Acta Biomaterialia 6 (2010) 2181–2188

Contents lists available at ScienceDirect

Acta Biomaterialia

journal homepage: www.elsevier .com/locate /actabiomat

Influence of moisture on the mechanical behavior of a natural composite

M. Johnson a, S.L. Walter b, B.D. Flinn c, G. Mayer c,*

a The Boeing Company-Mail Code 8M93, 20403 68th Avenue South, Kent, WA 98032-2316, USAb The Boeing Company, Seattle campus, P.O. Box 34787, Seattle, WA 98124-1787, USAc Department of Materials Science & Engineering, University of Washington, Seattle, WA 98195-2120, USA

a r t i c l e i n f o

Article history:Received 16 March 2009Received in revised form 8 November 2009Accepted 3 December 2009Available online 11 December 2009

Keywords:MoistureSpongesDampingEnergy dissipation

1742-7061/$ - see front matter � 2009 Acta Materialdoi:10.1016/j.actbio.2009.12.006

* Corresponding author. Tel.: +1 206 616 2832.E-mail address: gmayer@u.washington.edu (G. Ma

a b s t r a c t

The effects of moisture on the mechanical properties of the spicules of the sponge Euplectella aspergillumhave been investigated. Determinations were made with the aid of a dynamic mechanical analyzer inboth the static and dynamic modes, as well as imaging of the failed surfaces with scanning electronmicroscopy. For comparison purposes, melt-grown glass fibers of similar diameters were also studiedin both distilled water and seawater. That exposure reduced both the stiffness and strength of the spic-ules. In addition, the energy required to achieve complete failure decreased in moist environments. Thedata for the wet spicules in both aqueous media showed decreasing values of energy dissipated until cat-astrophic failure compared to dry samples. The strength of wet glass decreased when compared with thedry condition, and the elastic modulus was also reduced. The most marked influence of moisture wasseen in the damping effects in moist spicule samples that were nearly an order of magnitude larger thanthe damping of dry spicules. This effect was attributed mainly to plasticization of the thin organic layers.

� 2009 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.

1. Introduction

Moisture is known to have a substantial influence on themechanical behavior of natural materials as well as of syntheticmaterials, including glass [1]. The role of moisture in the viscoelas-tictity of biomaterials was reviewed broadly by Dorrington [2].Work has been done on the effects of moisture on a variety of nat-ural materials, ranging from spider silk [3] to the effects of mois-ture on insect cuticle and pine-cone bract by Vincent [4]. Themechanical properties of other rigid biological composite materialshave also been shown to have substantially different values in wetthan in dry conditions. For the case of mollusk shells [5], resultsgenerally showed lower strength and stiffness, but higher tough-ness. Such effects have been related to plasticization of the organicphases of the composite systems that were investigated.

Since siliceous sponges exist in seawater as well as in freshwa-ter environments, they are always immersed in water, often coldand deep. Because of this fact, and in view of the observed effectsof moisture on other rigid biological composites, it was consideredessential to study such effects on the silica/organic composites thatcomprise sponge spicules and skeletons.

The spicules of the sea sponge Euplectella aspergillum consist of alayered structure of alternating concentric cylinders of hydrated sil-ica, SiO2�nH2O, and thin organic constituents, probably proteins, inaddition to an inner cylindrical hydrated silica core (a schematic

ia Inc. Published by Elsevier Ltd. A

yer).

diagram is shown in Fig. 1). Due to this unique structure, the spic-ules of this sponge have been shown to possess attractive combina-tions of mechanical properties, including strength, stiffness,resilience and energy dissipation, that exceed those of similarman-made materials, such as monolithic silica [6,7]. The spiculesof a number of hexactinellid sponges have structures with interest-ing and potentially important optical and mechanical properties[8–10], although in deep ocean environments, where many of thesesponges live, the ability to harvest energy from sunlight may bedoubtful. However, nature has found ways to utilize the strengthand stiffness of glass, and the flexibility and resilience of the thin or-ganic interlayers of the spicules. These characteristics confer on thestructure of the composite fiber an ability to dissipate largeamounts of energy on both the static and dynamic levels [11,12].

The sponge E. aspergillum is usually found at ocean depths from36 to 5050 m [13]. The skeletons of siliceous sponges are formedthrough biomineralization at cold temperatures (rather than atthe high temperatures, on the order of 1500 �C, that are normally re-quired to melt and form synthetic glass fibers). As the spicule fibersform the lattice-like skeleton, they either are fused together or areconnected by a glue, to form the complex structure seen in Fig. 2.

Spicule fibers of mature sponges of E. aspergillum generallyrange from �30 to �70 lm in diameter and contain a central cylin-drical structure that typically represents �1/3 of the total diameterof the fiber, as shown in Fig. 1. In the center of this hydrated silicacore is an axial filament of square cross-section that is about 1 lmon each side (probably proteinaceous, based on work on anotherhexactinellid sponge by Müller et al. [14]. Moving radially outward

ll rights reserved.

Fig. 3. Skeleton of Euplectella sponge, indicating position of smooth fibers used inmechanical tests. ([9]; copyright 2004, National Academy of Sciences, USA) usedwith permission.

Fig. 1. Cross-section of E. aspergillum spicule fiber: the cylindrical fibers are �30–70lm in diameter; rings generally increase in thickness (0.1–1.0 lm progressing fromouter to inner rings; solid silica cylinder is �1/3 of the diameter; axial cylinder with�1 lm square core of protein constituents; number of rings varies (typically 15–30); thin protein rings between hydrated silica glass cylinders of the order of 5–10 nm.

Fig. 2. Scale bar: 1 cm. E. aspergillum skeleton; length usually ranges from 6 to32 cm and diameters from 1.5 to 5 cm.

2182 M. Johnson et al. / Acta Biomaterialia 6 (2010) 2181–2188

from the central core is a series of concentric layers alternating be-tween a hydrated silica, SiO2�nH2O, and a very thin (of the order of5–10 nm) protein phase. The protein phase makes up only about2–3 vol.% of the spicule.

Organic phases have been found to play important roles inaltering and controlling the mechanical behavior of moist biologi-cal composites, as noted earlier. Generally, the effects have been tolower the strength and stiffness, but with an observable increase intoughness. The major influence is believed to stem from plasticiza-tion of the organic constituent(s) of the composites [4].

Previous experiments have shown that, in bending as well as intension, the capacity to dissipate mechanical energy in the spongespicules was considerably higher than that shown by monolithicsilicate glass [11]. Energy dissipation, in this case, is defined asthe amount of energy that is expended up to the starting point ofcatastrophic failure. Walter et al. [12] found that a number of fac-tors contributed to such increases in these fibers in the dry state.The primary focus of this phase of study was to define the extentthat moisture affected the mechanical behavior of the complexcomposite that is the spicule, and perhaps, by extension, similarnatural composites.

2. Materials and methods

2.1. General background

This study focussed on the mechanical response of both dry andwet spicule fibers of E. aspergillum in comparison with a baseline of

fibers of melt-drawn glass fibers in dry and wet states under bothstatic and dynamic conditions. The skeletons of the sponges fromwhich the spicules were obtained had been donated by Natures’Creations of Sammamish, WA, and originated from a private col-lecting source in the Philippine Islands. There was no prior knowl-edge of handling, other than that the skeletons had been cleaned ina diluted bleach solution. The outer soft biological tissues had beenremoved from the skeletons. The spicule fibers were (originally ob-served as smooth) straight sections taken from near the base of theskeleton shown in Fig. 3. They ranged in diameter from about 30 to70 lm. Although a baseline of opal (hydrated silica) would havebeen preferred for comparison, the scatter in the structures andproperties of opal, combined with the fact that opal does not existas a fiber, prevented such a choice. Thus, the glass fibers used asthe base of comparison had been pulled from the melt through diesby NIST from an EMGO 360 lead-free electronic glass (comprisingprimarily SiO2, but also small levels of Na2O and BaO). The diame-ters of those melt-processed fibers ranged from 50 to 90 lm.

Each of the spicule fibers for analysis was taken from the samesponge in an attempt to avoid variations between different speci-mens of the same species. For example, it had been determinedearlier [15] that the numbers of concentric rings in a spicule fibercould vary from about 15 to 30 or more in different specimens of E.aspergillum that were examined, and that those differences did notcorrelate with external fiber diameter.

2.2. Fiber selection and exposure media

The moisture media to which the spicule fibers and the melt-drawn glass fibers were exposed were seawater and distilledwater. The saturation point was to be defined by soaking the sam-

Fig. 4. DMA loading system for fiber testing in bending, showing channel (inaluminum base) and triangular anvil [7].

150

200

250

300

350

400

inee

ring

Stre

ss (M

Pa)

Distilled Sea Dry

M. Johnson et al. / Acta Biomaterialia 6 (2010) 2181–2188 2183

ples incrementally over time in either seawater or distilled water.After 1, 24 and 48 h, the spicules were loaded to failure in bending.The point of saturation was taken to occur at the point where nosignificant change in elastic modulus occurred in the samples.

Specimens for mechanical testing were carefully extracted fromthe sponge skeleton with tweezers, selecting only smooth-sidedsegments (Fig. 3c). Samples no smaller in diameter than 40 lmwere selected during optical microscopic examination to ensurethat they would fracture completely, rather than simply bend, dur-ing testing. This led to S/d (span to depth) ratios between 20 and25. Fibers for mechanical testing were confirmed to be free fromspines or branches (see Fig. 3b). During handling of both the spic-ules and the melt-processed glass fibers, extreme care was taken toavoid damage, particularly at the point of application of the dy-namic mechanical analysis (DMA) load.

2.3. Sample preparations

Samples of both spicules and glass fibers were soaked in twosolutions – one of specially filtered (to remove any microbial spe-cies) seawater that was collected off the coast of Hawaii, and theother of distilled water – for a series of time intervals. The stiffnessof the fibers was then measured with the DMA apparatus in thestatic mode and, when stabilized, taken as the indicator ofsaturation.

2.4. DMA procedures

The DMA apparatus was a Perkin-Elmer Model 7e, and was setin either the static or the dynamic mode for three-point bend test-ing. The bending fixture was fabricated from a short cylindricalelectron microscope sample holder that was machined across thediameter to contain a shallow square central channel of widthand depth 1 mm. A specially fabricated anvil was prepared andadapted to apply the load, as shown in Fig. 4. The samples wereplaced perpendicular to the stage channel and the electromagneticloading system of the DMA apparatus was used to apply the load.Careful visual observation of the fiber alignment was done to en-sure proper orientation of samples for bend testing. These weretested at a loading rate of 20 mN min–1. Failed samples were col-lected and stored for subsequent examination of the surfaces.

Dynamic tests on both the sets of dry and wet spicules and dryand wet glass fibers were performed over a frequency range of1–51 Hz, and a peak load of 10 mN. Stress and strain in the DMAtests were calculated from the three-point beam bending solutionfor a point at x = L/2 and y = �d/2 (where L is the length of the fiberbetween supports, and y the distance from the center to the outersurface: see the Appendix).

2.5. Scanning electron microscopy (SEM)

Failed specimens of both the wet and dry specimens for boththe spicule fibers and the melt-drawn glass fibers were carefullycollected after DMA testing, and prepared for examination in aJEOL JSM-7000F field emission scanning electron microscope usingstandard preparation procedures. Additional SEM examinations offiber failures (performed in the dry state) in tension were also donefor comparison.

0

50

100

0 0.01 0.02 0.03 0.04 0.05 0.06

Engineering Strain (mm/mm)

Eng

Fig. 5. Representative engineering stress vs. strain curves for spicule fibers [7].

3. Results and discussion

Analysis of the samples was performed using the results frombending tests taken from the DMA records. These yielded engineer-ing stress and strain values. Stiffness (from elastic moduli) wasused to determine the saturation period after immersion in seawa-

ter and distilled water environments. It was difficult to make quan-titative comparisons of the moduli from the stress–strain behaviorof the spicules vs. glass fibers in the wet and dry conditions. Asseen in Fig. 5 [7], there are erratic changes in that behavior, espe-cially at small loads in the spicules, while the glass fibers do notshow such behavior. Furthermore, there was much scatter in thedata for both dry and wet conditions, especially in the spicule sam-ples. This was judged to be due to the presence of a large variety ofsurface defects, even in what were originally thought to be smoothsections of the spicule fibers (Fig. 3c), and those defects can be seenin the laser scanning confocal microscopy image (Fig. 7), which af-fords better resolution. It is at low deflections that elastic modulishould be determined, and so a quantitative comparison was notpossible with the behavior indicated. However, the trends for

Fig. 7. Laser scanning confocal microscope image of an untested spicule fiber,showing surface defects.

0

50

10

15

20

25

30

35

40

45

0.00 0.01 0.02 0.03 0.04 0.05 0.06

Strain

True

Stre

ss (M

Pa)

Dry

Sea

Distilled

Fig. 6. True (or engineering) stress–strain behavior of dry and wet melt-grownglass fibers [7].

Fig. 8. Telescoping model of the response of a spicule fiber tested in tension [6],used with permission of MRS).

2184 M. Johnson et al. / Acta Biomaterialia 6 (2010) 2181–2188

elastic modulus appear clear from Table 1. The two things that areevident from the curves (Figs. 5 and 6) and data are a significantdrop in stiffness after 1 h exposures in both aqueous environments,and the decreasing slopes of the curves for both sets of data (glassand spicule) going from dry to distilled water and seawater envi-ronments. For the glass fibers, the explanation for such softeningis indicated by the reactions of moisture with silica and the variousadditives in the EM360 fibers, as described in the article by Doveand Rimstidt [16], and by Doremus [17]. For the spicules, the mois-ture effect is more complicated. It is expected that plasticization ofthe thin protein layers occurs, but this is not necessarily a degrada-tion effect, since sponges and their spicules are known to exist forsome years in marine environments. It is expected that reactionkinetics would also be much slower at deep-ocean temperatures,where the sponge normally lives, than at the surface, or in a labo-

Table 1Mechanical properties of spicules in bending to determine moisture effects in both seawa

Soak time (h) Dry Seawater

0 1 24

Elastic modulus (GPa) 13.7 ± 2.2 9.3 ± 2.7 9.1 ± 1.0Maximum strain 0.044 ± 0.005 0.041 ± 0.008 0.044 ± 0Engineering fracture stress (MPa) 431 ± 87 320 ± 72 303 ± 41No. of samples 4 9 4

ratory environment. In addition, because of the nature of the for-mation of biomineralized silica (that was shown in E. aspergillumby the aggregation of colloidal nanoparticles of a size of the orderof 2.8 nm, and clustered in groupings of �50–200 nm [18]), poresshould exist between the aggregated clusters, so moisture effectson the silica phase itself within the concentric ring structure ofthe spicules may also affect the stiffness of the spicules.

The point at which the elastic modulus appeared to stabilizewas taken to be the point of saturation, and this occurred after1 h of exposure in both moist environments. At least four speci-mens were used for each test condition and time.

Additionally, two different ways were employed to analyze thefailed surfaces of the samples from the SEM images, and those pro-vided the means to estimate the final cross-sectional areas of thesamples. In tension, the model proposed forth by Mayer et al. [6]was used (Fig. 8). This model indicated that a telescoping mecha-nism for the concentric cylinders operates in tension, with the out-er cylinders failing either individually or in packets (why theseoccur in different ways is not known at present), the load–exten-sion curve increasing and decreasing in a serrated fashion, anddropping suddenly when the last fast fracture occurred in the spic-ules (not shown here). In tension, the final cross-sectional area wasregarded as the flat central region of the sample where the loadwas judged to drop precipitously. In bending, where most of thedata in this work were taken, the stress state in the outer fibersthat are in tension is thought to bring about the first failure ofthe outermost cylinder, or packet of cylinders, on the tensile sideof the bent spicule. In order to clarify the complex failure mecha-nisms during deformation in bending in the DMA apparatus, thefollowing sequence is proposed based upon fractographic evidence,based on a cylindrical rod of the cross-section shown in Fig. 1:

(1) cracks initiate at flaws at the outermost layer on the tensilesurface;

(2) cracks then propagate radially and circumferentially in thedecreasing tensile stress field, and are blunted and divertedby thin organic layers;

ter and distilled water.

Distilled water

48 1 24 48

8.9 ± 1.9 8.4 ± 1.2 9.2 ± 1.1 8.3 ± 0.6.005 0.044 ± 0.005 0.046 ± 0.003 0.039 ± 0.005 0.040 ± 0.005

300 ± 33 328 ± 50 284 ± 37 294 ± 846 9 3 4

(Initial condition) Failure of Spicule Failure of Spicule(tension specimen) (bending specimen)

a b c

Fig. 9. Schematic depiction of areas of spicules, untested (a), failed in tension (b),and failed in bending (c).

Fig. 10. Representative SEM image of failed surfaces of a dry spicule fiber tested intension.

Fig. 11. Representative SEM image of the failed surface of a dry spicule sample (inbending).

M. Johnson et al. / Acta Biomaterialia 6 (2010) 2181–2188 2185

(3) radial crack growth is arrested or diverted at an organiclayer, and viscoelastic deformation of that layer may retardcrack growth;

(4) circumferential crack growth continues around the spiculeuntil KI (fracture toughness in mode I) at the crack tip < KIC

(local critical fracture toughness in mode I) of the inorganiclayer. Some axial delamination may start due to rotations inbending;

(5) new crack initiation occurs in the second most outer layer(or packet) at a flaw along the tensile surface;

(6) circumferential cracks start to turn in the axial direction dueto shear forces and eventually propagate in the axial direc-tion accompanied by delamination;

(7) repeat steps 2–6 until a critical crack size is reached and fastfracture occurs through the inner core section.

However, the cylinders probably separate and break away un-der increased bending load until somewhere above the mid-pointof the bent fiber. Thus, when the load drops precipitously, for sim-plicity, we postulate that the final cross-section at the load drop isthe central core, plus half of the area of the spicule fiber (Fig. 9).This generalization provided the necessary information to calcu-late a true stress using a modification of the moment of inertia inthe beam-bending equation (indicated in the Appendix) at thepoint of final failure in the spicules. This true stress takes into ac-count the decreased cross-sectional area that carries the load dur-ing the course of bending.

Table 1 shows that the elastic modulus of the spicules de-creased in the presence of moisture. In addition, after a 1 h soak,the modulus and maximum strain remained fairly constant,regardless of additional moisture exposure. A 1 h soak was thusinterpreted to be saturation, and was used as a guide in subsequenttesting of the spicules. Since the melt-drawn glass fibers, unlike thehydrated silica spicules, were not porous, they were tested at theearlier established saturation time of 1 h.

In the presence of moisture, melt-grown glass exhibited a smalldecrease in maximum strain when compared to the dry condition.Fig. 6 displays the extent to which moisture affects the wet and dryglass samples (engineering stress and strain are expected to beidentical to true stress and strain for glass). The mechanism under-lying the decrease of the mechanical properties is associated withmoisture facilitating breakage of Si–O–Si bonds [1] and also lower-ing the fracture strength by lowering the surface energy accordingto the Griffith equation, which applies to brittle materials such asmonolithic glass:

rf ¼ ð2cE=pcÞ1=2 ð1Þ

where rf is fracture strength, c is surface energy, E is elastic modu-lus and c is crack length. In such materials, this also results in thereduction of the energy required for fracture. Interestingly, in the

case of the spicule fibers with the concentric ring structure, the out-er layers are very thin (of the order of 0.1 lm, out of a radius of per-haps 25 lm). The flaws and ensuing cracks that exist in this outerlayer may propagate under stress, but are expected to be bluntedat the first thin organic layer. The next inner siliceous layer is with-out damage, so the fracture strength of the remaining part of thespicule fiber is quite high.

3.1. Interpretation of failure modes from SEM observations

It was apparent that the failures of the spicule fibers that wereobserved in tension (Fig. 10) were very different from the failuresthat were observed in bending. The failures of spicules in tensiletesting were described by Mayer et al. [6]. The data that form thebasis of this study were taken in bending, and the appearance ofthe specimens that were exposed to moist conditions did not ap-pear substantially different from those in dry conditions. The bend-ing test results were from the DMA, operating in either the static ordynamic mode.

Fig. 11 is a SEM image of a typical failed surface of a spiculespecimen tested in a wet condition (the surfaces that failed underthe dry condition or in seawater were similar in appearance tothose tested in distilled water). Estimations of the cross-sectionalareas just before final failure were from the model shown in

0

5

10

15

20

25

0 10 20 30 40 50 60

Frequency (Hz)

Tan

δ

Moist Spicule

Dry Spicule

Fig. 13. Representative tan d vs. frequency for wet and dry spicules from DMA tests.

2186 M. Johnson et al. / Acta Biomaterialia 6 (2010) 2181–2188

Fig. 9, and subsequent calculations were made using the assump-tion that the spicule’s final diameter was as indicated schemati-cally. (Calculations that were made for energy dissipation in thespicules are shown in the Appendix) The reason for this approxi-mation was because, although many of the broken samples wereexamined, some were lost in the DMA apparatus at the point of dy-namic failure during final fracture, and our best estimate of thearea at that juncture was based on examination of the retrievedfailed sample sections.

The extent of the environmental effect on both the spicule andglass specimens is shown in the energy to fracture values in the bargraphs of Fig. 12. It is important to note that there are large errorbars in the energy dissipation data for the glass fibers under bothwet and dry conditions, and that this scatter is much larger thanthe energy dissipation data for the spicules. Two conclusions canbe drawn from these findings. The first is that the surfaces of theglass fibers were not especially protected after processing, andwere susceptible to damage in handling, despite careful proce-dures. The second is that the spicule fibers show less scatter in en-ergy dissipation because of their sacrificial, protective outer layers.The specimens that were exposed in distilled water showedslightly lower energy dissipation than those exposed to seawater.At present, no explanation is evident for this behavior.

A similar trend was exhibited in the mechanical behavior of thesponge spicule samples in the presence of moisture. Table 1 showsvalues of the maximum stresses in the spicules under dry and wetconditions (measured in the DMA record as engineering stress).Those maximum stresses were much lower in the samples thatwere exposed to either fluid. The decrease in maximum stresswithout a significant increase in the extension to failure bringsabout the result that the soaked spicules showed a lower energyto failure than the dry samples, as shown in Fig. 12. This decreasewas attributed to the plasticization of the protein interlayers in thespicules by moisture, and, in this instance, the deformation of theseries of many thin plasticized layers did not result in increaseddeflection in bending.

With regard to the scatter in the data of these samples, it wasobserved that there was less scatter in the spicule data than inthe glass results. We propose that part of the scatter may be attrib-uted to small variations in positioning of the fine spicule and glassfibers before testing in the DMA apparatus, part to the formationsof SiOH groups as described above, but mainly to the observationthat the outer spicule fiber layers, while sustaining initial damage,are very thin (�0.1 lm), and are sacrificed during initial loadingwithout having much effect on the total energy dissipation. Suchsacrificial layers are not available in the glass fibers, and the lattershow larger scatter in the data.

0

5

10

15

20

25

30

Ene

rgy

to F

ailu

re (M

pa/m

^2) Dry

SpiculeGlass

1 Hr Sea

1 Hr Distilled

Fig. 12. Energy dissipation for spicules and glass under dry and wet conditions inbending.

3.2. Frequency scans

Frequency scans provided information on the viscoelastic prop-erties and damping characteristics of both the spicules and theglass fibers. Three of the values that can be obtained from fre-quency scans are E0, E0 0 and tan d. E0 is the elastic portion of themodulus and indicates how much energy is stored. E0 0 is the lossportion of the modulus, which indicates how much energy is lost.Tan d is the ratio of the two modulus values (E0 0/E’) and is a mea-sure of the damping capacity of the material. This value indicateshow well the material loses energy to friction, heat or molecularrearrangements.

When wet, there is little observed change in the behavior ofglass. However, for spicules, the amount of damping changes dra-matically. The tan d spectrum showed peaks in saturated spiculesin the intermediate range of frequency, from �25 to �35 Hz, andwas more than five times larger than in the dry condition(Fig. 13). From the presence of many discontinuities and suddenchanges in the tan d curves of Fig. 13 and in the E0 0 values ofFig. 14, it was concluded that a number of physical events, suchas misalignment of the spicule fiber during the testing, delamina-tion, crack diversion, cracking of packets of cylinders, etc., couldbe taking place in the spicules during deformation. However, dueto the inability to observe the specimens in the DMA apparatusduring testing, and the dynamic nature of the events, specificmodes that were contributing to the deformation could not be

Fig. 14. Representative measured values for E’ and E” taken from DMA data forspicule and glass fibers after 1 h exposure to moisture (seawater), and in the drystate for the spicules (no significant changes noted for glass in the dry state).

M. Johnson et al. / Acta Biomaterialia 6 (2010) 2181–2188 2187

identified before final failure occurred. In the moist condition, theincreased damping in the spicules is probably related to softeningdue to plasticization of the thin organic layers that has been refer-enced earlier. The presence of several peaks may reflect that fail-ures occur in packets of cylinders, as seen in the SEM image ofFig. 11, rather than by failure of individual cylinders. Since the cyl-inders that comprise these packets are at different distances fromthe central axis of the spicule fibers, the stress levels at which theyactivate for slippage would differ. At the present time, there is noexplanation of why breakage of cylinders seems to occur in pack-ets, rather than as individual cylinders.

L

P

Fig. A1. Bending load on spicule.

R2

4. Conclusions

It was interesting to note that rough, but very thin, outer layersof what had been originally judged as smooth spicule sections havebeen recently shown to be sacrificial, and thus protective of the in-ner structure [19]. It was, however, concluded that the failure ofthe outer layers at low stresses had a deleterious effect on the abil-ity to quantify modulus changes.

The damping characteristics of spicule fibers, comprising astructure of concentric cylinders of hydrated biosilica separatedby very thin layers of organic material, were found to be signifi-cantly greater than those of melt-drawn silica fibers of similardiameters.

Furthermore, the results of this work have shown very signifi-cant effects of moisture on the damping of the spicule fibers, asindicated by the large changes in tan d that were measured in bothdistilled water and in seawater. In the spicules, those changesshowed at least a 5-fold increase between the wet and the drystates. The damping has been interpreted as stemming from theviscoelastic behavior of the spicule fibers and to plasticization bymoisture of the thin organic layers between the silica cylinders.It should be emphasized that viscoelastic damping was seen inboth wet and dry spicule specimens, a fact that suggests that otherwork on mechanical behavior of hexactinellid sponges should bere-examined to account for such effects in both dry and wet condi-tions. The small shift in peak tan d between seawater and distilledwater has not yet been addressed. In addition, the results of nano-indentation tests that have been reported on the cross-sections ofspicule fibers (e.g. [8,20] should be re-evaluated because of time-dependent relaxations.

Upon examination of many failed surfaces, there were no signif-icant differences discerned between dry and wet samples of spic-ules or between dry and wet glass specimens. The spongespicules exhibited a weakening process similar to that seen inthe glass samples. Soaking the samples before testing proved to de-crease the maximum stress of fracture, with little increase in themaximum extension until failure. This led to a decrease in energydissipation in both liquids, though it was slightly more pronouncedin seawater. The cause of this decrease in toughness has been re-lated to the degradation of Si–O–Si bonds in the glass fibers andspicules and to the plasticization of the organic interlayers in thespicules.

Finally, it should be mentioned that while the mechanicalbehavior of wet vs. dry spicules, and especially the damping char-acteristics, are substantially different, the properties of the spiculesof living, as compared to non-living, sponges, may be different yetagain.

R1

Fig. A2. Cross-section of spicule showing central core.

Acknowledgments

The authors would like to thank Kathy Krogsland of the Univer-sity of Washington Marine Science Center for supplying the seawa-ter, Dr. Tom Murray, former Director of the Electron Microscopy

Center of the MSE Department of the University of Washington,for critical assistance with SEM, and Paul Kannikeberg for helpwith the experiments. This work was based on the MS thesis ofS.L.W. at the University of Washington.

Appendix A. Calculations of energy dissipation in bending ofspicule fibers

The calculations of true stress in bending were determined fromthe following equations. In the equation, 1/3 of the original R2 wasused for R1, which represents the outside and central core failureregion of the spicule samples (see Figs. A1 and A2).

True stressðbendingÞ; r ¼ McI

Flexural strain; e ¼ 6Dd

L2

Bending moment; M ¼ PL4

Moment of inertia; I ¼ 0:25pR42 � 0:125pðR4

2 � R41Þ

Distance from the neutral axis from to the point of maximum

stress on the periphery of the spicule; c ¼ 4ðR31 þ R3

2Þ3pðR2

1 þ R22Þ

where: r = true stress in bending; D = deflection; d = depth of thebeam (from the sample diameter); P = load; R2 = radius of the out-side of the spicule; R1 = radius of the central core; M = bending mo-ment; c = distance from the neutral axis from to the point ofmaximum stress on the periphery of the spicule as seen in Fig. 9([21]); L = length between supports; I = moment of inertia (from[21,22]); W = work; Af = cross-sectional area of the spicule at fail-ure; ef = strain at failure; E = energy dissipated in bending.

Final equation:

r ¼ 8PLðR31 þ R3

2Þ3pððR2

1 þ R22Þð0:25pR4 � 0:125pðR4

2 � R41ÞÞ

Work; W ¼ 0:5ref

2188 M. Johnson et al. / Acta Biomaterialia 6 (2010) 2181–2188

Cross-sectional area of the spicule at fracture; Af

¼ 0:5pR21 þ 0:5pR2

2

Energy dissipated in bending; E ¼WAf

Appendix B

Figures with essential colour discrimination. Certain figures inthis article, particularly Figures 4, 5, 6, 7, 8, 12, 13, and 14, are dif-ficult to interpret in black and white. The full colour images can befound in the on-line version, at doi: 10.1016/j.actbio.2009.12.006.

References

[1] Doremus RH. Glass science. 2nd ed. New York: Wiley-Interscience; 1994. pp.185–198.

[2] Dorrington KL. The theory of viscoelasticity in biomaterials. In: Vincent JFV,Currey JD, editors. The mechanical properties of biological materials (Symp.Soc. Exp. Biol. 34). Cambridge: Cambridge University Press; 1980. p. 289–314.

[3] Bonthrone KM, Vollrath F, Hunter BK, Sanders JKM. The elasticity of spiders’webs is due to water-induced mobility at a molecular level. Proc R Soc Lond B1992;248:141–4.

[4] Vincent JFV. Functional aspects of the hydration of two natural materials-insect cuticle and pine-cone bract. Proc MRS Symp, 2006;898E:L07-01.1-01.6.

[5] Vincent JFV. Structural biomaterials. Princeton, NJ: Princeton University Press;1990. p. 169.

[6] Mayer G, Rodriguez M, Tran K, Song H, Ma WH, Trejo, R, et al. Lessons for newclasses of inorganic/organic composites from the spicules and skeleton of thesea sponge Euplectella aspergillum. In: Mechanical properties of bioinspiredand biological materials, Mater. Res. Soc. Symp. Proc. Vol. 844, 83; 2005.

[7] Walter SL. Studies of the mechanical behavior of the spicules of the spongeEuplectella aspergillum, MS thesis, Department of Materials Science &Engineering, University of Washington, Seattle, WA; 2006.

[8] Sarikaya M, Fong H, Sunderland N, Flinn BD, Mayer G, Mescher A, et al.Biomimetic model of a sponge–spicular optical fiber – mechanical propertiesand structure. J Mater Res 2001;16:1420–8.

[9] Aizenberg J, Sundar VC, Yablon AD, Weaver JC, Chen G. Biological glass fibers:correlation between optical and structural properties. Proc Nat Acad Sci (USA)2004;101:3358–63.

[10] Müller WEG, Wendt K, Geppert C, Wiens M, Reiber A, Schröder HC. Novelphotoreception system in sponges? Unique transmission properties of thestalk spicules from the hexactinellid Hyalonema sieboldi. Biosens Bioelectron2006;21:1149–55.

[11] Walter SL, Flinn BD, Mayer G. Mechanisms of toughening of a natural rigidcomposite. Mat Sci Eng C 2007;27:570–4.

[12] Walter SL, Flinn BD, Mayer G. Effects of loading rate on the mechanicalbehavior of a natural rigid composite. Acta Biomater 2007;3:377–82.

[13] Hooper JNA, Van Soest RWM, editors. Systema porifera: a guide to theclassification of sponges, Vol. 2. Dordrecht: Kluwer; 2002. p. p. 1393.

[14] Müller WEG, Eckert C, Kropf F, Wang X, Schlossmacher U, Seckert C, et al.Formation of giant spicules in the deep-sea hexactinellid Monorhaphis chuni(Schulze 1904): electron-microscopic and biochemical studies. Cell Tissue Res2007;329:363–78.

[15] Ma W-H. Mechanical properties and physical characterization of the spiculesof the sponge Euplectella aspergillum, BS thesis, University of Washington,Seattle, WA; 2006.

[16] Dove PM, Rimstidt JD. Silica–water interactions. In: Silica – physical behavior,geochemistry, and materials applications, Reviews in Mineralogy, Vol. 29.Washington, DC: Mineralogical Society of America; 1994. p. 259–308.

[17] Doremus RH. Private communication; 2007.[18] Aizenberg J, Weaver JC, Thanawala MS, Sundar VC, Morse DE, Fratzl P. Skeleton

of Euplectella sp.: structural hierarchy from the nanoscale to the macroscale.Science 2005;309:275–8.

[19] Mayer G, Zhou J. The role of the organic component in the mechanical behaviorof biomineralized composites. Mater. Res. Symp. Proc. Vol. 1187, KK07-01;2009.

[20] Woesz A, Weaver JC, Kazanci M, Dauphin Y, Aizenberg J, Morse DE, et al.Micromechanical properties of biological silica in skeletons of deep-seasponges. J Mater Res 2006;21:2068–78.

[21] Alexander NA. Course notes for structural mechanics – centroids, secondmoments of inertia, University of Bristol, UK, 1996. http://www.cen.bris.ac.uk/personal/cenaa/Nick’s%20Web%20site/notes/Mechanics%20I/tutorial.pdf.

[22] Luebkeman CH, Peting D. Course notes for structures I. University of Oregon,Eugene, OR; 1996. http://www.uoregon.edu/~struct/courseware/461/461_lectures/461_lecture28/461_lecture28.html.