author's personal copy · f. jroundi et al./acta biomaterialia 10 (2014) 3844 3854 3845....

This article appeared in a journal published by Elsevier. The attachedcopy is furnished to the author for internal non-commercial researchand education use, including for instruction at the authors institution

and sharing with colleagues.

Other uses, including reproduction and distribution, or selling orlicensing copies, or posting to personal, institutional or third party

websites are prohibited.

In most cases authors are permitted to post their version of thearticle (e.g. in Word or Tex form) to their personal website orinstitutional repository. Authors requiring further information

regarding Elsevier’s archiving and manuscript policies areencouraged to visit:

http://www.elsevier.com/authorsrights

http://www.elsevier.com/authorsrights

Author's personal copy

Consolidation of archaeological gypsum plaster by bacterialbiomineralization of calcium carbonate q

Fadwa Jroundi a, Maria Teresa Gonzalez-Muñoz a, Ana Garcia-Bueno b, Carlos Rodriguez-Navarro c,⇑a Department of Microbiology, Faculty of Sciences, University of Granada, Avda Fuentenueva s/n, 18002 Granada, Spainb Department of Painting, Faculty of Fine Arts, University of Granada, Avda Andalucía s/n, 18071 Granada, Spainc Department of Mineralogy and Petrology, Faculty of Sciences, University of Granada, Avda Fuentenueva s/n, 18071 Granada, Spain

a r t i c l e i n f o

Article history:Available online 18 March 2014

Keywords:BiomineralizationCalcium carbonateVateriteGypsum plasterConservation

a b s t r a c t

Gypsum plasterworks and decorative surfaces are easily degraded, especially when exposed to humidity,and thus they require protection and/or consolidation. However, the conservation of historical gypsum-based structural and decorative materials by conventional organic and inorganic consolidants shows lim-ited efficacy. Here, a new method based on the bioconsolidation capacity of carbonatogenic bacteriainhabiting the material was assayed on historical gypsum plasters and compared with conventional con-solidation treatments (ethyl silicate; methylacrylate–ethylmethacrylate copolymer and polyvinyl buty-ral). Conventional products do not reach in-depth consolidation, typically forming a thin impervioussurface layer which blocks pores. In contrast, the bacterial treatment produces vaterite (CaCO3) bioce-ment, which does not block pores and produces a good level of consolidation, both at the surface andin-depth, as shown by drilling resistance measurement system analyses. Transmission electron micros-copy analyses show that bacterial vaterite cement formed via oriented aggregation of CaCO3 nanoparti-cles (�20 nm in size), resulting in mesocrystals which incorporate bacterial biopolymers. Such abiocomposite has superior mechanical properties, thus explaining the fact that drilling resistance of bio-consolidated gypsum plasters is within the range of inorganic calcite materials of equivalent porosity,despite the fact that the bacterial vaterite cement accounts for only a 0.02 solid volume fraction. Bacterialbioconsolidation is proposed for the effective consolidation of this type of material. The potential appli-cations of bacterial calcium carbonate consolidation of gypsum biomaterials used as bone graft substi-tutes are discussed.

� 2014 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.

1. Introduction

Gypsum-based mortars and plasters have been used since theadvent of pyrotechnology (from around 12000 BC) [1] in architec-tural applications as joint mortars, to cover masonry or to supportpaintings and decorations [2]. Gypsum was a rather commonbuilding and decorative material in the Levant [1] and ancientEgypt [3,4], as well as in the Mediterranean countries, especiallyduring the Middle Ages [2,5]. There are numerous examples ofmedieval Islamic architecture where gypsum was cast and/orcarved as highly sophisticated decorative structures, such as themocarabes from the Alhambra in Granada, Spain (twelfth to fif-teenth centuries) (Fig. 1A). Up to now, gypsum has been widely

used in plasterworks found in old and modern buildings world-wide [6–8].

In addition to its applications as a building material, calciumsulfate has been used as a ceramic bone graft substitute for morethan 100 years [9–11]. This biomaterial has been applied for bonerepair either as a self-setting paste or as preset solid bodies (blocksor pellets) [10,12]. Its biomedical applications include treatment ofbone defects, endodontic surgery, guided tissue regeneration, sinusaugmentation and drug delivery [10]. Among the main advantagesof calcium sulfate as a biomaterial are its high biocompatibility, ra-pid and complete resorption and remarkable osteoconductivity[10]. However, its low strength and rapid resorption can be disad-vantageous in some circumstances [10–12]. This has prompted itsuse as a composite biomaterial in combination with inorganic (e.g.,calcium phosphates, carbonates and/or silicates) and organic mate-rials (e.g., carboxymethylcellulose, gelatin, hyaluronic acid, chito-san and polyacrylic acid) [10–13].

http://dx.doi.org/10.1016/j.actbio.2014.03.0071742-7061/� 2014 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.

q Part of the Biomineralization Special Issue, organized by Professor HermannEhrlich.⇑ Corresponding author. Tel.: +34 958 246616; fax: +34 958 243368.

E-mail address: [email protected] (C. Rodriguez-Navarro).

Acta Biomaterialia 10 (2014) 3844–3854

Contents lists available at ScienceDirect

Acta Biomaterialia

journal homepage: www.elsevier .com/locate /ac tabiomat


Gypsum-binder technology involves an initial thermal treat-ment of gypsum (heating to �120–160 �C) to form the hemihy-drate (the mineral bassanite, also called ‘‘plaster of Paris’’)according to the reaction [7]:

CaSO4 � 2H2O$ CaSO4 � 0:5H2Oþ 1:5H2O ð1Þ

When the hemihydrate is mixed with water, the fresh mix setsand rapidly hardens via the formation of an interconnected porousstructure of acicular and/or prismatic gypsum crystals that impartstrength to the plasterwork (i.e., Reaction (1) is reversible) [14].Gypsum is thus suitable for many architectural and decorativeapplications because, in addition to its fast setting, it possesseshigh plasticity, good adhesion, ease of application and adequateprotective behavior. However, its softness (Mohs hardness of 2),poor mechanical behavior (typical values for gypsum plaster with

�30–50% porosity are: �4–7.5GPa Young’s modulus, �7–18 MPauniaxial compressive strength, 63 MPa uniaxial tensile strength)[14,15] and relatively high solubility (pKsp = 4.62) [16], favor itsdegradation, especially when exposed to humidity (rain, condensa-tion and/or rising damp) [5]. The implementation of conservationstrategies, typically including consolidation, is thus needed to limitdeterioration [8,17]. However, consolidation of artistic and histor-ical gypsum plaster is a poorly explored and challenging issue.Consolidation could be achieved through the application of organicor inorganic consolidants commonly used in stone conservation[18,19]. This is the case of acrylic, epoxy or (poly)vinyl polymers,and their copolymers [4]. However, polymer protectives and con-solidants are generally incompatible (physically and chemically)with the inorganic substrate (i.e., stone or plaster) they are appliedto, and may result in exacerbated damage [20]. Alkoxysilanes havealso been applied for consolidation, especially on silicate sub-strates such as sandstones [21]. The formation of amorphous silicaupon sol–gel transition offers some advantages as the end productis inorganic and can act as a new cement. However, the formationof superficial films and/or hard crusts [20], drying-induced crack-ing of the silica gel [22,23] and poor bonding with non-silicate sub-strates [16] may undermine the effectiveness of such a treatment.Recently, nanolimes have been proposed as a consolidant for gyp-sum stucco [24], yet little is known about its effectiveness. To theseshortcomings, one has to add economic, environmental, health andsafety issues that must be considered in determining the suitabilityand compatibility of a given treatment [25].

An alternative to standard chemical treatments is the use of thecarbonatogenic capacity of some bacteria to consolidate/protectbuildings and decorative inorganic materials [26–33]. It is an envi-ronmentally friendly bioconsolidation strategy based on the capac-ity of bacteria to induce the new formation of calcium carbonatebiocement within the porous system of materials [28]. Bacteriallyinduced mineralization occurs because bacteria can change thechemistry of their environment as a result of their metabolic activ-ity, and their cell structures and debris, as well as the exopolymericsubstances (EPS) they secrete, can act as nuclei for heterogeneousCaCO3 crystallization [28,29,34]. In the last few decades, bacterialbiomineralization processes and their application for the conserva-tion of ornamental stones have been studied by a number ofresearchers (see reviews in Refs. [35,36]). Gonzalez-Muñoz et al.[37] developed a new bioconsolidation method based on the selec-tive activation of carbonatogenic microbiota inhabiting stone bythe application of a suitable nutritional solution. Both the airbornebacteria as well as the stone-inhabiting carbonatogenic microor-ganisms can be stimulated to grow and to induce the in situ forma-tion of new calcium carbonate biocement, which effectivelyconsolidates decayed stones [29,30]. This last method possessesthe principal advantage that it does not require the application ofa bacterial culture, thereby circumventing any potential problem/complexity associated with the culture and application of foreignbacteria for calcium carbonate production.

Despite the numerous studies on bacterial carbonatogenesis,little is known about gypsum plaster consolidation using bacterialcalcium carbonate precipitation. To our knowledge, only one paperreports on the application of a commercial bacterial conservationtreatment (Bacillus cereus cultures) on laboratory-made gypsumplaster samples [38]. However, the aim of such a study was tounambiguously identify bacterial CaCO3 precipitates. No attemptto evaluate the effectiveness of the treatment in terms of consoli-dation efficacy was made. Up to now, the application of this newconsolidation strategy was focused on the conservation of carbon-ate stones [29,30] and sandstones [39], as well as on the protectionand repair of concrete and cement structures [40]. Little is knownon the influence of a gypsum substrate on bacterial CaCO3 precip-itation, on the effectiveness of the bacterial conservation treatment

B

A

B

Fig. 1. Ancient gypsum plasterwork: (A) example of decorative gypsum plaster atthe Lions Court in the Alhambra (Granada, Spain); (B) representative decoratedgypsum plaster recovered from the archaeological site ‘‘Alcázar de Guadalajara’’(Guadalajara, Spain).

F. Jroundi et al. / Acta Biomaterialia 10 (2014) 3844–3854 3845


itself or on the effectiveness of activating the autochthonous car-bonatogenic bacteria present in historical gypsum plasters.

Here, we use this new conservation methodology to bioconsol-idate archaeological gypsum plasters. We take into account thatthe traditional preparation of gypsum plasters often required theaddition of lime, which positively contributes to the mechanicalproperties to the material once gypsum and CaCO3 precipitate[6]. The formation of bacterial carbonates is assumed to play a sim-ilar cementing role as that associated with the traditional additionof lime. Application of such a new conservation methodology ongypsum plasters will involve both the consolidation of the weath-ered material with a biocement compatible with the substrate andthe possibility of differencing the newly formed bacterial calciumcarbonate from the original calcite present in the treated material,which is not trivial [38]. The main goal of the present study is tocompare the consolidation effectiveness of different conventionalconsolidants and our newly developed method based on CaCO3

production in situ, via the activation of the carbonatogenic micro-biota present in historical gypsum plasters. Special emphasis isplaced on evaluating the mechanical properties of the bacterialbiocement and its relationship with the mesostructure of newlyformed CaCO3 precipitates. Note that while many studies have fo-cused on the evaluation of the mechanical properties of severalCaCO3 biominerals such as seashells [41–45] and their relationwith the biomineral micro- and nanostructure [43,46,47], little isknown regarding the mechanical properties and micro/nanostruc-ture of bacterial CaCO3.

2. Materials and methods

Remains of historical gypsum plasters were collected from thearcheological site ‘‘Alcázar de Guadalajara’’ (thirteenth to fifteenthcenturies) in Guadalajara, Spain. The size of the selected pieces (15samples) was �5 cm � 2 cm � 4 cm and some of them were carvedin decorative reliefs (Fig. 1B).

2.1. Treatment with conventional consolidants

Three sets of gypsum plaster pieces (three samples per set; a to-tal of nine samples) were treated with the following conventionalconsolidants: (i) commercially available tetraethoxysilane (TEOS),commonly known as ethyl silicate (Bio Estel, from CTS) (as-re-ceived ethyl silicate solution was further diluted to 50% v/v in eth-anol); (ii) a copolymer of ethylmethacrylate and methylacrylatemonomers (PEMA/PMA) (Paraloid™ B-72, from Dow Chemical)prepared as a 2.5% w/v toluene solution; and (iii) polyvinyl butyral(PVB) (Mowital™ B60 HH, supplied by CTS) prepared as a 2.5% w/vethanol solution. The three consolidants were applied by brush onthe upper surface of each sample. Because the material was highlyweathered, as shown by its powdering and tendency to disinte-grate (even under careful handling), the treatment involved sevensuccessive impregnations. Between each impregnation, plastersamples were left interposed with Japanese paper at room temper-ature for 24 h until they were completely dry. Following treatmentapplication, specimens were kept at room temperature for 7 daysprior to testing in order to ensure complete solvent evaporation.

2.2. Bioconsolidation methodology

Bacterial bioconsolidation was carried out as described in detailin Ref. [32]. Briefly, pieces of gypsum plaster (four samples) weretreated with sterile M-3P nutritional solution (1 wt.% Bacto Casi-tone (a pancreatic digest of casein), 1 wt.% Ca(CH3COO)2�4H2O,0.2 wt.% K2CO3�½H2O in a 10 mM phosphate buffer, pH 8), accord-ing to a patent by Gonzalez-Muñoz et al. [37]. The solution was

sprayed on the upper surface of the samples until saturation, twicea day, 6 days in succession. Control samples (two blocks) weretreated with sterile distilled water.

It could be argued that the presence of calcium and carbonateions in M-3P culture medium would eventually lead to the precip-itation of calcium carbonate, without the need for bacterial pres-ence/activity. However, our previous studies have shown nocalcium carbonate formation in M-3P in the absence of bacterialactivity. Rather, a poorly crystalline (nearly amorphous) calciumphosphate (hydroxylapatite) phase formed, which produced noconsolidation effect [28].

In order to analyze the microbiota present in gypsum samples,grains (�300 mg) were collected from three different spots (usingsterile tweezers) before and after biotreatment. These grains wereadded to 1 ml sterile physiological solution (0.9 wt.% NaCl), gentlydispersed, and aliquots were taken to perform serial dilutions andinoculations on different selective culture media to test acid pro-ducing bacteria, nitrifying and sulfur-oxidizing bacteria, as wellas the possible presence of fungi [32].

To test the carbonatogenic capacity of the microbiota presentbefore and after the bioconsolidation treatment, plates of M-3P so-lid culture medium were inoculated with aliquots from the above-mentioned serial dilutions and incubated at 28 �C for 1 week. Petridishes were periodically examined for the presence of CaCO3 crys-tals by optical microscopy for up to 25 days. Controls consisting ofuninoculated culture medium were incubated along with inocu-lated samples.

2.3. Assessment of consolidation effects

Drilling resistance (DR) was measured by a drilling resistancemeasurement system (DRMS; Sint Technology). The DRMS is anew portable instrument designed to perform a precise ‘‘drillingresistance’’ test through the continuous measurement of the forcenecessary to drill a hole in the material under specific operatingconditions [48–51]. Recent results place this technology as themost promising for the evaluation of consolidation performance,particularly in relatively soft porous materials [48,50,52–54]. Dur-ing testing, both rotational speed (x) and penetration rate (t) aremaintained constant. In this study, test conditions were as follows:(i) 5 mm diameter Diaber drill bit (Sint Technology) with flat-edged diamond-tip (a Macor� calibration standard of known DRwas drilled to ensure that no variation in DR associated with wearof the drill bit tip took place during the course of the tests); (ii) tset at 20 mm min–1; (iii) x set at 200 rpm (the selected low rota-tional speed yielded best results in terms of reproducibility whendrilling a soft material such as gypsum). Under this conditions, ddefined as the cutting depth per revolution (d = 2pt/x) is equalto 0.62 mm per revolution. Between three and seven drill testswere carried out on each historical specimen before and aftertreatment (11 samples tested). DR measurements were also per-formed on geologic centimeter-sized (optical quality) gypsumcrystals (two samples) in order to obtain a reference value for azero-porosity gypsum material. We also tested the DR of marble(calcitic Macael marble; 0.5–1% porosity) and porous limestone(calcarenite; 25–28% porosity) in order to have reference valuesfor carbonate stones. DR results were normalized to drill bit diam-eter: i.e., the final DR units were N mm–1 [51]. It should be notedthat DR results show a good linear correlation with mechanicalproperties such as uniaxial compressive strength and biaxial flex-ural strength, which includes the combined effect of compressiveand tensile strength [51]. As expected, DR shows an exponentialvariation with Mohs hardness [51]. DR values have been shownto increase linearly with d [55]. This enables comparison of pub-lished DR values obtained using different t and x values.

3846 F. Jroundi et al. / Acta Biomaterialia 10 (2014) 3844–3854


The porosity and pore size distribution of gypsum plasters be-fore and after treatment were determined by means of mercuryintrusion porosimetry (MIP) using a Micromeritics Autopore5510 device. A total of 13 samples were analyzed by MIP. The tex-ture and distribution of the different consolidants on and withintreated plaster samples were studied by scanning electron micros-copy (SEM) using a Gemini SEM (Carl Zeiss SMT) equipped with afield emission gun and coupled with energy dispersive X-ray spec-trometry (EDS). Both the surface and cross-sections of each samplewere observed to evaluate the penetration depth of each treat-ment. All specimens were carbon-coated prior to observation.

Color changes were evaluated by measuring the CIEL⁄a⁄b⁄ colorparameters of untreated and treated samples (five measurementsper sample) [56]. Measurements were done using a MinoltaChroma Meter portable spectrophotometer equipped with axenon lamp (illuminant C) and diffuse reflectance geometry. Totalcolor variations were reported as DE = (DL⁄2 + Da⁄2 + Db⁄2)½, whereDL⁄, Da⁄ and Db⁄ are the difference between untreated and treatedstone of L⁄ (lightness: 0 being black and 100 being diffuse white),a⁄ (negative values indicate green whilst positive values indicatemagenta) and b⁄ (negative values indicate blue and positive valuesindicate yellow) values, respectively. Note that the human eye isonly able to detect chromatic changes for DE > 3 [56].

2.4. Analysis of bacterial calcium carbonate

Mineralogy and crystallite size/strain of bacterial calcium car-bonate(s) were determined by X-ray diffraction (XRD) on an X’Pertsystem (PANalytical, The Netherlands) with Ni filter. Measurementparameters: CuKa radiation k = 1.5406 Å, 45 kV, 40 mA, explorationrange from 4� to 70� 2h, steps of 0.001� 2h and goniometer speed of0.01� 2h s�1. Powders were deposited in zero-background Si sam-ple holders. Mineral phases were identified by comparison withJCPDS (Joint Committee on Powder Diffraction Standards) powderspectra. Crystallinity or crystallite size Dhkl (i.e., coherent X-rayscattering domains) of bacterial CaCO3 was calculated using theWarren modification of the Scherrer equation [57]:

Dhkl ¼kk

bhkl cos hð2Þ

where k is the crystallite-shape factor (�0.9), k is the X-ray wave-length, bhkl is the X-ray diffraction broadening (i.e., hkl Bragg peakfull width at half maximum: FWHM, in radians) and h is the Braggangle. Peak broadening is also affected by lattice strain e, which isgiven by [58]:

e ¼ bhkl

4 tan hð3Þ

Assuming that the particle size and strain contributions to peakbroadening are independent, a combination of Eqs. (2) and (3)gives the Williamson–Hall equation [59]:

bhkl cos h ¼ kkDhklþ 4e tan h ð4Þ

From the linear fit to the data (plot of 4 tan h vs. bhkl cos h), Dhkl

is determined from the y-intercept, and e from the slope of the fit.Peak broadening analysis was performed using the XPowder 1.2software package, which enables background subtraction, Ka2

stripping and instrumental broadening correction [60]. Peak profilefitting was performed using Gaussian, Lorentzian and pseudo-Voi-gt functions. Best fits were systematically observed when using aLorentzian function. After fitting, Williamson–Hall plots were per-formed to obtain mean crystallite size and strain (latticedistortion).

Thermogravimetric (TG) and differential scanning calorimetry(DSC) analyses of untreated and bacterially treated gypsum

plasters were performed to quantify both the amount of gypsumand carbonate phases originally present in the plasters, as well asthe amount of newly formed bacterial CaCO3 and organic by-prod-ucts of bacterial activity. Powder samples with a mass of �50 mgwere analyzed on a METTLER-TOLEDO TGA/DSC1 under the follow-ing experimental conditions: 50 ml min–1 air flow, 25–950 �C heat-ing interval and 20 �C min–1 heating rate.

Analysis of the morphology, size and nanostructure of bacterialcarbonates was performed by means of transmission electronmicroscopy (TEM) using a Philips CM20, operated at 200 kV. Priorto analysis, bacterial precipitates were dispersed in ethanol anddeposited on carbon/Formvar� film coated copper grids. TEMobservations were performed using a 40 lm objective aperture. Se-lected area electron diffraction (SAED) patterns were collectedusing a 10 lm aperture, which allowed collection of diffractiondata from a circular area �0.2 lm in diameter. In order to unam-biguously identify the zone-axis of bacterial carbonate SAED pat-terns, modeling of SAED patterns was performed using theWebEMAPS software (developed by the Material Research Labora-tory, Department of Materials Science and Engineering, Universityof Illinois at Urbana-Champaign, available at: http://emaps.mr-l.uiuc.edu/).

3. Results

DR curves of gypsum plasters before and after treatments appli-cation are shown in Fig. 2. (Note that DRMS curves systematicallyshowed an increase in DR from zero to the average value of thematerial within the first 0–1 mm, as reported elsewhere[48,50,52–54].) The studied plasters were relatively soft materials,as shown by the low average DR force (0.24–0.82 N mm–1) of un-treated specimens. This value was well below the 3.8 N mm–1 DRvalue of geologic gypsum crystals. The low DR values of unconsol-idated archaeological gypsum plasters are due to their high poros-ity (29–48%). Note that mechanical properties such as Young’smodulus, compressive strength and tensile strength, as well asdrilling resistance, of a range of porous brittle materials, includinggypsum plasters, show an exponential decay with increasingporosity according to the equation [54,61]:

S ¼ S0e�fP ð5Þ

where S and S0 are the strengths at porosity P and zero, respectively,and f is an empirical constant.

A good exponential fit according to Eq. (5) (correlation coeffi-cient of 0.959) was observed when plotting experimental DR val-ues vs. gypsum plaster porosity, P (Fig. 3). The experimentalfitting equation was:

DR ¼ 4:04e�0:05P ð6Þ

Plaster pieces treated with PEMA/PMA showed a clear increasein DR (up to �2.4 N mm–1) until a depth of �3 mm. However, be-low this hardened surface layer, DR decreased to a low value of�0.8 N mm–1, similar to that of the untreated material (with29.1% porosity). Samples treated with PVB showed a DR patternsimilar to that of the previous treatment. In this latter case, how-ever, the increase in DR was limited to a depth of �2 mm. Belowthis strengthened surface layer, DR decreased to a residual valueof �0.6 N mm–1, corresponding to that of the untreated material(porosity of 38.3%). Samples treated with TEOS showed DR valuesof �0.8 N mm–1 throughout the whole profile of the treated frag-ment. Considering that the average porosity of this sample was33.8%, the experimental DR value is in good agreement with thatcalculated using Eq. (6) (0.74 N mm–1), and matches the DR valueof an unconsolidated plaster of similar porosity. The best consoli-dation effects in terms of DR were observed following treatment



with M-3P culture medium (bioconsolidation treatment) (Fig. 3).As shown in Fig. 2A, an average drilling resistance of 1.7 N mm–1

was reached throughout the whole profile of the treated fragment.This value is higher by a factor of two than the DR value(0.87 N mm–1) corresponding to an unconsolidated gypsum plasterwith the porosity of this sample (31.5%). In gypsum plasters withthe highest porosity of 48% and with average DR of 0.24 N mm–1,the bacterial treatment decreased the porosity to 41% and in-creased DR up to 1.38 N mm–1 (Fig. 2B). The calculated DR (accord-ing to Eq. (6)) for such a latter porosity is 0.52 N mm–1, a valuesignificantly lower (by factor of �3) than the experimental one.Note, however, that in this latter case the strengthening effectwas not homogeneously distributed with depth: i.e., it extendedto a depth of �6 mm and the maximum consolidation was reachedat a depth of 1.5 mm.

SEM photomicrographs of the untreated plaster surface showedtangled gypsum crystals with acicular and/or bulky habit, thus pro-ducing rather rough surface features (Fig. 4A). Crystals tended toshow corrosion features indicative of chemical weathering (disso-lution) (Fig. 4B). After the biotreatment with M-3P, the surfaceappearance drastically changed due to the formation of an homo-geneous coating of newly precipitated micrometer-sized

spherulites of vaterite (see XRD results below), in contact withmineralized bacterial cells and bacterial EPS (Fig. 4C). These fea-tures were also observed in cross-sections, indicating a good pen-etration of the bioconsolidation treatment and a homogeneousdistribution of the bacterial carbonate cement throughout the poresystem of the treated plaster. Both surface and cross-sectionimages also showed that the M-3P treatment did not block thepores. A completely different texture was observed following treat-ment with PEMA/PMA (Fig. 4D), PVB (Fig. 4E) and TEOS (Fig. 4F).The SEM images show that in all these cases consolidant depositionwas limited to the formation of a surface film, which blocked thepores. Pervasive cracks were observed in the surface film formedafter treatment with ethyl silicate (Fig. 4F).

Fig. 5 shows the porosity and pore size distribution (determinedusing MIP) of a gypsum plaster before and after bacterial treat-ment. No significant shift in pore size was observed after treatment(i.e., changes were within error), although a porosity reduction(ranging from 5% to 15%) was detected. Such a porosity reductionis consistent with the deposition of bacterial CaCO3 cement plusthe organic by-products of bacterial activity (EPS). Due to the factthat conventional consolidants formed a surface film, no detectablechanges in porosity and pore size distribution of treated plasterwere observed. Note that samples used for standard MIP analysisare �1 cm3. As a consequence, the observed thin surface coating(a few micrometers thick) formed after treatment with conven-tional consolidants does not contribute to a detectable porosityreduction.

Table 1 shows the results of color measurements before andafter the treatment of gypsum plasters. Untreated plasters showeda medium–high L⁄ parameter (between 65 and 75) and a light yel-lowish color (i.e. b⁄ values were positive and ranged between 9 and13, while a⁄ values ranged between 2 and 5). After treatment, nosignificant changes in L⁄, a⁄ and b⁄ values were observed. In almostall cases, total color change values (DE) were �5 or slightly above(in the case of PEMA/PMA- and TEOS-treated plasters) and wereslightly higher than those of the control plasters (treated withsterile distilled water), which show an average DE value of 0.9.In summary, color measurements confirm that the treatments donot significantly alter the appearance of the gypsum plaster, thuscomplying with current conservation guidelines which indicatethat a DE value 65 is acceptable [62].

Fig. 6 shows XRD patterns of untreated and M-3P treated gyp-sum plaster specimens. The untreated gypsum plasters used in thisstudy consisted of gypsum (CaSO4�2H2O), with minor amounts of

0

0.5

1

1.5

2

2.5

0 2 4 6 8 10 12 14

Dril

ling

Res

ista

nce

(N/m

m)

Depth (mm)

M-3P

PEMA/PMA

Unconsolidated plasterPVB

TEOS

0

0.3

0.6

0.9

1.2

1.5

0 2 4 6 8 10 12 14

Dril

ling

Res

ista

nce

(N/m

m)

Depth (mm)

Unconsolidated plaster

M-3P

A

B

Fig. 2. DR of gypsum plasters. (A) DR measured on gypsum plaster pieces withinitial porosity of 29–38% treated with three conventional consolidants and M-3Pculture medium. (B) DR of untreated and bacterially treated gypsum plaster with aninitial porosity of 48%. Values are average of three to five drill holes.

0

1

2

3

4

0 10 20 30 40 50

Dril

ling

resi

stan

ce (N

/mm

)

Porosity (%)

Fig. 3. Variation of gypsum plasters DR vs. porosity. The solid line shows the best fitto Eq. (5) for non-porous gypsum crystals (N) and unconsolidated gypsum plasters(4). DR values corresponding to bacterially consolidated gypsum plasters are alsoshown (j). Values represent the average of three to five drilling tests.



quartz (SiO2) and calcite (CaCO3) (note that the main 104Calcite

Bragg peak at 3.034 Å is masked by the strong �141Gypsum peak at3.053 Å). As expected, no mineralogical changes were detected fol-lowing application of the three conventional consolidants. The bio-treated specimens showed the presence of vaterite in addition tothe previous phases (gypsum, quartz and calcite). Vaterite is ametastable polymorph of calcium carbonate, commonly producedduring bacterial carbonatogenesis [34]. Williamson–Hall plots cal-culated using the 112 and 224 harmonic Bragg reflections showedthat bacterial vaterite crystallite size was 21 nm, with a low latticestrain contribution of 0.05% (Fig. 7). The average crystallite size cal-culated using all vaterite Bragg reflections was 23 ± 1 nm with astrain of 0.09 ± 0.03% (inset in Fig. 7), values consistent with theabove results.

TG analyses of unconsolidated plasters showed a significantweight loss in the temperature range 95–250 �C, and a minimalbut continuous weight loss in the temperature range 250–450 �C(Fig. 8). In addition, the DSC trace showed an endothermal (split)band at 120–220 �C, followed by a smaller endothermal broadband at 280–380 �C (Fig. 8). These are standard features of thethermal decomposition of gypsum [63] and confirm that gypsumis the main phase (81–93 wt.%) of unconsolidated plaster. Thepresence of CaCO3 (calcite, according to XRD) was shown by the

Fig. 4. SEM photomicrographs of untreated and treated gypsum plasters: (A) untreated control; (B) detail of gypsum crystals showing corrosion features. Samples treatedwith: (C) M-3P, showing bacterial vaterite surrounded by EPS, and bacterial cells (bc; inset); (D) PEMA/PMA; (E); PVB; and (F) TEOS.

0

0.01

0.02

0.03

0.04

0.05

0.06

0

5

10

15

20

25

30

35

40

45

50

0.0010.010.11101001000

dV/lo

gR (c

m3 /g

)

Por

osity

(%)

Radius (µm)

Fig. 5. Representative MIP plots showing the cumulative intrusion curves (poros-ity) and pore size distribution curves (i.e., differential of intruded volume (dV) vs.log of the pore radius, R) for non-consolidated (solid lines) and biomineralized(dotted lines) gypsum plaster samples.



weight loss due to decarbonation in the temperature range 600–780 �C and the associated endothermal peak in the DSC trace(Fig. 8). The average amount of calcite in the untreated plaster ran-ged from 4 to 12 wt.%. TG analyses showed that the bacterial con-solidation treatment increased the CaCO3 content. The amount ofbacterial carbonate (vaterite) was 1.5–2 wt.%. The correspondingsolid volume fraction of bacterial vaterite was �0.02

(vaterite density 2.65 g cm–3) [34]. The presence of organics(�1.1–3.5 wt.%) was detected by TG/DSC, as shown by the weightloss in the temperature range 250–450 �C and the associated exo-thermal band in the DSC trace (Fig. 8). Assuming a density of1.3 g cm–3 for the organics [64], it can be estimated that the organ-ics formed following the bioconsolidation treatment (bacterial celldebries and EPS) account for a solid volume fraction of up to 0.07.The sum of vaterite and organics volume fractions is equal to 0.09,a value in good agreement (within error) with the solid volumefraction (0.11) calculated from the porosity difference (MIP results)between bacterially treated and untreated gypsum plasters.

TEM-SAED analyses showed that bacterial vaterite typicallyformed spherulitic aggregates a few micrometers in size (Fig. 9A).Higher magnification images showed that each spherulite wasmade up of an aggregate of rod-shaped vaterite crystals �20 nmthick and 100–200 nm long (Fig. 9B and E). The aggregate dif-fracted as a single crystal, with Debye spots slightly ellipsoidaldue to a small missorientation among nanoparticles (angularspreading <5�) (Fig. 9C). The SAED pattern in Fig. 9C correspondedto the vaterite [001] zone axis, as confirmed by SAED simulationusing the WebEMAPS software (Fig. 9D). The vaterite structure re-ported by Wang and Becker [65] was considered for SAED simula-tion. Fig. 9E shows a detail of the vaterite nanounits making up thebacterial CaCO3 structure. The rod-shaped vaterite nanocrystalswere typically elongated along [001]. These observations showthat vaterite structures are made up of an oriented aggregate ofnanocrystals. Within nanocrystals, low electron-absorbing areaswere observed. Energy dispersive X-ray spectroscopy (EDS)microanalysis showed the presence of S and P in these low-electron-absorbing areas. This is consistent with the incorporationof organics between vaterite nanounits: i.e., S and P are present inbacterial EPS and other byproducts of bacterial activity which getincorporated into the CaCO3 structure [34].

Table 1Spectrophotometric color measurements of gypsum plasters before and after bioconservation and conventional treatments.a

Sample/treatment Before treatment After treatment DE

L� a⁄ b⁄ L� a⁄ b⁄

Controlb 75.0 ± 4.4 3.4 ± 0.8 13.4 ± 1.9 75.4 ± 5.5 3.2 ± 0.7 12.6 ± 2.2 0.9 ± 0.5M-3P 65.1 ± 11.7 5.1 ± 2.6 12.9 ± 2.5 63.4 ± 16.8 7.6 ± 3.2 16.2 ± 2.3 4.5 ± 2.6PVB 65.1 ± 8.7 4.9 ± 1.3 12.9 ± 1.4 60.2 ± 5.9 5.9 ± 1.4 15.8 ± 1.8 5.8 ± 4.6PEMA/PMA 76.6 ± 5.3 2.5 ± 0.4 9.6 ± 0.8 76.1 ± 1.5 3.7 ± 0.7 14.6 ± 1.2 5.2 ± 1.8TEOS 67.1 ± 9.1 4.8 ± 1.8 13.1 ± 1.6 62.9 ± 7.7 6.3 ± 1.5 17.8 ± 1.5 6.5 ± 4.3

a Average values of 5 measurements per sample are presented.b Treated with sterile distilled water.

0

200

400

600

800

1000

1200

1400

1600

1800

0 10 20 30 40 50 60

°2θ

Inte

nsity

(a.u

.)

020Gy

121Gy

031Gy

141Gy

112Gy

121Gy

211Gy

022Gy

202Gy

200Gy

110Vat

112Vat 114Vat

_

_

_

_

_

_

101Qtz

222Gy141Gy

_

__152Gy

123Gy 062Gy

262Gy

A

B

Fig. 6. XRD patterns of gypsum plaster before (A) and after (B) M-3P treatment.Main Bragg reflections of gypsum (Gy; JPDF #360432), quartz (Qtz; JPDF #461045)and vaterite (arrows) (Vat; JPDF #330268) are indicated.

Fig. 7. Williamson–Hall plot for bacterial vaterite calculated considering broaden-ing of two order Bragg reflections (i.e., 112 and 224). Inset shows average Dhkl valuecalculated considering all vaterite Bragg reflections.

70

75

80

85

90

95

100

0 200 400 600 800 1000

T (°C)

Mas

s (%

)

-140

-120

-100

-80

-60

-40

-20

0

20

40

DSC

(mW

)

Fig. 8. Representative TG and DSC curves for untreated (solid lines) and bacteriallytreated (dashed lines) gypsum plasters.



Microbial growth and carbonatogenic activity tests showed that95% of the microbiota inhabiting the plaster and activated by theapplication of M-3P culture medium was carbonatogenic bacteria.No nitrifying bacteria, sulfur oxidizers or fungi were detected afterthe biotreatment. The total number of bacteria (culturable in M-3Pmedium) present before and after completion of the bioconsolida-tion treatment was not altered (�5.2 � 104 ufc g–1) and 10% ofthese bacteria were able to produce acids in a glucose-containingmedium. The latter clearly demonstrates that the glucose-freemedium (M-3P) used here is more suitable than the glucose-con-taining medium used by others [26].

4. Discussion

Our results show that, in general, treated gypsum plasters arestrengthened by both conventional and biomineralization consoli-dation treatments. However, conventional products show impor-tant limitations. Although TEOS is considered by some as themost promising consolidant for porous materials because of itsgood penetration [52], on gypsum plasters its efficacy is limitedby the formation of a highly cracked surface coating of amorphoussilica (SEM results), which produces negligible consolidation(DRMS results). Such a cracking is associated with drying stressesgenerated during solvent evaporation [22]. Drying-induced crackdevelopment is a mayor handicap of alkoxysilane consolidants[21]. In addition, the poor consolidation effectiveness of TEOS iswidely known for non-silicate stones such as limestones and isassociated with the low chemical affinity between calcite and silicaformed after hydrolysis and condensation of alkoxysilanes [21]. Apoor bonding is also expected between silica and gypsum due to

their structural and chemical dissimilarities. As an additional det-rimental effect, this treatment produces the highest change in col-or (yellowing; DE = 6.54). In summary, this treatment shows verylittle efficacy when applied for the consolidation of archaeologicalgypsum plasters. The consolidation effect of PEMA/PMA is limitedto a depth of �3 mm. According to SEM analysis, this consolidantleads to the formation of an impervious polymer film on the sur-face of the samples, which blocks the pores. The same detrimentaleffect has been observed when using PEMA/PMA to consolidateother porous materials such as old ceramic tiles [66]. Treatmentwith PEMA/PMA leads to a relative increase in strength, but nega-tively modifies the water permeability [66]. The effects of PVBapplication are very similar to those of PEMA/PMA: it forms animpervious surface film and only consolidates the surface downto a depth of a few mm. Overall, these results were not unexpected,since it is known that when applied for stone consolidation, allthese treatments typically increase strength, but they are only ableto impregnate a thin outer layer [52]. In summary, the limited pen-etration, formation of superficial (impervious) films, lack of appro-priate bonding with gypsum crystals and cracking (in the case ofethyl silicate) limit the effectiveness of these conventional treat-ments for archaeological gypsum plaster consolidation.

In contrast, the bacterial bioconsolidation treatment shows thehighest consolidation effectiveness. This type of treatment has pre-viously been demonstrated to be effective for the consolidation ofporous limestone, having the main advantage (if compared withprevious bacterial consolidation treatments) of not introducingany exogenous microorganism into the stone [29–33]. A significantstrengthening of gypsum plasters was achieved following treat-ment with M-3P (DR results). In most cases (i.e., samples withporosity ranging from 29 to 38%), the whole profile of the samples

Fig. 9. TEM photomicrographs of bacterial vaterite: (A) general view of bacterial vaterite spherulites forming micrometer sized aggregates; (B) detail of a vaterite structureshowing its internal mesostructure; (C) [001] zone axis SAED pattern of vaterite structure in (B). Note that the nanostructure diffracts as a single crystal. Extra spotscorrespond to an area with a slightly different orientation. (D) Vaterite [001] zone axis pattern simulated using WebEMAPS software; (E) high magnification image of vateritenanounits making up the structure in (B); (F) shows the EDS spectrum of the central area in (E).



was consolidated. M-3P penetration/consolidation depth could beassessed to be as great as 2 cm, taking into account that the piecesused in this study were 2 cm thick. This is due to the fact that bac-terial cells can colonize not only the surface of porous materials,but also deeper areas. Such a remarkable penetration depth maybe associated with two phenomena: (a) irrespective of their posi-tion in a depth profile, bacteria inhabiting the gypsum plaster areactivated as soon as the nutritional solution reaches their location;and (b) activated bacteria are transported in suspension (in the cul-ture medium) through the stone porous system [39]. Note thatmotility and growth of bacteria (under static conditions) have beenfound to result in transport rates of up to 0.4 cm h�1 in nutrient-saturated sand and sandstone [67]. The latter effect is facilitatedby the mean pore size (�2–4 lm diameter) of the gypsum plasters.Although the strengthening was also substantial in samples withthe highest porosity (48%), it was, however. limited to a thick sur-face slab of �6 mm. This is most probably due to the fast drying ofthis highly porous substrate which could have favored accumula-tion (back migration) of the nutritive solution and bacterial activitywithin this near-surface zone.

Although there is a porosity reduction of 5–15%, SEM observa-tions show that no plugging or blocking of pores takes place duringthe biocementation and no increase in the bacterial activity orchange in the microbiota inhabiting the stone are observed aftercompletion of the biotreatment. Apparently, the culture mediumcomposition, in particular the lack of glucose which could lead tounwanted development of acid-producing bacteria, is a key factorin the success of this treatment. It is crucial because determiningthe compatibility and success of a given treatment must take intoaccount economic, environmental and human safety issues, as wellas the lack of other detrimental side-effects [25]. Unlike conven-tional consolidants, which often result in the formation of incom-patible and often harmful surface films [20], bacterialcarbonatogenesis provides a high level of consolidation with nonegative environmental or health impacts because it is an environ-mentally friendly method which produces a compatible CaCO3 bio-cement following the activation of non-pathogenic carbonateproducing bacteria. This cement attaches gypsum grains, makingthe plasters more resistant to mechanical stress, simulated hereby the drilling forces [54].

The observed strengthening is associated with the formation ofbacterial vaterite biocement. Vaterite typically forms when bacte-rial carbonatogenesis occurs on substrates that do not share a com-positional or structural similarity with the most stable CaCO3

polymorph, which is calcite. This is the case of silicate substrates(glass and sandstones) where bacterial vaterite typically forms,while on calcitic substrates, bacterial calcite grows in crystallo-graphic continuity (epitaxy) [39]. A similar effect may help to ex-plain the formation of vaterite on gypsum, which can beconsidered both compositionally and crystallographically differentfrom calcite. Note however, that growth of calcite on gypsum withthe following (epitaxial) orientation relationship [001]gypsum//[001]calcite has been reported [68]. Thus, the above-mentionedsubstrate effect could have been at work here, resulting in bacterialcalcite production. However, in this particular case, the presence ofsulfate ions associated with the partial dissolution of gypsum fol-lowing M-3P treatment application appears to have favored the ki-netic stabilization of vaterite, preventing its transformation intomore stable calcite (according to the Ostwald’s rule of stages)[69]. Most important, the incorporation of organic byproducts ofbacterial activity into vaterite structures also appears to have fa-vored the kinetic stabilization of this metastable phase [34]. Itcould be argued that the formation of metastable vaterite (insteadof stable calcite) on the studied gypsum plasters is a handicap toreach the maximum consolidation potential of the treatment.However, DRMS results show that a significant strength gain is

achieved. In fact, bacterially consolidated gypsum samples displayDR values of the same magnitude as those of limestones and limemortars of equivalent porosity (Fig. 10). This is unexpected as gyp-sum strength is systematically (much) lower than that of lime-stones [51–53,70–72] or lime mortars [73] of equivalent porosity,and the amount of CaCO3 introduced by the bacterial treatmentis low: �2% of the solid volume of the plaster. Below we will showthat the mesostructural features of bacterial vaterite help to ex-plain this conundrum.

Crystallite size measurements and TEM-SAED analyses showthat the newly formed bacterial vaterite displays a mesocrystallinestructure formed by oriented rod-shaped nanocrystals �20 nmthick and 100–200 nm long. The nanocrystals are assembled inan oriented manner (with their c-axis parallel) and gaps betweenthese building units are filled with organics (mainly EPS). Theseare general features of bacterial vaterite, as we have reported ear-lier [34]. Bacterial vaterite precipitates are thus mesocrystals, i.e.crystals composed of mutually oriented nanoparticle mineral unitsembedded within an organic matrix [74]. These mesostructuralfeatures of bacterial vaterite are also shared by several CaCO3

biominerals, including aragonite in mollusk shell nacreous[42,75,76] and crossed-lamellar layers [76], as well as calcite inmollusk prismatic layer [44] and echinoderms spicles [77,78].The organic–inorganic hybrid (mortar and brick) structure, as wellas the nanostructured features (i.e., mesocrystal structure) ofbiominerals, has been claimed to be responsible for their amazingand unique combination of stiffness, hardness and extremely hightoughness (work of fracture), which can reach values orders ofmagnitude higher than those of their inorganic counterparts[41,46,76]. For instance, the unusual mechanical properties of mol-lusk nacre are related to the combination of stiffness imparted bythe CaCO3 mineral component, which absorbs the externally ap-plied loads, and the toughness, prevention of crack spreading andcapacity for recovery after deformation imparted by the organiclayer [79]. The high strength and toughness of CaCO3 biomineralshave also been associated with the presence of inorganic nanounits[46], because materials become insensitive to flaws at the nano-scale [80]. For materials such as calcite and aragonite (andvaterite), whose Young moduli are �100 GPa [81], a critical lengthscale of �30 nm is calculated, below which the fracture strength ofa flawed prism-shaped crystal is identical to that of a perfect crys-tal [80]. This size is close to the �20 nm average size of bacterial

0

1

2

3

4

5

6

7

15 20 25 30 35 40 45 50

Dril

ling

resi

stan

ce (N

/mm

)

Porosity (%)

Fig. 10. DR vs. porosity of carbonate stones (filled circles) and lime plasters (emptycircles) of different porosity, compared to gypsum plasters, both unconsolidated(4) and consolidated by bacterial vaterite (j). Solid lines show best fits to Eq. (5).Data for carbonate stones are from: this study; Ref. [52]; Ref. [71]; Ref.[72]. Data for lime plaster are from: s Ref. [73]. Values have been normalized tod = 0.65 using the reported linear relationship between DR and d [55].



vaterite nanounits determined here by means of XRD and TEM. Allin all, bacterial vaterite in consolidated gypsum plasters impart avery high mechanical strength and toughness due its mesostruc-tured inorganic–organic nature. Note that vaterite biomineralscan reach a high elastic modulus (e.g., up to 57 GPa, Ref. [82]),being as high as those of calcite and aragonite biominerals [45].

In addition to its strengthening effect, the bacterial treatmentmight also be effective to increase the resistance of gypsum plas-ters against chemical weathering (dissolution). It has been shownthat organics present in bacterial calcium carbonate cement(including EPS) contribute to the reduction of the dissolution rateof bioconsolidated limestones [28]. Apparently, bacterial biofilms(EPS) strongly inhibit the dissolution of calcite crystals [83]. Inthe case of bacterially consolidated gypsum plasters, we have ob-served that gypsum crystals are covered by EPS and bacterial vate-rite cement containing organics. Such organics may protectgypsum in a similar way as they do in the case of calcite, therebyreducing its dissolution rate. Further tests will be performed toevaluate this protective effect.

Bacterial mineralization of calcium carbonate may find interest-ing biomaterial applications. The poor mechanical behavior of pre-set gypsum bone graft substitutes [10] could be improved by theformation of bacterial CaCO3 cement. In this particular case, inocu-lation of the culture medium with carbonatogenic bacteria, such asMyxococcus xanthus [28] or Brevundimonas diminuta [39], should beperformed to ensure maximum bacterial biocementation. The factthat such a consolidation treatment results in the formation ofvaterite, which includes organic by-products of bacterial activity,and EPS coatings, may pose an additional advantage: these organ-ics will likely reduce the dissolution rate of gypsum, facilitating itsslow resorption. Bacterial consolidation of gypsum may thus helpto overcome the main limitations of this biomaterial and expandits biomedical applications. Extensive in vitro and in vivo testsshould be performed to evaluate the biocompatibility and perfor-mance of bacterially consolidated gypsum biomaterials.

Finally, it should be underlined that DRMS may find applica-tions in the evaluation of the mechanical properties of biomateri-als. As a main advantage, DRMS will enable the evaluation of themechanical properties of biomaterials along depth profiles, animportant endpoint not properly addressed by most standardmechanical testing systems.

5. Conclusion

Although consolidants have been extensively used for over acentury in cultural heritage conservation, their selection is still lar-gely based on empirical considerations without proper determina-tion of the consolidant effectiveness and compatibility with thematerial it is applied to. In the case of the studied archaeologicalgypsum plasters, conventional consolidation treatments using analkoxysilane (TEOS) and organic polymers and copolymers(PEMA/PMA and PVB) at best consolidate only a few millimetersnear the surface. Better results are obtained using a bioconsolida-tion treatment based on the application of M-3P culture mediumto activate the carbonatogenic microbiota present in the plasters.These bacteria produce newly formed calcium carbonate bioce-ment. This biocement is formed by vaterite, which in turn is madeup of oriented inorganic nanounits embedded in an organic matrix.The brick and mortar mesostructure of bacterial vaterite cementprovides a high mechanical resistance. On the other hand, thedepth of consolidation achieved by this biotreatment, which doesnot alter the appearance of the treated material, is noteworthy(up to 2 cm). Our results show that the bacterial biotreatmentwould be applicable for in situ consolidation of archeological gyp-sum plasters. Finally, we would like to point out that such a

bacterial bioconsolidation treatment may help strengthen and re-duce the resorption rate of gypsum biomaterials.

Acknowledgements

This work was financially supported by the Spanish Govern-ment (Grant CGL2012-35992), the Junta de Andalucía (ResearchGroups RNM-179 and BIO-103, and Projects P11-RNM-7550 andP08-RNM-3943) and the Instituto FCICOP de Conservación y Res-tauración de Bienes Culturales (ICON-FCICOP). We thank the Cen-tro de Instrumentación Científica (CIC; University of Granada) forassistance with SEM, TEM and TG/DSC analyses. We thank M.J.Mosquera and L. Pihno for their help with DRMS measurements,A. Hernández Pablos for her help with sample treatments and K.Elert for comments and suggestions.

Appendix A. Figures with essential colour discrimination

Certain figures in this article, particularly Figs. 1, 2, 5, 6 and 8are difficult to interpret in black and white. The full colour imagescan be found in the on-line version, at http://dx.doi.org/10.1016/j.actbio.2014.03.007.

References

[1] Kingery WD, Vandiver PB, Prickett M. The beginnings of pyrotechnology, partII: production and use of lime and gypsum plaster in the pre-pottery neolithicNear East. J Field Archaeol 1988;15:219–44.

[2] Elsen J. Microscopy of historic mortars—a review. Cem Concr Res2006;36:1416–24.

[3] Lucas A, Harris JR. Ancient egyptian materials and industries. 4th ed. NewYork: Dover Publications; 1999.

[4] Agnew N, Maekawa S. Preserving Nefertari’s legacy. Sci Am 1999;281:75–9.[5] Blasco-López FJ, Alejandre Sánchez FJ. Porosity and surface hardness as

indicators of the state of conservation of Mudéjar plasterwork in the RealAlcázar in Seville. J Cult Heritage 2013;14:169–73.

[6] Adams J, Kneller W, Dollimore D. Thermal analysis (TA) of lime and gypsum-based medieval mortars. Thermochim Acta 1992;211:93–106.

[7] Rodriguez-Navarro C. Binders in historical buildings: traditional lime inconservation. In: Herrero JM, Vendrell M, editors. International seminar onarchaeometry and cultural heritage: the contribution ofmineralogy. Bilbao: Sociedad Española de Mineralogía; 2012. p. 91–112.

[8] Ashurst J, Ashurst N. Practical building conservation: mortars, plasters andrenders. English heritage technical handbook, vol. 3. New York: Halsted Press;1988.

[9] Peltier LF, Bickel EY, Lillo R, Thein MS. The use of plaster of Paris to fill defectsin bone. Ann Surg 1957;146:61–9.

[10] Thomas MV, Puleo DA. Calcium sulfate: properties and clinical applications. JBiomed Mater Res B Appl Biomater 2008;88B:597–610.

[11] Bohner M. Design of ceramic-based cements and putties for bone graftsubstitution. Eur Cells Mater 2010;20:1–12.

[12] Lewis KN, Thomas MV, Puleo DA. Mechanical and degradation behavior ofpolymer–calcium sulfate composites. J Mater Sci Mater Med 2006;17:531–7.

[13] Huan Z, Chang J. Self-setting properties and in vitro bioactivity of calciumsulfate hemihydrate–tricalcium silicate composite bone cements. ActaBiomater 2007;3:952–60.

[14] Singh NB, Middendorf B. Calcium sulphate hemihydrate hydration leading togypsum crystallization. Progr Cryst Growth Charact Mater 2007;53:57–77.

[15] Vekinis G, Ashby MF, Beaumont PWR. Plaster of Paris as a model material forbrittle porous materials. J Mater Sci 1993;28:3221–7.

[16] Charola AE, Pühringer J, Steiger M. Gypsum: a review of its role in thedeterioration of building materials. Environ Geol 2007;52:339–52.

[17] Mora P, Mora L, Philippot P. La conservation des peinturesmurales. Bologna: Editrice Compositori; 1999.

[18] Lazzarini L, Tabasso ML. Il Restauro della Pietra. Padova: Cedam; 1986.[19] Doehne E, Price C. Stone conservation: an overview of current research. 2nd

ed. Los Angeles, CA: The Getty Conservation Institute; 2010.[20] Chelazzi D, Poggi G, Jaidar Y, Toccafondi N, Giorgi R, Baglioni P. Hydroxide

nanoparticles for cultural heritage: consolidation and protection of wallpaintings and carbonate materials. J Colloid Interface Sci 2013;392:42–9.

[21] Wheeler G. Alkoxysilanes and the consolidation of stone. Los Angeles, CA: TheGetty Conservation Institute; 2005.

[22] Brinker CF, Scherer GW. Sol-gel science: the physics and chemistry of sol–gelprocessing. San Diego, CA: Academic Press; 1990.

[23] Mosquera MJ, Pozo J, Esquivias L. Stress during drying of two stoneconsolidants applied in monumental conservation. J Sol–Gel Sci Tech2003;26:1227–31.



[24] Daehne A, Herm C. Calcium hydroxide nanosols for the consolidation of porousbuilding materials – results from EU-STONECORE. Heritage Sci 2013;1:11.

[25] Striegel MF, Guin EB, Hallett K, Sandoval D, Swingle R, Knox K, et al. Airpollution, coatings, and cultural resources. Progr Org Coatings 2003;48:281–8.

[26] Le Metayer-Levrel G, Castanier S, Orial G, Loubiere JF, Perthuisot JP.Applications of bacterial carbonatogenesis to the protection andregeneration of limestones in buildings and historic patrimony. SedimentGeol 1999;126:25–34.

[27] Tianno P, Biagiotti L, Mastromei G. Bacterial bio-mediated calcite precipitationfor monumental stones conservation: methods of evaluation. J MicrobiolMethods 1999;36:139–45.

[28] Rodriguez-Navarro C, Rodriguez-Gallego M, Ben Chekroun K, Gonzalez-MuñozMT. Conservation of ornamental stone by Myxococcus xanthus-inducedcarbonate biomineralization. Appl Environ Microbiol 2003;69:2182–93.

[29] Jimenez-Lopez C, Rodriguez-Navarro C, Piñar G, Carrillo-Rosúa FJ, Rodriguez-Gallego M, González-Muñoz MT. Consolidation of degraded ornamentalporous limestone stone by calcium carbonate precipitation induced by themicrobiota inhabiting the stone. Chemosphere 2007;68:1929–36.

[30] Jimenez-Lopez C, Jroundi F, Pascolini C, Rodriguez-Navarro C, Piñar G,Rodriguez-Gallego M, et al. Consolidation of quarry calcarenite by calciumcarbonate precipitation induced by bacteria activated among the microbiotainhabiting the stone. Int Biodeter Biodegrad 2008;62:352–63.

[31] González-Muñoz MT. Bacterial biomineralization applied to the protectionconsolidation of ornamental stone: current development and perspectives.Coalition 2008;15:12–8.

[32] Jroundi F, Fernández-Vivas A, Rodriguez-Navarro C, Bedmar EJ, Gonzalez-Muñoz MT. Bioconservation of deteriorated monumental calcarenite stone andidentification of bacteria with carbonatogenic activity. Microb Ecol2010;60:39–54.

[33] Jroundi F, Gómez-Suaga P, Jimenez-Lopez C, Gonzalez-Muñoz MT, Fernandez-Vivas MA. Stone-isolated carbonatogenic bacteria as inoculants in bioconsolidationtreatments for historical limestone. Sci Total Environ 2012;425:89–98.

[34] Rodriguez-Navarro C, Jimenez-Lopez C, Rodriguez-Navarro A, Gonzalez-Muñoz MT, Rodriguez-Gallego M. Bacterially mediated mineralization ofvaterite. Geochim Cosmochim Acta 2007;71:1197–213.

[35] Webster A, May E. Bioremediation of weathered-building stone surfaces.Trends Biotechnol 2006;24:255–60.

[36] De Muynck W, De Belie N, Verstraete W. Microbial carbonate precipitation inconstruction materials: a review. Ecol Eng 2010;36:118–36.

[37] Gonzalez-Muñoz MT, Rodriguez-Navarro C, Jimenez-Lopez C, Rodriguez-Gallego M. Method and product for protecting and reinforcing constructionand ornamental materials. Patent WO 2008/009771 A1, 2008.

[38] Anne S, Rozenbaum O, Andreazza P, Rouet JL. Evidence of a bacterial carbonatecoating on plaster samples subjected to the calcite bioconceptbiomineralization technique. Constr. Building Mater 2010;24:1036–42.

[39] Rodriguez-Navarro C, Jroundi F, Schiro M, Ruiz-Agudo E, Gonzalez-Muñoz MT.Influence of substrate mineralogy on bacterial mineralization of calciumcarbonate: implications for stone conservation. Appl Environ Microbiol2012;78:4017–29.

[40] De Muynck W, Debrouwer D, De Belie N, Verstraete W. Bacterial carbonateprecipitation improves the durability of cementitious materials. Cem ConcrRes 2008;38:1005–14.

[41] Kamat S, Su X, Ballarini R, Heuer AH. Structural basis for the fracture toughnessof the shell of the conch Strombus gigas. Nature 2000;405:1036–40.

[42] Fleischli FD, Dietiker M, Borgia C, Spolenak R. The influence of internal lengthscales on mechanical properties in natural nanocomposites: a comparativestudy on inner layers of seashells. Acta Biomater 2008;4:1694–706.

[43] Meyers MA, Chen PY, Lin AYM, Seki Y. Biological materials: structure andmechanical properties. Progr Mater Sci 2008;53:1–206.

[44] Kunitake ME, Mangano LM, Peloquin JM, Baker SP, Estroff LA. Evaluation ofstrengthening mechanisms in calcite single crystals from mollusk shells. ActaBiomater 2013;9:5353–9.

[45] Yang W, Zhang GP, Zhu XF, Li XW, Meyers MA. Structure and mechanicalproperties of Saxidomus purpuratus biological shells. J Mech Behav BiomedMater 2011;4:1514–30.

[46] Li X, Chang WC, Chao YJ, Wang R, Chang M. Nanoscale structural andmechanical characterization of a natural nanocomposite material: the shell ofred abalone. Nanoletters 2004;4:613–7.

[47] Rodriguez-Navarro AB, Checa A, Willinger MG, Bolmaro R, Bonarski J.Crystallographic relationships in the crossed lamellar microstructure of theshell of the gastropod Conus marmoreus. Acta Biomater 2012;8:830–5.

[48] Tiano P, Delgad Rodrigues J, De Witte E, Vergés-Belmin V, Massey S, SnethlageR, et al. The conservation of monuments: a new method to evaluateconsolidating treatments. Int J Restor Buildings Monum 2000;6:133–50.

[49] Tiano P, Cantisani E, Sutherland I, Paget JM. Biomediated reinforcement ofweathered calcareous stones. J Cult Heritage 2006;7:49–55.

[50] Delgado Rodrigues J, Pinto AF, da Costa DR. Tracing of decay profiles andevaluation of stone treatments by means of microdrilling techniques. J CultHeritage 2002;3:117–25.

[51] Pamplona M, Kocher M, Snethlage R, Aires Barros L. Drilling resistance:overview and outlook. Z dt Ges Geowiss 2007;158:665–76.

[52] Ferreira Pinto AP, Delgado Rodrigues J. Stone consolidation: the role oftreatment procedures. J Cult Herit 2008;9:38–53.

[53] Ferreira Pinto AP, Delgado Rodrigues J. Consolidation of carbonate stones:Influence of treatment procedures on the strengthening action of consolidants.J Cult Herit 2012;13:154–66.

[54] Fratini F, Rescic S, Tiano P. A new portable system for determining the state ofconservation of monumental stones. Mater Struct 2006;39:139–47.

[55] Stavropoulou M. Modeling of small-diameter rotary drilling tests on marbles.Int J Rock Mech Mining Sci 2006;43:1034–51.

[56] Benavente D, Martinez-Verdú F, Bernabeu A, Vigueira V, Fort R, García del CuraMA, et al. Influence of surface roughness on color changes in building stones.Color Res Appl 2003;28:343–51.

[57] Warren BE. X-ray diffraction. Reading, MA: Addison-Wesley; 1969.[58] Cullity BD, Stock SR. Elements of X-ray diffraction. 3rd ed. New York: Prentice

Hall; 2001.[59] Williamson GK, Hall WH. X-ray line broadening from filed aluminium and

wolfram. Acta Metal 1953;1:22–31.[60] Martin Ramos JD. XPowder: a software package for powder X-ray diffraction

analysis, Version 2004.03, 2004 [www.xpowder.com].[61] Phani KK, Niyogi SK, Maitra AK, Roychaudhury M. Strength and elastic

modulus of a porous brittle solid: an acousto-ultrasonic study. J Mater Sci1986;21:4335–41.

[62] Sasse HR, Snethlage R. Methods for the evaluation of stone conservationtreatments. In: Baer NS, Snethlage R, editors. Saving our architectural heritage:the conservation of historic stone structures. New York: John Wiley & Sons;1997. p. 223–43.

[63] Strydom CA, Hudson-Lamb DL, Potgieter JH, Dagg E. The thermal dehydrationof synthetic gypsum. Thermochim Acta 1995;269–270:631–8.

[64] Grassmann O, Müller G, Löbmann P. Organic–inorganic hybrid structure ofcalcite crystalline assemblies grown in a gelatin hydrogel matrix: relevance tobiomineralization. Chem Mater 2002;14:4530–5.

[65] Wang JW, Becker U. Structure and carbonate orientation of vaterite (CaCO3).Am Miner 2009;94:380–6.

[66] Vaz MF, Pires J, Carvalho AP. Effect of the impregnation treatment withparaloid B-72 on the properties of old Portuguese ceramic tiles. J Cult Herit2008;9:269–76.

[67] Scholl MA, Mills AL, Herman JS, Hornberger GM. The influence of mineralogyand solution chemistry on the attachment of bacteria to representative aquiferminerals. J Contamin Hydrol 1990;6:321–36.

[68] Fernández-Díaz L, Pina CM, Astilleros JM, Sánchez-Pastor N. The carbonation ofgypsum: pathways and pseudomorph formation. Am Miner 2009;94:1223–34.

[69] Fernández-Díaz L, Fernández-González A, Prieto M. The role of sulfate groupsin controlling CaCO3 polymorphism. Geochim Cosmochim Acta2010;74:6064–76.

[70] Al-Naddaf M, Wakid F, Abu Alhassan Y. Micro-drilling resistancemeasurement: a new technique to estimate the porosity of a building stone.Mediterranean Archaeol Archaeomet 2013;13:225–33.

[71] Illescas J. Nuevos materiales para la conservación del patrimonio monumentalAndaluz. PhD thesis, University of Cadiz, 2012.

[72] Theodoridou M. Physical and mineralogical changes of Hungarianmonumental stones exposed to different conditions: stone-testing in-situand under laboratory conditions. PhD thesis, University of Bologne, 2009.

[73] Theodoridou M, Ioannou I, Philokyprou M. New evidence of the early use ofartificial pozzolanic material in mortars. J Archaeol Sci 2013;40:3263–9.

[74] Cölfen H, Antonietti M. Mesocrystals: inorganic superstructures made byhighly parallel crystallization and controlled alignment. Angew Chem Int Ed2005;44:5576–91.

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

[76] Barthelat F. Nacre from mollusk shells: a model for high-performancestructural materials. Bioinsp Biomin 2010;5:035001.

[77] Oaki Y, Imai H. Nanoengineering in echinoderms: the emergence ofmorphology from nanobricks. Small 2006;2:66–70.

[78] Seto J, Ma Y, Davis SA, Meldrum F, Gourrier A, Kim YY, et al. Structure–propertyrelationships of a biological mesocrystal in the adult sea urchin spine. ProcNatl Acad Sci USA 2012;109:3699–704.

[79] Aizenberg J, Weaver JC, Thanawala MS, Sundar V, Morse DE, Fratzl P. Skeletonof Euplectella sp.: structural hierarchy from the nanoscale to the macroscale.Science 2005;309:275–8.

[80] Gao H, Ji B, Jäger IL, Artz E, Fratzl P. Materials become insensitive to flaws atnanoscale: lessons from nature. Proc Natl Acad Sci USA 2003;100:5597–600.

[81] Sekkal W, Zaoui A. Nanoscale analysis of the morphology and surface stabilityof calcium carbonate polymorphs. Sci Rep 2013;3:1587.

[82] Ren D, Meyers MA, Zhou B, Feng Q. Comparative study of carp otolithhardness: lapillus and asteriscus. Mater Sci Eng C 2013;33:1876–81.

[83] Lüttge A, Conrad PG. Direct observation of microbial inhibition of calcitedissolution. Appl Environ Microbiol 2004;70:1627–32.


author's personal copy · f. jroundi et al./acta biomaterialia 10 (2014) 3844 3854 3845....

Documents