perpendicular

Origin of the variability of the mechanical properties of silk fibres

high coherence length values are not consistent with the experi-mental X-ray patterns nor with very broad silk Differential ScaningCalorimetry (DSC) endothermic melting (decomposition) peak,which is five times larger than that of synthetic polyamide.[7,8,38]

Figure 4 shows the Raman signature of the N. madagascarensissingle fibre. The strong polarisation of the ~85 cm–1 chain modeand of the nC–H triplet is obvious. The intensity of the latter trip-let is less strong than in B. mori silk and depends on the laser line.In particular the intensity of the ~3028 cm–1 band, characteristicof aromatic rings, is increased. Raman (pre)resonance effects arelikely because of the gold yellow colour of this fibre and a seriesof measurements with various exciting lines are required to gofurther in the understanding of the intensity variation.

Low wavenumber components

Figure 5 compares the decomposition of the low wavenumberregion. Six components are observed in the 50–250 cm–1 rangefor the different materials and their centres of gravity wavenumber

0 700 1400 2100 2800

para

perp

1383

1356

1236

1092

223

83

Ram

an In

tens

ity

Wavenumber / cm-1

//

828

641

568

303

86

850

903

1070

1330

1446

1550

1668

2032

2513

2935 30

28

Figure 4. Polarised Raman spectra of an N. madagascarensis fibre.

50 100 150 200 250 300 350

Rayleigh

para

Wavenumber / cm-1 Wavenum

Ram

an in

tens

ity

perp.88 cm-1

Bombyx moriRayleigh

50 100 150 20

Figure 5. Low-wavenumber polarised Raman spectra of a B. mori degummGaussian (G) band shapes are given.

J. Raman Spectrosc. 2012, 43, 1042–1048 Copyright © 2012 Joh

seem not to depend on the origin of the silk. Fittings have beenconducted by fixing the wavenumber positions; the intensitiesand bandwidths have only been fitted. Results are summarisedin Fig. 6. Polarization ratio are given in Table S1. Even if thedescription shows some subjective characteristics because ofthe assumptions made, the comparison of the fits points out thesimilarity of the Raman signatures, with very polarised bands atpositions rather close to those previously observed for syntheticpolyamide PA66 fibres. By comparison with synthetic polyamidefibres[20,21] the strongest Lorentzian component at ~75 cm–1 isassigned to a translational mode of the ‘crystalline’ chains. TheGaussian character of the ~140 cm–1 component indicates a linkwith ‘amorphous’ chains.

On the contrary, the Raman signatures of gland and regeneratedfilm (Figs 2 and 7) do not show marked differences in agreement

ber / cm-1 Wavenumber / cm-10 250 300 350

para

Crins de Florence

perp

50 100 150 200 250 300 350

para

Nephila

perp

ed, a Crins de Florence and a N. madagascarensis fibre. Lorentzian (L) and

72 88 110 140 230 2400

10

20

30

40Bombyx mori

Component wavenumber / cm-1

0

10

20

30

40

Crins de Florence

Rel

ativ

e A

rea

/ %

0

10

20

30Nephila

Figure 6. Comparison of relative area of different low-wavenumbercomponents for Fig. 5 polarised spectra.

n Wiley & Sons, Ltd. wileyonlinelibrary.com/journal/jrs

1045

Page 5

a) b)

100 200 300 400

Chain mode

Wavenumber / cm-1

Film BmdFibre Bmd

T' Chain modes

c)

Figure 7. Low-wavenumber polarised Raman spectra of a B. mori dried gland (a) and film (b, perpendicular polarisation); the narrow peak is the sig-nature of air bubbles in the film, note the broad component. The evolution of the n N–H/O–H domain is given in (c) as a function of the drying ofthe gland.

6 8 10 12 14 16 18 20 22 24 26

5

10

15

20

25

30

35

BmCd

Bmd

BmYBmC

BmOd

BmO

BmO*d

BmO*

Tussah

AAANeph

a) Bm OGMBm OGMBm* Anth Neph

You

ng's

mod

ulus

/ G

Pa

1200

Nephb) Bm

BmO

Diameter / μm

Ph. Colomban et al.

1046

with the very amorphous X-ray pattern of Fig. 3(a). Obviously, thestretching is at the origin of the polarised behaviour because ofthe alignment of the macromolecular fibroin chains along the fibreaxis. It can be expected that an optimisation of the stretching pro-cedure of the gland content, especially by reduction of the fibrediameter will increase the orientation of the silk chains along thefibre axis. Previous studies on synthetic fibres, e.g. on polyethyleneteraphthalate (PET) fibre[21] pointed out that hand stretchingincreases the polarisation degree but combined thermal and me-chanical treatments are required to increase both orientation andcrystallinity. The increase of the crystallinity reduces the bandwidthand hence the comparison of the bandwidth gives valuable infor-mation on the short-range structure.The full width at half maximum of the ~85 cm–1 Lorentzian peak

is ~30 cm–1, i.e. a bandwidth similar to thatmeasured for the crystal-line component of PA 66 and PET fibres. The values of the Gaussiancomponents are ~50 to 70 cm–1 as observed for the amorphouscomponents in the above synthetic fibres.[20,21] The short-range or-der in these fibres is thus very similar. The low relative intensity ofthe ‘narrow’ Lorentzian component versus the broad Gaussian onesin silk is consistent with a lower degree of crystallinity than that ob-served in synthetic homologous polyamide fibres.

0 2 4 6 8 10

0

10

20

30

40

50

60

70

80

Load

/ g

Strain / %

Figure 8. Load/strain behaviour of a Crins de Florence fibre (ultimatestress ~20MPa).

wileyonlinelibrary.com/journal/jrs Copyright © 2012 John

The comparison of the nN–H/O–H stretching domains (Fig. 7(c)) shows that the drying led, as expected, to a decrease in inten-sity of the nH2O band, but also a narrowing of the nN–H band ofN–H vibrators in ‘ordered’ chains. The narrowing of the N–H bandpoints out the high sensitivity of vibrational spectroscopy for

6 8 10 12 14 16 18 20 22 24 260

200

400

600

800

1000

OGMBmd

BmdAAA

Bmc

BmO*Anth

Neph

Ulti

mat

e st

ress

/ M

Pa

Diameter / μm

Figure 9. Dependence of mechanical parameters on silk diameter of dif-ferent creature fibres as a function of the fibre diameter: B. mori (differentbatches: Bm, BmO, BmO*), Antharea (Anth) and Nephila (Neph). BmdO isdegummed fibre of BmO, Bmd is very old degummed fibre of Bm, AAAis extracted fibre from textile yarn of Anth, BmC is extracted fibre fromcocoon of Bm.

Wiley & Sons, Ltd. J. Raman Spectrosc. 2012, 43, 1042–1048

Page 6

Origin of the variability of the mechanical properties of silk fibres

detecting small structural changes at the local scale. In conse-quence, the fresh gland specimen is more amorphous than thedried one. The small wavenumber shift from 3291 cm–1 (fresh)to 3280 cm–1 (dried) arises from the drying inducing shrinkageand an increase in density. By comparison, the correspondingwavenumber is 3286 cm–1 for the silk fibre.

Local structure and mechanical properties

The comparison of the vibrational signatures of B. mori andNephila fibres indicates that there are differences in the short-range structure between these silks and between natural andhand stretched fibres. The preparation procedures and the result-ing diameters were very different. It is well established that thestretching (�100, �500%, etc.) modifies the macromolecule ori-entation level and the crystallinity.[21,40] The alignment of thefibroin chains will depend on the side-chain grafts, i.e. the aminoacids. The important difference in the band intensity and degreeof polarisation between Bombyx and Nephila fibres may be re-lated to the difference in amino acid composition (high contentin glutamic acid, arginine and proline of spidroin). The observa-tion of two strong nN–H bands (~3255 and 3290 cm–1) in the caseof Nephila and a single one (3285 cm–1) for Bombyx is consistentwith differences between fibroin and spidroin backbone compo-sition and conformation. However, this has no influence on thestress/strain behaviour.[9]

Figure 8 shows a representative load/strain curve measured on a‘small’ diameter Crins de Florence fibre (~100mm). The elasticbehaviour limit (1.5%) is smaller than that observed with naturalsilk fibres (2%) revealing a reduced alignment of the chains alongthe fibre axis. Note that studies of fresh silk gland point out thepossible coil structure of silk with a hydrophobic part at the coil

Figure 10. (a) Longitudinal photomicrographs of a fibre extracted from a Bming (c) measured respectively for fibre segment Raman spectra in segment exacter. The mean spectrum is also given above.

J. Raman Spectrosc. 2012, 43, 1042–1048 Copyright © 2012 Joh

surface.[26,41] The coil-chain conformation transition may be theresult of the spinneret action and the stretching. Then, a quasi pla-teau is observed, very similar to that measured for natural fibres.[5]

The calculation of the ultimate stress gives values close to5–20MPa, i.e. much smaller values. Such a low value is related tothe big fibre diameter; Fig. 9 shows the evolution of the Young’smodulus and the ultimate tensile strength as a function of the fibrediameter. The variety of the materials studied at the laboratoryallows data for silk single fibres with diameter ranging from 8 to24mm to be obtained. We observe a regular decrease of the ulti-mate strength with the diameter increase (the ultimate strengthis a mechanical parameter depending on the artefact shape andsize, not on the material nature). The values measured for Crins(5 to 20MPa) are consistent with the extrapolation of the Fig. 8(b) data. The Young’s modulus seems to be less sensitive to thefibre diameter because of its intrinsic character; the Young’s mod-ulus is directly related to the strength of the chemical bond. How-ever, the decrease observed for big diameters is also consistentwith the low value measured for Crins, ~1GPa. Very similar valueswere measured for the tensile properties of regenerated films.[42,43]

Figure 10 shows microphotographs of a fibre segment of~40mm, i.e. the length required to make a complete cocoon loopwhen the creature makes its cocoon. The difference in opticalproperties is obvious. Raman spectra have been recorded in thedifferent locations, baseline subtracted, normalised and treatedwith MATLAB PCA software. The calculation made on these pre-liminary results underlines the difference in relation with the dif-ferent optical characteristics: the most affected bands are theAmide I (~1685 cm–1) and III (~1260 cm–1), according to someconformation changes along the fibre axis. More work is neededto go further in the understanding of the periodicity and natureof the variability along the fibre. This variability along the fibre

cocoon showing the different optical behaviour; PCA scores (b) and load-hibiting a high axial optical (1–4 zones) or a disordered (5–9 zones) char-

n Wiley & Sons, Ltd. wileyonlinelibrary.com/journal/jrs

1047

Page 7

Ph. Colomban et al.

1048

axis may also explain the low symmetry deduced from X-ray dif-fraction studies that require fibre bundles (monoclinic symmetryis claimed both for PA66 and silk).[32,44]

Conclusion

The decomposition of the low wavenumber Raman spectrumallows information on the chain alignment (fibre axial character)and degree of crystallinity to be obtained by comparing respec-tively the degree of polarisation and the bandwidth of the differ-ent components. The high number of components (at least sixrequired for the decomposition to be compared for the tworequired for PA66) indicates a low unit-cell symmetry and/orthe coexistence of different structures.The high polarisation of the ‘Crins de Florence’ spectrum and

the nonpolarisation of the gland spectrum prove that alignmentof the silk macromolecules is due to the stretching action. Thehighest degree of order is observed for the more dried materials.A quantitative comparison of the polarisation ratio and band-width will offer tools for comparing the silk orientation andcrystallinity and their relationship with mechanical properties.The complex motion of the head of the silkworm as it forms itscocoon, in a pattern of 8, involves modifications of the stressapplied to the fresh silk and variation of the molecule alignmentis expected.

Acknowledgements

Dr Bernard Mauchamp (INRA) is kindly acknowledged for the silksamples and for many discussions. Thanks to Dr Anthony Bunsellfor many advices regarding the mechanical testing. This workwas partly supported by ANR Blanc NANOSOIE and the project‘Training Scientific and Technical Cadres in Institutions Overseaswith the State Budget, Vietnam’.

Supporting information

Supporting information may be found in the online version ofthis article.

References[1] C. Vepari, D. L. Kaplan, Progr. Polym. Sci.-Polym. Biomedical Appl.

2007, 32, 991.[2] D. A. Tirrell, M. J. Fournier, T. L. Mason, Curr. Opin. Struct. Biol. 1991,

1, 638.[3] H. Heslot, Biochimie 1998, 80, 19.[4] Y. Liu, X. Chen, J. Quian, H. Liu, Z. Shao, J. Deng, T. Yu, Appl. Biochem.

Biotechnol. 1997, 62, 105.[5] V. Jauzein, Ph. Colomban, in A. R. Bunsell, P. Schwartz (Eds),

Handbook of Tensile Properties of Textile and Technical Fibres, CRCWoodhead Publishing Ltd, Oxford, 2009, pp 144–178.

[6] Ph. Colomban, H. M. Dinh, J. Riand, L. C. Prinsloo, B. Mauchamp,J. Raman Spectrosc. 2008, 39, 1749.

[7] H. M. Dinh, C. Paris, Ph. Colomban, B. Mauchamp, Comptes-RendusJNC 16, 16èmes journées nationales sur les composites, Toulouse, (Eds:

wileyonlinelibrary.com/journal/jrs Copyright © 2012 John

Ph. Olivier, J. Lamon), AMAC, Paris, 2009. http://hal.archives-ouvertes.fr/docs/00/39/75/75/PDF/244b.pdf

[8] H. M. Dinh, H. El Baghli, Ph. Colomban, Proc. APCTP-ASEAN Work-shop on Advanced Materials Science and Nanotechnology-AMSN2008: Nha Trang (Vietnam) September 15th-21th, 2008.

[9] Ph. Colomban, H. M. Dinh, A. R. Bunsell, B. Mauchamp, J. RamanSpectrosc. 2012, DOI: 10.1002/jrs.3044

[10] Ph. Colomban, H. M. Dinh, J. Raman Spectrosc. 2012, DOI: 10.1002/jrs.3123

[11] M. E. Rousseau, T. Lefèvre, L. Beaulieu, T. Asakura, M. Pézolet, Bioma-cromolecules 2004, 5, 2247.

[12] M. E. Rousseau, L. Beaulieu, T. Lefèvre, J. Parais, T. Asakura, M. Pézolet,Biomacromolecules 2006, 7, 2012.

[13] P. Monti, P. Taddei, G. Freddi, T. Asakura, M. Tsukada, J. Raman Spec-trosc. 2001, 32, 103.

[14] J. Magoshi, Y. Magoshi, S. Nakamura, Polymer Commun. 1985,26, 309.

[15] S. Zheng, G. Li, W. Yao, T. Yu, Appl. Spectrosc. 1989, 43, 1269.[16] M. Preghenella, G. Pezzotti, G. Migliaresi, J. Raman Spectrosc. 2007,

38, 522.[17] T. Lefèvre, M. E. Rousseau, M. Pézolet, Biophys. J. 2007, 92, 2885.[18] J. Sirichaisit, R.J. Young, F. Vollrath, Polymer 2000, 41, 1223.[19] A. R. Bunsell, P. Schwartz (Eds), Handbook of Tensile Properties of

Textile and Technical Fibres, CRC Woodhead Publishing Ltd, Oxford,2009.

[20] A. Marcellan, Ph. Colomban, A. Bunsell, J. Raman Spectrosc. 2004,35, 308.

[21] Ph. Colomban, J. M. Herrera Ramirez, R. Paquin, A. Marcellan, A. Bunsell,Eng. Fract. Mech. 2006, 73, 2463.

[22] H. M. Dinh, Raman/IR study of the variability of proteic fibres: rela-tionship between local structure, treatments and (nano)mechanicalproperties of silk fibres and films, Université-Pierre-et-Marie-Curie(UPMC Paris 6) PhD thesis: Paris, 2010. http://www.ladir.cnrs.fr/pages/colomban/Manh-Thesis.pdf

[23] V. Jauzein, Etude de la microstructure et du comportement mécani-que de la fibre de soie, Mines ParisTech PhD Thesis, Paris, 2010.

[24] R. Paquin, Ph. Colomban, J. Raman Spectrosc. 2007, 38, 504.[25] J.M. Herrera Ramirez, A.R. Bunsell, Ph. Colomban, J. Mater. Sci. 2006,

41, 7261.[26] A. Percot, Ph. Colomban, C. Magnier, 2009, unpublished results.[27] J. M. Herrera Ramirez, Ph. Colomban, A. R. Bunsell, J. Raman Spectrosc.

2004, 35, 1063.[28] Ph. Colomban, G. Gouadec, Comp. Sci. Technol. 2009, 69, 10.[29] Ph. Colomban, Comp. Sci. Technol. 2009, 69, 1437.[30] G. Gouadec, Ph. Colomban, Progr. Cryst. Growth Charact. Mater.

2007, 53, 1.[31] R. Marsh, R. Corey, L. Pauling, Biochim. Biophys. Acta 1955, 16, 1.[32] Y. Sheng, A. Johnson, D. Martin, Macromolecules 1998, 31, 8857.[33] J. Warwicker, Acta Crystallogr. 1954, 7, 565.[34] A. Marcellan, Microstructure, Micromécanismes et comportement à

rupture de fibre PA66, PhD thesis., Ecole des Mines de Paris, 2003 ;http://bib.rilk.com/634/01/Marcellan.pdf

[35] J. P. Pouget, M. E. Jozefowicz, A. J. Epstein, X. Tang, A. G. MacDiarmid,Macromolecules 1991, 24, 7779.

[36] A. El Khalki, Ph. Colomban, B. Hennion,Macromolecules 2002, 35, 5203.[37] L. Mazerolles, S. Folch, Ph. Colomban,Macromolecules 1999, 32, 8504.[38] M. Dobb, R. Fraser, T. Macrae, J. Cell Biology 1967, 32, 289.[39] K. Numada, P. Cebe, D. Kaplan, Biomaterials 2010, 31, 2926.[40] K. Perekelpin, Fibre Chemistry 2007, 39, 308.[41] H. M. Dinh, A. Percot, Ph. Colomban, 2012, to be submitted.[42] Ph. Colomban, S.E. Moukit, H.M. Dinh,M. Hassine, J. Riand, B. Mauchamp,

Rev. Comp. Mater. Av. 2008, 18, 163.[43] M. Tiennot, unpublished UPMC Master Report, Paris, 2011.[44] A. Marcellan, A. R. Bunsell, R. Piques, Ph. Colomban, J. Mater. Sci.

2003, 38, 2117.

Wiley & Sons, Ltd. J. Raman Spectrosc. 2012, 43, 1042–1048