Investigation of the Composition of Historical and Modern Italian Papers by Energy Dispersive X-ray Fluorescence (EDXRF), X-ray Diffraction (XRD), and Scanning Electron Microscopy

Investigation of the Composition of Historical and ModernItalian Papers by Energy Dispersive X-ray Fluorescence(EDXRF), X-ray Diffraction (XRD), and Scanning ElectronMicroscopy Energy Dispersive Spectrometry (SEM-EDS)

MARTA MANSO, MARIA LUISA CARVALHO,* IGNACIO QUERALT, SILVIA VICINI,and ELISABETTA PRINCICentro de Fısica Atomica da Universidade de Lisboa, Av. Professor Gama Pinto, 2, 1649-003 Lisboa, Portugal (M.M., M.L.C.); Laboratory of

X-ray Analytical Applications, Institute of Earth Sciences ‘‘Jaume Almera’’, CSIC, Sole Sabaris s/n. 08028 Barcelona Spain (I.Q.); and

Dipartimento di Chimica e Chimica Industriale, University of Genova, Via Dodecaneso 31, 16146, Genova, Italy (S.V., E.P.)

In this work, a study concerning the composition of Italian papers from

the seventeenth to the twentieth centuries was carried out using energy

dispersive X-ray fluorescence spectrometry (EDXRF), X-ray diffraction

(XRD), and scanning electron microscopy coupled to energy dispersive

spectrometry (SEM-EDS). The analyzed samples consisted of papers

employed for drawing, writing, printing, and absorbance. Observations

carried out by SEM magnified the typical paper morphology. EDXRF in

combination with XRD and SEM-EDS allowed the determination of

calcite, gypsum, kaolin, talc, magnesite, and dolomite, used as fillers in the

production of the papers studied herein. The inks present in the

handwritten and printed papers, investigated by SEM-EDS and l-

EDXRF, were synthetic, Fe based, and iron gall inks.

Index Headings: Art conservation; Paper; Ink; Energy dispersive X-ray

fluorescence; EDXRF; X-ray diffraction; XRD; Scanning electron

microscopy energy dispersive spectrometry; SEM-EDS.

INTRODUCTION

For many centuries paper was the only material for recordingcultural and historical information all over the world. The art ofpapermaking started in the first century B.C. in China; in theeighth century it arrived in Samarqand and Baghdad, spreadingto Europe three centuries later. The manufacture of paper inEurope was first established in Islamic Spain in the middle ofthe eleventh century. In the second half of the fourteenthcentury the use of paper had become well accepted in all ofwestern Europe.1–4

Prior to 1850 the raw materials used to produce paper werehemp and linen rags composed only of cellulose fibers. Duringpapermaking, cellulose fibers rapidly accumulated metalsdissolved in the processing water and originated in thepapermaking machines. In order to obtain a writable sheet,sizing compounds were also added, such as gelatine (animalglue).4 Since the seventeenth century, papermakers have addedalum (Al2(SO4)3�18H2O) to the gelatine not only to enhancepaper strength, but also to protect it from micro organic attack.Indeed, alum controls the growth of bacteria and mold;unfortunately, it also induces the formation of sulfuric acid,which promotes acid hydrolysis in paper.5

With the increasing demand for paper and the shortage ofcellulose fibers from flax and cotton, wood became the main

source of cellulose.4 Hence, by the end of the nineteenthcentury, paper was composed of cellulose, hemicellulose, andlignin, originating from pulpwood. Moreover, non-fibrouscomponents including various coloring agents, sizing, fillers,coatings, and synthetic inks began to be employed inpapermaking.6–8 In the same period, gelatine, formerly usedfor sizing, was replaced by rosin,5 always applied with alum,and more recently by other synthetic products, such asalkylketene dimer and carboxy methyl cellulose (CMC).

It is evident how the evolution of papermaking throughoutthe centuries has contributed to a modification in papercomposition, although the principles remained the same.

The characterization of paper components plays an impor-tant role in the identification of paper documents; their identityand their source are not only important for the history ofhumanity but also in forensic science concerning judicialproblems encountered in forgery, felonious use of certificates,securities, wills, etc.9,10

Several analytical techniques can be used to identify fillers,sizing materials, and inks present in handwritten or printeddocuments.11,12 Among them, energy dispersive X-ray fluo-rescence (EDXRF) spectrometry is the most important for thequalitative and quantitative analysis of the elements in papersamples.13–15 These results can be supported by those resultsderived from energy dispersive spectrometry (EDS), whichallows determination of both the elemental composition and thenature of fillers and inks present in particular areas of samplesby means of coupling with scanning electron microscopy(SEM).16 Furthermore, the observations carried out by SEMmagnify the typical paper morphology, highlighting the effectsof weathering on the cellulose matrix.17,18 X-ray diffraction(XRD) is an important experimental technique for phaseidentification and determination of crystalline structures, interms of geometry and crystalline system parameters. Regard-ing paper, XRD can provide information about the crystallinitydegree of cellulose, related to its degradation level, and canconfirm the nature of fillers detected therein.19–21

In this work, a study concerning the inorganic compositionof some historical (seventeenth to nineteenth centuries) andmodern (twentieth century) Italian papers was carried out usingEDXRF, XRD, and SEM-EDS. Inks present in some of thesepapers were also investigated by SEM-EDS and l-EDXRF. Inthis way a micro-destructive investigation of paper compositionwas performed, whose results could be helpful for conservationpurposes. Indeed, a deep knowledge of paper artworks in termsof the main constitutive elements (i.e., cellulose, fillers, inks)

Received 25 August 2010; accepted 12 October 2010.

* Author towhom correspondence should be sent. E-mail: [email protected].

DOI: 10.1366/10-06105

52 Volume 65, Number 1, 2011 APPLIED SPECTROSCOPY0003-7028/11/6501-0052$2.00/0

� 2011 Society for Applied Spectroscopy

can support the choice of the best methods of conservationintervention. For example, to have information on thecomposition of the inks present may support the selection ofthe best restoration method that also avoids any worsening ofthe state of the paper or any variation in its formal identity.

It is worth noting that to completely characterize a paperartwork, its conservative state should also be assessed byapplying other analytical techniques, thus confirming the needfor a multi-analytical approach in paper conservation. Fromthis point of view, along with the information coming from theelemental analysis and the morphological observation, theknowledge of the molecular characteristics of the paper and inkbecomes essential. For example, Raman, infrared, and nuclearmagnetic resonance (NMR) spectroscopy give valuableinformation on the degradation level of cellulose.

EXPERIMENTAL

Sample Description. The selected papers, summarized inTable I, have different provenances and ages. They are mostlyfrom two well-known sites of papermaking in Italy, the LiguriaRegion (northern Italy) and Fabriano (central Italy).4 Papersamples from 14 to 17 were produced in Italy, but the exactregion of provenance is unknown.

Samples 1 through 12 do not contain any ink; samples 13through 18 were handwritten, printed, or have ink spots. Eventhough the papers were stored in libraries, they presentedevident traces of photo-degradation and oxidation due both tothe additives used in the production (inks, sizing, fillers, etc.)and to the storage conditions.

Energy Dispersive X-ray Fluorescence Measurements.The elemental concentration of samples from 1 to 18 wasobtained by EDXRF analysis. The spectrometer used for thispurpose consisted of an X-ray tube (PW 1140, 100 kV, 80 mA)equipped with a secondary target of Mo. The X-ray tube, thesecondary target, and the sample are in a triaxial geometry.22

The characteristic radiation emitted by the elements present inthe sample was detected by means of a Si(Li) detector. Theenergy resolution is 138 eV at 5.9 keV and the acquisition

system is a Nucleus PCA card. The quantitative evaluation wasmade by the fundamental parameters method.23,24 The X-raygenerator was operated at 50 kV and 20 mA, and theacquisition time was 1000 s. The accuracy was checked byanalysis of reference materials: orchard leaves, NBS standardreference material 1571, the matrix of which is cellulose.25 Allpapers were directly measured by EDXRF without any kind ofsample preparation for elemental characterization.

Micro-Energy Dispersive X-ray Fluorescence Measure-ments. Some samples were handwritten (13), printed (14 to17), or carried ink traces (18). In order to better distinguishbetween the ink components and the cellulose matrix, a systemwith a high lateral resolution was necessary. For this purpose, al-EDXRF spectrometer was used. This equipment consisted ofan OXFORD X-Ray generator (Model XTF 5011, 50 kV, 1mA). The characteristic radiation emitted by the elementspresent in the sample was detected by means of a VORTEX-EX SDD Si(Li) detector. The geometry between the X-ray tubeand detector was 458. The sample was positioned at the focalpoint of two laser beams. The focusing spot size was 100 lmobtained by a polycapillary lens. The spectra acquisitions wereobtained by the PI-SpecA software. The detector and X-raytube are coupled to a vacuum chamber (10 mbar). The X-raygenerator was operated at 40 kV and 0.5 mA, and anacquisition time of 500 s was used. All papers and inks weredirectly measured by l-EDXRF, without any kind of samplepreparation for elemental characterization.

Energy Dispersive X-ray Fluorescence Mapping. Sample18 consisted of a blotting paper containing several traces ofinks. In order to identify the inks, EDXRF chemical mappingwas carried out using an EDX spectrometer (XDV-SD model,Helmut Fischer GmbH, Sindelfingen, Germany). It consists ofa micro-focus tungsten anode X-ray tube operating at fixedvoltage of 50 kV and 1 mA of tube current, and a Si–Pinsemiconductor detector (Peltier cooling at �50 8C; energyresolution 180 eV). A color video microscope allowed theselection and viewing of the irradiated area with up to 453magnification. The impinging X-ray beam reaches the samplesurface at 908, and the detection system collecting the emergingsecondary X-rays from the sample is placed at 458. Theinstrument is controlled by the WinFTM software, which isalso used both for spectra acquisition and for spectral datatreatment. Quantitative results were obtained using the standardfree fundamental parameters method.26

X-ray Diffraction Measurements. X-ray diffraction anal-yses were carried out at a controlled temperature of 25 8C,using a diffractometer Bruker, type D8-ADVANCE with CuKa radiation (40 kV and 40 mA). The incident radiation passedthrough a nickel filter and was collimated by a Gobel mirror,providing a high flux of photons to the sample surface. The X-rays emerging from the sample were collected by a solid-statedetector (Sol-X, Bruker/AXS, Germany) with a detectionefficiency of 98% for the Cu Ka radiation. The use of such adetector avoided intensity losses associated with othersecondary filters or crystals in the beam path, reducing thecounting time and improving the peak-to-background ratio.This configuration resulted in a very good performance for theanalysis of samples with low X-ray absorption, for thin films,for not very flat surfaces, and also for small samples. All paperswere directly measured without any kind of sample prepara-tion. The XRD spectra of papers were recorded using the h/2hgeometry, collecting the data in the range (2h) between 48 and

TABLE I. List of the analyzed paper samples.

Paper samplePapergrade

Regionof origin Age

1 (sheet) Drawing Italy (Liguria) XVIII century2 (sheet) Drawing Italy (Liguria) Beginning XIX

century3 (sheet) Drawing Italy (Liguria) XVII century4 (sheet) Drawing Italy (Fabriano) XIX century5 (sheet) Drawing Italy (Fabriano) XIX century6 (sheet) Drawing Italy (Fabriano) XIX century7 (sheet) Drawing Italy (Fabriano) XVIII century8 (sheet) Writing Italy (Fabriano) 1760–17689 (sheet) Writing Italy (Fabriano) 1860–186410 (sheet) Writing Italy (Fabriano) 1815–182711 (sheet) Writing Italy (Fabriano) 1890–190012 (sheet) Writing Italy (Fabriano) 1805–181013 (manuscript) Writing

(handwritten)Italy (Liguria) 1815–1827

14 (newspaper) Printing Italy 193315 (newspaper) Printing Italy 194916 (magazine, picture) Printing Italy 194017 (magazine, printing) Printing Italy 194018 (blotting paper) Absorbency

(ink spots)Italy (Liguria) 1815

APPLIED SPECTROSCOPY 53

608 with a step scan of 0.058, at a rate of 1 s per step. Theevaluation of the spectra was made using the Diffrac.SuiteTM

software and the identification of the inorganic chemicalcompounds (when they were present) was made by use of theInternational Centre for Diffraction Data Powder DiffractionFiles database (ICDD PDF).

For determining the percentage of crystallinity of thecellulose matrix a simple approach was applied.27 It consistsof taking from the diffractogram a suitable maximum and aminimum intensity to give the ‘‘crystallinity degree’’ (CrI),defined as:

CrI ¼ I002 � Iam

I002

100

where I002 is the intensity of the crystalline peak at themaximum at 2h between 228 and 238 for cellulose I (between188 and 228 for cellulose II) and Iam is the intensity of theamorphous reflection at the minimum at 2h between 188 and198 for cellulose I (between 138 and 158 for cellulose II).

Scanning Electron Microscopy Energy Dispersive Spec-trometry Measurements. Morphological observations werecarried out directly on the paper samples by SEM associatedwith an EDS microprobe without any kind of samplepreparation. A scanning electron microscope Stereoscan 440Leica-Cambridge associated with an EDS microprobe Link-Gun Oxford have been used following metallization of thespecimen with a very thin layer of graphite to obtain a goodconductivity. SEM images were recorded using both asecondary electron (SEI) as well as a back scattering (BSE)detector at different magnifications.

RESULTS AND DISCUSSION

Cellulose Fibers. No substantial degradation of thecellulose fibers was observed in the selected papers by meansof SEM. In general, no broken fibers were noticed, asillustrated by Fig. 1. Moreover, the latter resembles the typical

morphology of a paper in good condition. A superficial coatingpresumably made by an organic compound, thus not revealedby EDS nor by EDXRF, was observed by SEM on samples 1,3, and 5 (Fig. 1b). This coating could be a superficial layeradded for drawing purposes.

Paper samples were analyzed by X-ray diffraction toevaluate the crystallinity degree of the cellulose matrix, whichcan be related to the degradation level of cellulose. It is wellknown that the degradation pathways (i.e., acid hydrolysis,oxidation, photo-degradation, etc.) first occur in the amorphousphase of cellulose, and only in a second stage is the crystallinephase attacked.28 Therefore, values of CrI lower than thosetypical of cellulose (more than 80%) indicate a degradedmaterial or a paper made of low-quality cellulose fibers, such asthose coming from wood.

The CrI data corresponding to the analyzed papers arecollected in Table II. Most samples from the seventeenth tonineteenth centuries were characterized by high values of CrI,indicating that they did not suffer degradation to a large extent.Again, these data confirm that the samples were made of high-quality paper, presumably coming from white cotton rags. Onthe contrary, the inked samples were characterized by lowervalues of CrI (less than 70%). This issue is consistent with theirage and their use (newspapers and magazines of the first half ofthe twentieth century) for which low-quality raw materials,such as pulpwood, were usually employed. In addition, thelignin could catalyze some degradation processes, determininga further reduction of the crystalline phase.

Fillers. Energy dispersive X-ray fluorescence analysis wasused to obtain information on the elemental composition of thestudied papers. In Fig. 2 the mean elemental concentration andstandard deviation (lg g�1) of the analyzed papers is presentedtaking into account the different paper grades. The highestconcentrations were observed for K, Ca, and Fe. Copper, Zn,and Pb were also identified as trace elements.

The presence of Ca was not surprising since this elementplays a role in several steps of the papermaking process.

FIG. 1. SEM images of: (a) sample 2 (SEI); (b) sample 5 (SEI); (c) sample 10 (BSE); (d) sample 16 (BSE); (e) sample 17 (SEI); and (f) sample 13 (SEI).

54 Volume 65, Number 1, 2011

Indeed, Ca in paper may come from the processing watersupply, bones and other animal parts used in the gelatine size,from the lime [Ca(OH)2] used to disintegrate the rags (usuallyknown as lyeing), or from the additives (fillers, coatings, andbleaches).6,7,29

The presence of K can be related to the paper production,when linen and cotton rags are used as raw materials. After a

vigorous shaking to eliminate dust and dirt, the rags werechopped into small pieces, then mixed with water and lye toeliminate the fatty impurities. This lye could contain lime, soda(Na2CO3), caustic soda (NaOH), or caustic potash (KOH).5

Some residuals of lye could remain entrapped in the cellulosepulp, allowing the detection of K in some papers (samples 1, 3,5, 8, 11–14, 16, and 17). It is interesting to note that both in theremaining historical and modern samples K is absent. In thefirst case, the papers produced from rags, the lye probablycontained no caustic potash, whereas the modern papers weremade of pulpwood.4 Unfortunately, EDXRF cannot give anyinformation on the lye composition, since Na is not detectableby the equipment used.

Iron was found both in the ‘‘blank’’ (not inked) and inkedpapers. The presence of Fe in the ‘‘blank’’ papers could beassociated with the processing water supply as well as withthe papermaking equipment.7 The highest concentrations ofFe were obtained in the case of printed papers (from samples14 to 17) presumably due to the presence of the ink. Similarresults were obtained in Refs. 30 and 31 for analogous papergrades.

Generally, Cu and Zn levels in the ‘‘blank’’ samples rangedbetween 15 and 70 lg g�1 and it was probably picked up fromthe processing water or the equipment. Zinc occurrence hasalready been reported in papers from the eighteenth century29

as possibly coming from white zinc oxide, a filler that becamemore frequent in the early twentieth century.32

TABLE II. Values of CrI obtained by XRD for the papers under study.

Papersample CrI Age

1 89 XVIII century2 85 Beginning XIX century3 86 XVII century4 88 XIX century5 83 XIX century6 87 XIX century7 88 XVIII century8 88 1760–17689 88 1860–186410 88 1815–182711 82 1890–190012 87 1805–181013 88 1815–182714 66 193315 70 194916 65 194017 68 194018 83 1815

FIG. 2. Mean elemental concentration of the papers for: (a) drawing; (b) writing; (c) printing; and (d) blotting paper.

APPLIED SPECTROSCOPY 55

Lead was found in almost all the analyzed papers, probablypicked up from the processing equipment.29

In order to completely identify the nature of fillers, the papersamples were analyzed by SEM-EDS and XRD as well.

Indeed, the X-ray diffractogram of some samples showednot only the characteristic peaks of crystalline cellulose at 2hbetween 228 and 238 and at 2h between 158 and 168, but alsothose corresponding to the fillers. Particularly, the character-istic peaks of calcite (CaCO3) are at 2h ¼ 308 and 368; forkaolin [Al2Si2O5(OH)4] at 2h¼ 128 to 138 and 258, and for talc[Mg3Si4O10(OH)2] at 2h¼ 108, 18–208, and 288. As shown inFig. 3, the XRD diffractogram of sample 16 highlights thepresence of kaolin, calcite, and talc.

The mineral composition of the samples, as well as thetechniques that allowed their identification, are summarized inTable III.

The oldest paper investigated herein contained a smallamount of filler (Fig. 1c) with respect to the usual commercial-grade paper.16,17 Indeed, in modern paper characterized as lowgrade, such as newsprint paper, the amount of cellulose contentis reduced to decrease their cost, whereas the content of the lessexpensive inorganic fillers increases.

In paper samples 1, 2 (Fig. 1a), and 7 no fillers or otherinorganic additives are present, as revealed by SEM-EDS andXRD. This is consistent with their age; indeed in the oldestsamples, especially those made of rags, no other componentswere added to the cellulose pulp, except the sizing.4

The chemical composition of calcium-based fillers wasidentified by XRD and/or SEM-EDS in the form of calcite(samples 3–6, 8–13, and 16–18), gypsum (CaSO4) (samples 9and 11–13), and dolomite CaMg(CO3)2 (samples 15 and 16).The most frequent filler was calcite. Used as filler in thealkaline paper manufacturing process, calcite improves severalimportant paper characteristics, such as smoothness, bright-ness, opacity, and affinity for ink; it also reduces paper acidity.

The presence of other mineral fillers such as kaolin (samples4, 6, and 8–18), talc (samples 14–17), barium sulfate

(sample16), and magnesite (MgCO3) (samples 14 and 15)was also detected by XRD and/or SEM-EDS.

Combining the results coming from EDXRF, EDS, andXRD, it is evident that the main mineral fillers found in mostpapers consisted of calcite and kaolin; generally the modernpapers (twentieth century) presented the widest variety of fillersin their composition.

Inks. The inked samples were analyzed by l-EDXRF andSEM-EDS.

The similarity between the spectra obtained for ink andpaper is evidenced by paper sample 14 in Fig. 4. Since thespectrum obtained for the ink reveals no other elements ordifferent line intensities, we may conclude that it is mostcertainly an organic dye. A similar result was obtained for

FIG. 3. XRD spectrum of sample 16.

TABLE III List of the mineral composition of the papers under studydetected by SEM-EDS (6) and by XRD (�).

Papersample

Mineral composition

Calcite Gypsum Kaolin TalcBariumsulfate Magnesite Dolomite

123 64 �6 65 66 �6 678 �6 69 �6 �6 6

10 �6 611 6 �6 612 6 6 613 �6 614 � � 6 �6 615 6 � 616 �6 6 �6 �17 �6 6 �6 6 �18 �6 6

56 Volume 65, Number 1, 2011

paper sample 15. This evidence is consistent with the age andthe use of the papers under study; they came from two differentnewspapers belonging to the first half of the twentieth century,when in low-quality newsprint paper no expensive inks such asiron gall were employed.4,5 It is well known that from thebeginning of the twentieth century synthetic dyes and colorantswere mainly used for most applications.

Completely different was the composition of the ink comingfrom paper samples 16 and 17, both concerning the printingand the images. As shown in Fig. 1d, the cellulose fibers incorrespondence with the picture in sample 15 were covered bylittle spheres in which EDS analysis detected the presence ofFe. A similar result was obtained for sample 16; as shown inFig. 1e, in the printed area the cellulose matrix and the fillersare completely covered by the ink, in which EDS detected Fe.The presence of Fe, together with Ca (detected by EDS andEDXRF), suggested that the ink could be a black ink based oniron oxide and calcium compounds, produced and employed inthe first half of the twentieth century (an example is the ink‘‘Noille Taille Douce’’ produced by Charbonnel, France).33

In sample 13 the ink formed a uniform coating of the papersurface that hampered the observation of the cellulose matrix(Fig. 1f). Both l-EDXRF and EDS detected the presence of Fein the inked areas; EDS also revealed the presence of S. Thissuggested that the ink used for the writing was an iron gall ink,made by mixing iron vitriol (FeSO4) with a gallnut extract andArabic gum (to keep the black pigment suspended in theliquid). The indelible black iron galet complex was created byoxidation on exposure to air.34 The iron gall ink was already inuse in 1000 B.C. and it was substituted with synthetic dyes inthe beginning of the twentieth century.

Sample 18 showed many different traces of inks. Taking intoaccount that it is a blotting paper, it was probably used many

times on different occasions. A mapping of the concentrationlevels in mg cm�2 for Fe, Cu, and Zn obtained by EDXRFanalysis of this sample revealed the presence of at least twotypes of inks (Fig. 5). Both inks have Fe in common, but onlyink a contains Cu and Zn. EDS analysis of the inks confirmedthe presence of Fe, together with S and K (Fig. 6), suggestingthat both were iron gall inks. An overall comparison of Fe, Cu,and Zn concentration led to the confirmation that the two irongall inks had different composition, even if, unfortunately, itwas not possible to assign it exactly, since many recipes wereavailable in the past, essentially depending on the provenanceand the manufacturer.34 Because of the natural raw materials,the composition of the inks has variable amounts of the majoringredients, additives, and eventual contaminants.

Concerning the presence of Cu and Zn, some interestingconsiderations can be taken into account. In some recipes ironsulfate and copper sulfate (CuSO4) were used interchangeably,since the natural sources of the minerals were usually minedtogether.34 Cechak et al. state that copper was markedly presentin inks of the fourteenth and fifteenth century; its concentrationprogressively decreased in the sixteenth century and from thesecond half of the seventeenth century onward it is present onlyin low concentrations or is not present at all.35 The Zn contentdisplays similar features; Zn is preferably present in inks fromthe fourteenth century to the beginning of the seventeenthcentury.35

Most iron gall inks were made in water and hence, theirpurity varied widely; the oldest recipes indicate the use of rainwater. However, the water may be contaminated with somesalts, which justifies the presence of K and Ca in the inks.These elements could also come from the wine, beer, orvinegar sometimes used instead of water to prevent the inkfrom freezing in winter.34

FIG. 4. Spectra of ink and paper obtained by l-EDXRF for sample 14.

APPLIED SPECTROSCOPY 57

CONCLUSIONS

This work comprised the use of EDXRF, XRD, and SEM-EDS techniques for the study of the inorganic components inpapers from the seventeenth to the twentieth centuries,including the inks present on some of these paper samples.The combination of the three techniques allowed complemen-tary results to be obtained for the characterization of the papercomposition (cellulose matrix and fillers). It is interesting tonote that some compounds were detected only by sometechniques, whereas others, such as, for example, the calcite,were easily identified in any case, due essentially to their highconcentration in the samples. The results obtained by EDXRFand EDS not only pointed out the differences in the elementalcomposition of the paper samples but also facilitated theanalysis of the diffractograms and SEM micrographs.

In our study we used some invasive techniques such as SEM-EDS and XRD that require micro-sampling, but the informationgained using these techniques is essential to support theEDXRF and l-EDXRF results. It is evident that a multi-analytical approach is essential, because even being obvious,not all techniques can be applied for all purposes. Again, inmany cases the analytical approach cannot distinguish betweennondestructive techniques and methods requiring sampling,since at this moment no interchangeable methodologies exist.This situation has been partially overlapped thanks to thedevelopment of expensive portable instrumentation.

Scanning electron microscopy observations revealed that thepapers under study were in a good state of conservation in what

concerns the degradation of cellulose fiber. Furthermore, fromthe percentage of crystallinity of the cellulose matrix weconcluded that printing papers were not as well preserved asthe remaining paper grades.

The collected results revealed very low amounts of fillers inthe historical papers (before the twentieth century), essentiallycalcite and kaolin. This is an indication that the quality of these

FIG. 5. Chemical mapping obtained by EDXRF analysis (a) from paper 18 of: (b) Fe; (c) Cu; and (d) Zn (mg�cm�2).

FIG. 6. EDS spectrum of sample 18 corresponding to ink b.

58 Volume 65, Number 1, 2011

papers made of rags was high, since there was no need to fillout the spaces between fibers. These paper grades have a betterquality with respect to the ones made of pulp wood; thus, todaythey present fewer problems of conservation.

The use of fillers in papermaking may have unwanted effectsrelated to the loss of bonding between fibers. The modernpapers (twentieth century) came from newspapers andmagazines; thus their low quality was justified by the shortdurability required of these objects. They were made ofpulpwood and contained new fillers, such as dolomite, talc, andmagnesite, along with the usual ones (calcite and kaolin). Theuse of barium sulfate is curious in the magazine sample(sample 16), where it worked essentially as a pigment,imparting a white color to the sheets.

Iron-based inks were identified essentially by SEM-EDS andEDXRF techniques. Particularly, the use of a l-EDXRFcoupled to a vacuum chamber spectrometer was essential forink identification, not only because of its spot size but also forits ability to detect lighter elements. In addition, it is worthmentioning the suitability of l-EDXRF chemical mappinganalysis in the identification of different inks present in thesame sample.

ACKNOWLEDGMENTS

This work was partially funded by the European Project PAPERTECH(INCO-CT-2004-509095) and FCT projects POCTI/0303/2003 financed by theEuropean Community Fund FEDER; SFRH/BD/30259/2006; PTDC/EAT/71998/2006. The authors would like to thank Prof. Enrico Pedemonte for thesupport, the suggestions, and the fruitful discussions.

1. D. Hunter, Papermaking – The History and Technique of An Ancient Craft(Dover Publications, New York, 1978).

2. T. Collings and D. Milner, Paper Conserv. 14, 58 (1990).3. B. Rudin and R. G. Tanner, Making Paper: A Look into the History of an

Ancient Craft (Lyons Press, Sweden, 1990).4. E. Pedmonte, E. Princi, and S. Vicini, La Chimica e l’Industria 8, 62

(2005).5. J. Kolar and M. Strlic, Iron Gall Inks: On Manufacture, Characterisation,

Degradation and Stabilisation (National and University Library, Slovenia,2006).

6. K. Beazley, Paper Conserv. 15, 17 (1991).7. P. Reynard, Paper Conserv. 13, 7 (1989).8. D. Reyden, E. Mosier, and M. Bakerm, Pigment-Coated Papers I: History

and Technology (ICOM Committee for Conservation, 10th TriennialMeeting, Washington, D.C., 1993), p. 491.

9. P. J. Simon, B. C. Giessen, and T. R. Copeland, Anal. Chem. 49, 2285(1977).

10. L. D. Spence, A. T. Baker, and P. Byrne, J. Anal. At. Spectrom. 15, 813(2000).

11. K. Castro, S. Pessanha, N. Proietti, E. Princi, D. Capitani, M. L. Carvalho,and J. M. Madariaga, Anal. Bioanal. Chem. 391, 433 (2008).

12. K. Castro, N. Proietti, E. Princi, S. Pessanha, M. L. Carvalho, S. Vicini, D.Capitani, and J. M. Madariaga, Anal. Chim. Acta 623, 187 (2008).

13. M. Mantler and M. Schreiner, X-ray Spectrom. 29, 3 (2000).14. B. Wehling, P. Vandenabeele, L. Moens, R. Klockenkamper, A. von

Bohlen, G. Van Hooydonk, and M. de Reu, Mikrochim. Acta 130, 253(1999).

15. J. Kolar, A. Stolfa, M. Strilic, M. Pompe, B. Pihlar, M. Budnar, J. Simcic,and B. Reissland, Anal. Chim. Acta 555, 167 (2006).

16. H. Bennis, R. Benslimane, S. Vicini, A. Mairani, and E. Princi, J. ElectronMicrosc. 59, 91 (2010).

17. E. Princi, S. Vicini, E. Marsano, and V. Trefiletti, Thermochim. Acta 468,27 (2008).

18. B. Wagner, E. Bulska, A. Hulanieki, M. Heck, and H. M. Ortner, FreseniusJ. Anal. Chem. 369, 674 (2001).

19. K. Castro, E. Princi, N. Proietti, M. Manso, M. L. Carvalho, D. Capitani, S.Vicini, and J. M Madariaga, ‘‘Assessment of the weathering effects oncellulose based materials through a multianalytical approach’’, Anal.Bioanal., paper submitted (2010).

20. G. Piantanida, M. Bicchieri, and C. Coluzza, Polymer 46, 12313 (2005).21. E. Vassallo, L. Laguardia, D. Ricci, and G. Bonizzoni, Macromol. Symp.

46, 238 (2006).22. T. Magalhaes, A. von Bohlen, M. L. Carvalho, and M. Becker,

Spectrochim. Acta, Part B 61, 1185 (2006).23. A. Rindby, X-ray Spectrom. 18, 113 (1989).24. J. Bohman and J. Isakson, X-Ray Spectrom. 20, 305 (1991).25. M. Manso, M. Costa, and M. L. Carvalho, Appl. Phys. A 90, 43 (2008).26. I. Queralt, J. Ibanez, E. Marguı, and J. Pujol, Spectrochim. Acta, Part B 65,

583 (2010).27. L. Segal, J. J. Creely, A. E. Martin, and C. M. Conrad, Text. Res. J. 29, 786

(1959).28. R. H. Atalla and D. L. VanderHart, Science (Washington, D.C.) 223, 283

(1984).29. V. F. Hanson, ‘‘Determination of the Trace Elements in Paper by Energy

Dispersive X-Ray Fluorescence’’, in Advances in Chemistry Series, J. C.Williams, Ed. (American Chemical Society, Washington, D.C., 1981), p.53.

30. M. Manso, M. Costa, and M. L. Carvalho, Nucl. Instrum. Meth. A 580,732 (2007).

31. M. Manso and M. L. Carvalho, J. Anal. At. Spectrom. 22, 164 (2007).32. T. Collings and D. Milner, Paper Conserv. 14, 58 (1990).33. J. L. Ferrero, C. Roldan, D. Juanes, J. Carballo, J. Pereira, M. Ardid, J. L.

Lluch, and R. Vives, Nucl. Instrum. Meth. B 213, 729 (2004).34. The ink corrosion website (2008), http://www.knaw.nl/ecpa/ink/index.html

[accessed 1 January 2010].35. T. Cechack, T. Trojek, L. Musilek, and H. Paulusova, Appl. Rad. Isotopes

68, 875 (2010).

APPLIED SPECTROSCOPY 59