Characterisation and reproduction of yellow pigments ... - Springer Link

Appl. Phys. A 83, 557–565 (2006)

Applied Physics A

DOI: 10.1007/s00339-006-3525-0

Materials Science & Processing

g. bultrini1,u i. fragala` 1 g.m. ingo2 g. lanza 3

Characterisation and reproduction of yellow pigments used in central Italy for decorating ceramics during Renaissance 1

Dipartimento Scienze Chimiche, Università di Catania, v.le Doria 6, 95125 Catania, Italy Istituto per lo Studio dei Materiali Nanostrutturati del Consiglio Nazionale delle Ricerche, CP 10, Monterotondo Stazione, 00016 Rome, Italy 3 Dipartimento di Chimica, Università degli Studi di Potenza, via Nazario Sauro 85, 85100 Potenza, Italy 2

Received: 30 June 2005/Accepted: 18 January 2006 Published online: 9 March 2006 • © Springer-Verlag 2006 ABSTRACT This study presents the characterisation of prototypical yellow pigments used during the Renaissance period in Italy and the successful reproduction of homologous materials in accordance with the ancient recipes. Moreover, a large number of yellow decorative layers of Sicilian ceramic artefacts dated back from 13th to the 19th century have been selected and the main chemical, structural and minero-petrografic features have been studied by X-ray diffraction, optical microscopy and scanning electron microscopy-energy dispersive spectrometry. These results have been compared with literature data of some yellow decorations of Renaissance ceramics made in central Italy. Comparison has also been made with homologous materials that have been successfully reproduced in accordance with ancient recipes described by Cipriano Piccolpasso in the textbook: “I Tre Libri dell’Arte del Vasaio” using the same ingredients proposed by this artist. Such yellow materials reproduce the typical yellow colorants used by craftsmen of relevant sites for ceramic fabrication in central Italy, namely Citt`a di Castello, Urbino and Castel Durante, during the 16th century. Comparative arguments have shown some intriguing differences that are indicators of both technological transfer processes between central and southern Italy as well as of some local implementations likely due to specific raw materials locally available. PACS 81.05.Je; 82.80.-d; 68.37.Hk; 68.55.-a; 61.66.Fn; 61.10.Nz

1

Introduction

perspective, several attempts have been made by various researchers to correlate chemico-physical data of selected artefacts with ingredients of ancient recipes, thus proposing the raw materials used for their production. In this broad context, the present paper focuses on yellow decorations and compares the microchemical and microstructural peculiarities of some reproduced yellow ceramic pigments largely used during the Renaissance in central Italy, with those of several coeval ceramic decorations produced in various relevant production sites in Sicily (southern Italy) as well as with data of some related XVI century yellow decorations of central Italy artefacts [5]. Therefore, the yellow pigments typical of different towns of the central Italy, namely Città di Castello, Urbino and Casteldurante, have been successfully reproduced, in accordance with the ancient recipes reported in “I tre libri dell’arte del vasaio” by Cipriano Piccolpasso in the middle of the XVI century [7]. These reproduced pigments have been fully characterised, and then brushed over enamelled ceramic substrates and fired in the oven, using firing conditions similar to those used in Renaissance kiln (temperature range 900– 950 ◦ C; ∆T/t = 10 ◦ C/min and 80%Ar + 20%O atmosphere). The maturing temperatures have also been evaluated and possible tone variability depending on the atmosphere of the kiln during the firing has similarly been investigated. For this reason some sticks have been decorated with uniformly distributed coloured strips and fired inside an horizontal hot walls reactor (Fig. 1) adopting either a lightly oxidant or lightly reducing atmosphere, by tuning the gas flows into the reactor. In the same way, the yellow decorated surfaces of several Sicilian ceramic fragments [10], found in various Sicilian sites (Caltagirone, Syracuse, Piazza Armerina, Sciacca

In recent years there has been a large debate [1–6] on the chemical nature of ancient ceramic colorants in the perspective of a possible reproduction of old decorations as well as of a better understanding of materials and methods adopted by craftsmen for producing decorative layers. This matter is well addressed in some reference textbooks: “I tre libri dell’arte del vasaio” written in the 16th century by Cipriano Piccolpasso [7], in the “Persian treatise on pottery production” by Abu’l-Qasim (1301 A.D.) [8] and in the “De coloribus et artibus Romanorum” by Eraclius (10th century) [9]. In this u Fax: +39-095/580138, E-mail: [email protected]

FIGURE 1

Horizontal hot walls reactor scheme

558

Applied Physics A – Materials Science & Processing

Sample

Provenance site

Area

10 13 20 43 53 57 58 C1 P1 P2 P3 P4 P5 P6 P7 P8 P9 P11 P12

Syracuse Syracuse Syracuse Piazza Armerina, Villa del Casale Caltagirone Caltagirone Sciacca Caltagirone Paternò Paternò Paternò Paternò Paternò Paternò Paternò Paternò Paternò Paternò Paternò

Central Italy Local Local Local Local Local Local Local Local Local Local Local Local Local Local Local Local Local Local

TABLE 1

XVIII XIII XV/XVI X XVII XVIII XVII XVI XVIII XVI XVII XVIII XVI XVIII XVIII XVIII XVII XVII XVIII

Description Blue/yellow decorated Majolica Green/brown/yellow decorated Protomajolica Engobbed yellow/brown decorated Graffita Green/brown/yellow decorated fragment, Medieval Blue/yellow decorated Majolica Green/brown/yellow decorated Majolica Green/brown/yellow/blue decorated Majolica Blue/yellow decorated Majolica Green/brown/yellow decorated Majolica Blue/yellow decorated Majolica Blue/yellow decorated Majolica Green/brown/yellow/blue decorated Majolica Blue/yellow decorated Majolica Blue/yellow decorated Majolica Blue/yellow decorated Majolica Blue/yellow decorated Majolica Blue/yellow decorated Majolica Blue/yellow decorated Majolica Green/brown/yellow/blue decorated Majolica

Brief description of present ceramic samples

and Paternò) (Table 1) and those of a sample (labelled 10 in Table 1) whose decorative features suggest provenance from a central Italy production (probably Faenza, courtesy of “Museo Regionale della Ceramica di Caltagirone”) have been fully characterised. All these data have been compared with those of decorations of mentioned central Italy artefacts [5] as well as with those of decorations reproduced in our laboratory. Possible relationships between Piccolpasso’s recipes and Sicilian pigments have emerged. 2

Century AD

Materials and methods

Present reproductions of yellow pigments have required retrieval of the same materials available during Renaissance. Thus, the “tartaro di botte” (barrel tartar), present in many Piccolpasso recipes, has been recovered upon scraping the internal surfaces of some old wine barrels belonging to Sicilian wine industries. The XRD patterns of this ingredient revealed that it consists exclusively of potassium tartrate (Fig. 2). This ingredient was used by ancient crafts-

men as a flux in combination with other, generally leadbased fluxes. Similarly, the so-called “ruggine” (rust) has been obtained from old rusty iron ware, in accordance with the recipe of Piccolpasso [togliasi ferraccia, o vogliam ruggine di ferro, e la migliore è quella che si coglie dintorno alle ancore delle navi] “We want iron dust, and it is better when one can collect around ship anchors” by scraping with stainless steel lancets old iron anchors stored in the Catania port. The recipes presently adopted for the four yellow pigment are reported in the following Table 2. The ingredients are expressed in libre (pounds) and oncia (ounce). The “zallolino” of Città di Castello, besides the weights expressed in libre, appears the oncia (ounce) unit. It is worthy of note that during the Renaissance both “libra” and “oncia” were variable units for each city. For example in Rome the “libra” was equivalent to 339 g; in Tuscany to 348 g; and in

The “zallo” (yellow pigment) of Casteldurante (the Piccolpasso birth city): Piombo (Pb3 O4 ) Antimonio (Sb2 O3 ) Ferraccia (Fe2 O3 )

libra libra libra

5 3 2

The “zallolino” (light yellow pigment) of Città di Castello: Piombo (PbO) Antimonio (Sb2 O3 ) Sale (NaCl) Tartaro

libra libra oncia oncia

1.5 1 1 1

The “zallolino della Marca” light yellow typical of many cities of the Marche region: Piombo (PbO) Antimonio (Sb2 O3 ) Tartaro

libra libra libra

6 4 0.5

The “zallolino all’Urbinate”, light yellow typical of Urbino: Piombo (PbO) Antimonio (Sb2 O3 ) Sale (NaCl) Tartaro FIGURE 2 XRD pattern for the “Tartaro” showing the presence of potassium hydrogen tartrate

TABLE 2

libra libra libra libra

Ancient recipes for producing the yellow pigments

6 4 1 2

BULTRINI et al.

Characterisation and reproduction of yellow pigments used in central Italy

559 FIGURE 3 Typical aspects of the four reproduced yellow pigments after firing

Ferrara to 345.85 g [11]. In present reproduction of pigments the weight of 348 g and 29 g (the twelfth part of the libra) have been adopted for the libra and oncia, respectively. Moreover, Piccolpasso reports different proportions for the same recipe. Differences could be likely due to possible variations used in the same production centres. Here, only one variation for each production centre has been selected. In the present preparations of the four yellow pigments the various ingredients have been very well mixed, placed in little hand-made “terracotta” tubs and fired in oven at 920 ◦ C (heating rate of 10 ◦ C min−1 ) in accordance with ancient firing conditions [1, 12]. Samples have been kept at 920 ◦ C for 120 min. The obtained solids samples have been finely pulverised and characterised. XRD measurements (Bruker D5005 X-ray Diffractometer) have been made to detect the neo-crystalline phases present in the reproduced pigments. Identification of these species was carried out using a diffrac plus evaluation program (EVA) software. Microstructures and elemental compositions have been investigated by scanning electron microscopy and energy dispersive spectrometry (SEM + EDS) using a Leo Iridium 1450 microscope equipped with an EDS apparatus (IXRF SYSTEM) and a four sector back-scattered electron detectors. Reference decorated samples have been obtained by brushing previously enamelled sticks with reproduced pigments1 . Pigment strips were spatially distributed and fired inside an horizontal hot walls reactor with different temperature gradients adopting either a lightly oxidant or lightly reducing atmosphere by tuning the gas flows into the reactor. The surfaces of Sicilian ceramic fragments and of the reproduced decorations were gently and mechanically cleaned with a glass fibre brush. Then, coloured decorations were removed with a clean and new little diamond cut-off wheel was applied to a flex shaft and mounted onto electron microscope stubs with carbon cement. 1 The

ceramic colour has been prepared, in accordance with Piccolpasso’s suggestion, by admixing the pigment with H2 O diluted enamel.

The sample was coated with a thin layer of graphite to avoid charging effects. These coatings (3.0 nm) were deposited with a Emitech K450 sputter coater. Thin sections of the ancient Sicilian ceramic fragments were examined under transmitting light, using a polarising Leitz optical microscope, for both the petrographicalmineralogical characterisation of the yellow decorations and for the microscopic observation of different mineral phases in the glaze. Thin sections were prepared by cutting small ceramic slabs, taken from decorated areas of the samples, using a diamond saw blade to obtain a fragment with a flat surface. Then, the surface was first polished to flatten any roughness and make it as smooth as possible and then mounted onto the microscope slide with a resin or Canada balsam. Mounted samples were again cut parallel to the flat surface to produce a second parallel face as thin as possible. This face was lapped with low-grade (20 – 30 microns) silicon carbide paper and polished with diamond pastes until an approximate thickness of 50 microns. Finally, hand polishing using a diamond paste on a flat glass surface was adopted. The thickness of the resulting thin section was about 30 microns. 3

Results and discussion

Four Piccolpasso yellow pigments recipes were reproduced by mixing the different ingredients (Table 2). Then, the mixtures, placed in terracotta tubs, were fired into oven at 920 ◦ C, adopting a standstill time at 920 ◦ C of 120 min and a heating rate of 10 ◦ C min−1 . Figure 3 shows typical aspects of the four reproduced yellow pigments. After firing, several yellow tones appeared as well as different consistencies (Table 3). XRD diffraction patterns, beside the ubiquitous presence of bindheimite (Pb2 Sb2 O7 ), point to several differences between the four pigments, despite the close analogies among the adopted recipes. The most relevant crystalline phases are listed in Table 4. It is worthy to note that all the XRD patterns do not show background bumps, thus denoting a very efficient

560

Pigment

Applied Physics A – Materials Science & Processing Aspect of the reproduced pigments Colour Consistence

Giallo of Castedurante Giallolino all’Urbinate Giallolino della Marca Giallolino of Città di Castello TABLE 3

dark yellow yellow intense yellow yellow

compact but friable pulverulent compact but not vitrified slightly vitrified

Description of the consistence of the reproduced pigments after

firing

FIGURE 4

Prototypical XRD pattern for the so called “Giallolino della

Marca”

crystallite formation. The prototypical XRD pattern of “Giallolino de la Marca” is displayed in Fig. 4. In particular, the XRD reflections of the “zallo di Casteldurante” point to the exclusive presence of bindheimite, while

the “Giallolino all’Urbinate” consists of bindheimite and less frequent PbSb2 O4 /PbO · Sb2O3 . The “Giallolino de la Marca” pigment is mainly composed of bindheimite, Pb2 Sb2 O7 , PbSb2 O6 and Pb3+x Sb2 O8+x . Finally, the XRD pattern of the “Giallolino di Città di Castello” shows the presence of bindheimite, NaSbO3 and Sb6 O13 . It, therefore, comes out that the different proportions in the recipes result in slightly different crystalline phases, and, as a consequence, in different thermo-chemical reactions. Differences among the four pigments are also evident at the microscopical level. Thus SEM-EDS data of the surfaces of reproduced yellow decorations reveal different chemical composition and, even more interesting, differently structured pigment micro-particles. Thus, the micro-structural analyses of the “Giallo di Casteldurante” and of the “giallolino di Città di Castello” provide evidence of the contemporary presence of agglomerations of irregular crystals and of well-formed tetrahedral and hexagonal pigment particles embedded in the glaze (Fig. 5a and b, respectively). The dimensions of tetrahedral particles range from 1 to 3 microns, while, those of hexagonal crystallites range from 2 to 5 microns wide. In these samples a few zoned lead antimonate crystals are also present. By contrast, the SEM analysis of the “Giallolino all’Urbinate” pigment reveals an exclusive presence of irregular crystal agglomerates not homogeneously distributed in the glaze (Fig. 5c). Finally, the microstructural observation of the “Giallolino de la Marca” shows frequent large hexagonally shaped particles, whose size is about 4 µm, mixed with agglomerations of small (about 0.5 µm) irregular crystals (Fig. 5d). These micro-structural differences could be attributed to chemically inhomogeneous micro areas in the mixture.

FIGURE 5 Backscattered electron images showing the microstructures of the reproduced pigments of “Giallo di Casteldurante”, “Giallolino di Città di Castello”, “Giallolino all’Urbinate” and “Giallolino de la Marca”, images (a)–(d), respectively

BULTRINI et al.


Pigment Giallo of Castedurante Giallolino all’Urbinate Giallolino della Marca Giallolino of Città di Castello

TABLE 4 XRD mineralogical data of the reproduced pigments

Diffractometric analyses of the reproduced pigments Phases bindheimite bindheimite bindheimite bindheimite

– PbSb2 O4 /PbO · Sb2 O3 Pb2 Sb2 O7 NaSbO3

Average EDS quantitative chemical data (Table 5) as well as the PbO/Sb2 O3 ratios found in the pigment particles show a large spreading of compositional data. This observation suggests that the composition of the source materials remarkably affect microstructures of final pigments. Moreover the compositional variability, among microparticles of the same pigment (Table 5) points to micro-segregation effects due to local compositional fluctuation. Thus the PbO/Sb2 O3 ratios lie in wide range and differ from the 1.38 ratio expected for bindheimite (Pb2 Sb2 O7 ) [13]. Nevertheless, it is important to note that the PbO/Sb2 O5 ratio of the “Giallolino di Città di Castello” agrees well with the scarce data to date presented. In fact, it has been reported that in two Renaissance decoration yellow layers of tiles produced in Venice (Italy) PbO/Sb2 O5 ratio lies around 2.30 [5]. This value compares well with 2.20 found in the presently reproduced pigment. The differences found in the case of other oxides (Table 5) might be also associated to different enamel compositions. It, therefore, transpires that the Venetian craftsmen have used a recipe similar to that of the “Giallolino di Città di Castello”. These compositional fluctuations cannot, in any case, depend on firing conditions since identical procedures have been adopted for reproduction of the four pigments. Therefore, the simultaneous presence of areas rich of large Pb and Sb

– – PbSb2 O6 Sb6 O13

561

– – Pb3+x Sb2 O8+x –

oxide crystals and zones with agglomerations of small irregular crystals is probably due to the existence of particular micro-areas having favourable conditions for the growth of such large crystals. Similar argumentation can account for the different crystal typologies inside the same sample. Interesting enough, SEM-EDS microchemical and microstructural data of present ancient Sicilian yellow ceramic decorations similarly show non-uniform distribution of the pigment in the glaze and, in the majority of cases, the simultaneous presence of different crystalline phases. Thus, the study of 19 reference ancient ceramic fragments has revealed pigments consisting of small irregular crystals embedded in the glaze only in eight samples (labelled C1, P1, P4, P7, P11, P12, 53 and 58 in Table 1). By contrast, the remaining samples show different and, in some cases, complex structural situations. In particular, the ceramic fragments P5 and P9 present zones with association of tetrahedral and hexagonal crystals (Fig. 6a), and the samples P3, P8, 10 and 57 show the presence of abundant hexagonal crystals associated to irregular microcrystal agglomerations (Fig. 6b). The sample P2 presents areas with exclusive hexagonal crystals and areas with only octahedral crystals embedded in the glaze (Fig. 6c). Nevertheless, the most complex structural situation is present in the sample P6 that shows zones with

FIGURE 6 Backscattered electron images for the microstructures of the ceramic fragments: P5, P57, P2 and P6, images (a)–(d), respectively

562


co-existence of hexagonal and tetrahedral crystals and zones with combined octahedral and hexagonal crystals (Fig. 6d). It is worthy to note that in the samples P2, P6, P8 and P9 diffused crystals with dark zones are present. The chemical composition is very similar either in the centre or along the border, as shown in Fig. 7. Similarly to the reproduced pigments, the ancient ceramic samples also have different and not homogeneous micro-

chemical compositions of several crystal typologies, even inside the same ceramic decoration (Table 5). It is relevant to note that the comparison between the chemical data of the ancient yellow lead antimonate decorations emphasises the ubiquitous presence of iron oxide in the Sicilian samples (from 1.36% to 4.9%) and the almost total lack (0.36%) in the sample 10, attributed by archaeologist to central Italy production (courtesy of the Museo Regionale della Ceramica di Cal-

FIGURE 7 Backscattered electron images of the hexagonal crystals present in the ceramic fragment P2. EDS spectra show the chemical composition of the different zones. Only a very slight difference is appreciable

BULTRINI et al.


tagirone). This observation agrees well with literature data of central Italy decorations, where the iron content is generally low (at most 1.56%). As far as chemical data of the medieval sample 43, the protomajolica sample 13 and the graffita sample 20 (dated back to XIV-XV century) present yellow pigments with chemical compositions very different from those previously discussed and lacking of Sb content. This is an indication of the use of different recipes than those adopted

563

in the following centuries. Thus, SEM-EDS data of medieval samples show that the pigment phases consist of diffuse crystals with irregular shapes. It is worthy to note that these pigment particles, rich in lime, magnesia, silica and iron oxide, are much more uniform then the particles found in the present in Pb/Sb pigments. The SEM-EDS study in case of protomajolica sample (sample 13 Table 1) has shown zoned pigment particles with rounded shape and chemical composition rich

FIGURE 8 Backscattered electron image showing the different crystalline phases present in the ceramic fragment 20. EDS spectra evidence large compositional differences between the different phases

564


Sample

Description

Giallo Castdurante Giallo Castdurante Giallolino Castell Giallolino Castell Giallolino Marca Giallolino Marca Giallolino Urb Giallolino Urb 43 43 13 13 20 20 C1 C1 10 10 53 53 57 57 58 58 P1 P1 P2 P2 P2 P3 P3 P4 P4 P5 P5 P6 P6 P6 P7 P7 P8 P8 P9 P9 P9 P11 P11 P12 P12 142∗ 48∗ ∗

Hexagonal Phase Yellow-matrix Tetraedral phase Yellow-matrix Hexagonal Phase Yellow-matrix Yellow-phase Yellow-matrix Yellow Phase Yellow-matrix Yellow-phase Yellow-matrix Hexagonal Phase Yellow-matrix Yellow-phase Yellow-matrix Hexagonal Phase Yellow-matrix Yellow-phase Yellow-matrix Hexagonal Phase Yellow-matrix Yellow-phase Yellow-matrix Yellow-phase Yellow-matrix Hexagonal Phase Ottaedral phase Yellow-matrix Hexagonal Phase Yellow-matrix Yellow-phase Yellow-matrix Tetraedral phase Yellow-matrix Tetraedral phase Hexagonal Phase Yellow-matrix Yellow-phase Yellow-matrix Hexagonal Phase Yellow-matrix Tetraedral phase Hexagonal Phase Yellow-matrix Yellow-phase Yellow-matrix Yellow-phase Yellow-matrix Yellow Yellow

Na2 O

MgO

Al2 O3

SiO2

K2 O

0.33 1.68 1.44 1.94 0.51 1.26 1.16 2.30 0.81 1.41 0.74 0.31 0.13 0.42 0.92 1.16 0.14 0.77 0.28 1.04 0.99 1.35 0.41 1.25 0.63 1.00 0.12 0.26 0.62 0.46 0.73 0.22 1.39 0.42 0.97 0.45 0.56 0.36 0.90 1.65 0.53 1.24 0.29 0.55 0.84 0.42 0.22 0.73 1.33 n.d. n.d.

0.41 0.18 0.23 0.38 0.47 0.22 0.57 0.30 6.87 0.40 0.48 0.32 0.39 0.27 0.23 0.47 0.23 0.22 0.43 0.46 0.36 0.55 0.29 0.29 1.90 0.29 0.31 0.59 0.46 0.46 0.31 0.30 0.39 0.33 1.05 0.58 0.43 0.95 0.84 0.47 0.52 0.92 0.40 0.26 0.44 0.46 0.61 0.52 0.39 tr. 0.25

0.60 2.29 1.32 2.98 1.20 3.63 1.57 3.00 5.29 6.00 4.03 2.15 2.01 3.80 1.29 3.76 0.57 2.58 1.56 2.98 0.91 4.91 0.85 2.99 1.92 3.74 1.07 1.06 4.09 1.06 4.11 0.86 5.33 0.94 3.52 1.23 1.46 4.09 2.49 4.17 1.90 3.43 1.06 0.62 3.80 1.42 5.83 0.89 5.25 1.83 1.95

2.90 37.92 12.77 38.28 9.70 45.26 15.41 41.91 48.62 39.99 53.38 40.00 0.73 23.34 11.23 37.61 3.55 55.54 13.32 44.69 2.95 46.48 5.05 56.40 15.06 43.45 5.30 5.65 44.93 6.71 44.76 4.27 47.21 2.62 43.84 4.16 8.96 63.31 22.02 50.24 14.08 43.05 5.45 3.17 33.04 6.36 43.24 3.11 49.38 31.15 26.12

0.08 0.15 0.08 0.33 0.03 0.85 0.10 0.54 0.75 1.80 3.89 1.17 0.09 0.37 0.00 1.67 0.03 4.31 0.21 3.25 0.00 1.53 0.04 3.30 0.21 2.27 0.00 0.00 1.68 0.06 2.77 0.02 3.17 0.00 1.41 0.00 0.00 1.61 0.46 2.43 0.23 2.73 0.00 0.00 1.13 0.02 1.18 0.02 3.22 2.4 2.37

Oxide wt. % CaO MnO 1.00 1.26 1.92 0.37 2.59 0.00 1.48 0.59 18.69 4.71 4.66 1.83 0.18 0.55 4.20 1.51 13.96 1.17 5.88 1.15 6.49 2.89 2.73 0.27 1.02 0.19 5.19 5.92 1.71 4.20 1.96 3.99 0.27 7.06 2.92 4.53 3.34 2.15 4.81 2.17 5.01 3.53 2.82 3.26 0.87 3.76 5.31 7.02 1.40 5.23 5.3

– – – – – – – – 0.22 0.19 – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – tr. tr.

FeO

SnO2

Sb2 O5

PbO

PbO/Sb2 O5

3.39 0.54 0.45 0.34 0.45 0.39 1.81 0.26 11.58 4.40 14.79 4.02 52.42 4.60 4.90 4.56 0.32 0.33 5.37 7.81 2.88 3.19 3.23 0.49 3.39 0.57 2.39 2.73 2.47 1.88 2.04 1.40 2.90 2.33 2.96 2.51 1.91 3.30 1.36 1.61 4.07 7.21 3.05 3.83 2.72 2.88 10.83 3.02 3.37 1.56 1.32

0.78 0.60 0.73 0.65 1.07 0.69 0.80 1.58 1.82 0.78 5.27 1.66 0.31 0.64 3.11 0.45 1.60 1.91 6.28 0.98 10.94 0.47 3.64 0.63 2.23 0.80 1.86 2.70 0.86 0.80 0.34 12.11 0.71 2.24 0.65 4.34 9.04 0.57 0.48 0.69 5.70 1.07 6.46 3.21 0.45 11.01 0.66 0.76 0.96 8.22 8.6

22.88 0.96 25.49 1.21 42.84 0.72 18.63 0.39 – – – – – – 24.72 0.31 62.84 0.22 25.95 0.82 36.73 0.23 37.46 1.06 23.16 0.38 28.57 28.50 0.75 30.58 1.18 21.35 0.08 36.68 0.10 29.82 32.96 0.00 26.63 1.39 22.21 0.39 22.99 33.49 0.38 26.43 3.05 48.36 0.99 14.92 15.96

67.63 54.42 55.57 53.51 41.13 46.98 58.46 49.12 5.36 40.31 12.75 48.56 43.75 66.00 49.40 48.50 16.76 32.94 40.72 36.84 37.76 38.41 46.29 33.32 50.48 47.32 55.19 52.58 42.42 53.77 41.80 55.47 38.55 47.39 42.57 52.38 41.34 23.66 40.00 35.18 45.74 36.44 57.48 51.62 56.32 47.25 29.06 35.56 33.70 34.29 37.91

2.96 2.18 0.96 3.14

2.00 0.27 1.57 1.03 1.24 2.18 2.14 1.84 1.76 2.60 1.29 1.95 1.25 1.50 2.06 2.78 1.54 1.79 0.74 2.30 2.38

= Fabbri et al. [5]

TABLE 5

EDS analysis (expressed as weight percent, wt. %) of the pigment particles and relative glazed matrix

in aluminium, silicon, potassium, calcium and iron oxides relative to the embedding matrix. Interesting enough, data of the yellow decoration of the sample 20 (engobbed graffita pottery, XV–XVI century), provides evidence of hexagonal shaped crystals embedded in the glaze (Fig. 8), very similar to the hexagonal crystals found in present lead-antimonatebased yellow decoration. In this case, however, the Fe ion substitutes for Sb, and, moreover, the crystals are more rich of Fe and Pb oxides than the related matrix. In particular, the relative contents seem to point to hexagonal plumboferrite (PbFe4 O7 ) crystals (Table 5). Besides these hexagonal phases, other pigment particles are present into the decora-

tion layer (Fig. 8). EDX data indicate that crystals consist almost exclusively of iron oxide, whereas the zoned dark particles are rich of silicon, lead, iron, aluminium, calcium and potassium. The discussed microchemical differences among the various yellow decorated layers become also evident with optical microscopy observation of thin sections of the ceramic fragments. Thus examinations in transmitted light reveal similar thickness in the yellow decorations (about 50 µm) and almost total absence of bubbles. Under crossed nicols the various yellow layers show different gradations correlated to the chemical compositions.

BULTRINI et al.


Finally, the present reproduced yellow pigments brushed as strips on enamelled ceramic sticks and fired inside a controlled atmosphere, horizontal hot walls reactor reveal that these pigments do not undergo appreciable chromatic changes under different firing temperatures as well as under different composition of the gas fed into the oven during firing. 4

Conclusions

Compared data of microchemical/microstructural investigations of i) ancient Sicilian yellow decorations, dated back from 10th century to 19th century, ii) of presently reproduced pigments according to Piccolpasso recipes and, finally, iii) of current reported literature data of some yellow decoration of XVI century, central Italy artefacts provide evidence of remarkable differences. SEM-EDS data of Sicilian samples earlier than XVI century (X–XV century) show that iron oxide has been almost exclusively used, whereas in yellow pigments of Renaissance Sicilian samples, beside Pb and Sb oxides, the iron oxide was also constantly present. By contrast iron oxide represents an ingredient of only yellow (“giallo”) pigments used in central Italy in accordance to Piccolpasso recipes. Nevertheless, the content remains generally below that found in the Sicilian recipes. Furthermore, the Sb-oxide content in Sicilian yellow decorations earlier than XVI century (samples 43, 13 and 20) is absent, thus pointing to different recipes adopting Fe instead of Sb oxide. In particular the hexagonally shaped crystals embedded in the glaze of sample 20 (Table 5) consist almost exclusively of Fe and Pb oxides and certainly represent attempts to reproduce the Pb/Sb formulation using Fe instead of Sb. In fact, both in Sicily and, generally, in southern Italy, Sb is not easily available. In the same context there is evidence that the light yellow decoration (Giallolino) was obtained by the ancient Sicilian craftsmen by diluting the pigment. Different microstructures have been, in addition, presently highlighted among ancient Sicilian decorations as well as among the presently reproduced pigments. In particular, either hexagonal crystals or triangularly shaped crystals,

565

besides irregular small phase agglomerates, have been observed in reproduced pigments. In contrast, different morphologies, namely tetrahedral and octahedral crystals have been detected in the ancient Sicilian samples in addition to mentioned forms. These preliminary observations suggest that the hexagonal crystals result from the coalescence of triangular-shaped particles that, in turn, develop in extended plates upon cooling. These aspects of morphogenesis require further investigations and experiments, now in progress, adopting different firing conditions. In any case present data witness different implementations of known procedures by the ancient Sicilian craftsmen to produce, fire and apply their yellow ceramic pigments. ACKNOWLEDGEMENTS The Science and Technology Park of the Sicilian Regional Government (PSTS, Italy) and MIUR (Ministry of Italian University Research) are gratefully acknowledged for their financial support.

REFERENCES 1 R. Alaimo, G. Bultrini, I.L. Fragalà, R. Giarrusso, G. Montana, Appl. Phys. A 79, 263 (2004) 2 G. Padeletti, P. Fermo, S. Gilardoni, A. Galli, Appl. Phys. A 79, 335 (2004) 3 P. Comodi, M. Bernardi, A. Bentivoglio, G.D. Gatta, P.F. Zanazzi, Archaeometry 46, 405 (2004) 4 K. Sakellariou, C. Miliani, A. Morresi, M. Ombelli, J. Raman Spectrosc. 35, 61 (2004) 5 B. Fabbri, V Fassina, A. Rattazzi, D. Salvioni, in: Fourth EuroCeramics, Vol. 14, ed. by B. Fabbri (Gruppo Editoriale Faenza Editrice, 1995) p. 383 6 R.J.H. Clark, L. Curri, G.S. Henshaw, C. Laganara, J. Raman Spectrosc. 28, 105 (1997) 7 C. Piccolpasso, I Tre Libri dell’Arte del Vasaio, ed. by A. Nobili, Pesaro, Italy 1879, from the original manuscript (1557) 8 J.W. Allan, Abu’l-Qasim’s, Treatise on Ceramics, Iran IX, 111 (1973) 9 Eraclius, De Coloribus et Artibus Romanorum, translated by M.P. Merrifield in Original Treatises Dating from the XIIth to XVIIIth Centuries on the Arts of Painting in Oil, Paris, 1849 10 A. Ragona, La maiolica siciliana dall’origine all’ottocento (Edizioni Sellerio, Palermo, 1985) 11 B. Scappi, Opera (Michele Tramezzino, Venezia, 1570) 12 G. Padeletti, P. Fermo, Appl. Phys. A 76, 515 (2003) 13 A.J. Shortland, Archaeometry 44, 517 (2002)