1 Introduction and Evolution of EDXRF-Devises

The use of energy-dispersive X-ray fluorescence (EDXRF)-analysis to determine pigments in paintings and metals, to our knowledge, starts in 1971, and the first paper “Non-destructive analysis of chemical elements in paintings and enamels” by R. Cesareo et al., from the Institute of Physics, Faculty of Engineering, University of Rome, and M. Marabelli et al., from the Central Restoration Institute, Rome, was published in 1972 [1]. The paper was published in Archaeometry and reported EDXRF-analysis of various paintings, among them the “Deposition” by Raphael, located at Galleria Borghese (Fig. 1), Rome. The paintings were analyzed at the Istituto Centrale del Restauro, Rome, where they were under restoration.

Fig. 1
figure 1

The “Deposition” by Raphael painted in 1507. It is exposed in the Galleria Borghese in Rome

Full size image

The equipment used at the time, 48 years ago, may be considered today rather primitive, being composed of a radioactive source (of Cd-109, Am-241, Pm-147 or Co-57), or Kr, Ne, or Xe gas proportional counters with an energy resolution of about 18% at the energy of 6.4 keV (iron Kα-line) [2] and an electronic chain coupling the detector to a LABEN 400 pulse-height analyzer. The device had a total weight of about 50 kg and was transportable, not portable (Fig. 2).

Fig. 2
figure 2

Equipment employed to analyze the “Deposition” by Raphael and other paintings at the Central Restoration Institute of Rome. It is analyzing the “Madonna con bambino” by Lucas Cranach. Main Parts: (2): measuring head with radioisotopic source and gas proportional counter; (3) measuring head with radioisotopic source and NaI(Tl)-detector (not employed); (4) electronic chain; (5) LABEN multi-channel analyzer

Full size image

Due to the poor energy resolution, the X-ray spectra were not easy to process. However, it was possible to get the following:

  • identify the typical pigments used by Raphael: malachite, lead-tin yellow, white lead, lapis lazuli, a Cu-pigment, cinnabar, azurite, ochre.

  • identify areas of modern restoration, through identification of “modern pigments” based on the detection of Zn-white and Ti-white industrially produced around 1850 and 1920 respectively [3], but also by pigments based on chromium, selenium, and cadmium produced only recently. The last point was particularly important for the restoration of the painting, and the restorers removed the pigment layer containing “modern” pigments.

Approximately at the same time, a portable EDXRF equipment was developed and used to analyze ancient metals (bronzes, brasses, gold, silver….) in museums, churches, open air, and so on. This equipment was characterized by a radioactive source, a gas proportional counter, and a single-channel analyzer (SCA) coupled to a scaler. The SCA was modified in order to select, keeping fixed the window (interval of energy). Six switchable channels determine the energy of well-defined single elements (for example, in the case of bronzes Cu, Sn, Pb; in the case of gold Au, Ag, and Cu). In such a manner, a portable equipment was constructed, able to analyze only selected elements, but with poor sensitivity. The device is shown in Fig. 3, and a typical gold EDXRF-spectrum showing how the SCA works is shown in Fig. 4 [4,5,6].

Fig. 3
figure 3

Portable device for analysis of metals (1972). From left: proportional gas counter (B) with radioactive source in the middle of the Be-window (A); electronics for the detector. Home-modified S.C.A. (TC440M) with six windows prefixed on intervals of energy corresponding to the elements to be analyzed, and a switch (on the bottom). The S.C.A. is followed by a timer-scaler. An Etruscan bracelet from the Tomb Regolini-Galassi (Vatican Museum) is being analyzed

Full size image
Fig. 4
figure 4

EDXRF spectrum of a Au-alloy (Au = 75%. Ag = 18%, Cu = 7%) measured with a proportional gas counter and registered with a MCA. The energy intervals are also shown, which were used to determine Au, Ag, and Cu concentration with a S.C.A. in which the fixed window can be switched in four possible positions: Cu-counts (red); difference red-left minus red light; Au-counts (blue); Ag-counts (green). The Cu concentration, given by the difference between left and right red rectangles, must be 0. The very poor energy resolution is apparent and may be compared, for example, with the gold XRF spectrum shown in Fig. 10

Full size image

With such a device, Etruscan, Roman, and Greek bronzes were analyzed [4, 5], as well as Etruscan gold objects from various tombs and in museums, such as the pectoral (~ 675–650 B.C.) from the tomb Regolini-Galassi exposed at the Vatican Museum [6].

Soon after, the first revolution occurred: the new liquid N2-cooled semiconductor X-ray detectors, both Si(Li) and hyperpure Ge (HpGe) entered the market, with an energy resolution of 200–300 eV at 6.4 keV, much better than that of proportional gas counters, which is more than 1000 eV. The use of these detectors marked the first detector revolution and permitted to analyze pigments with much more resolution and accuracy. However, these detectors needed liquid nitrogen, and the total device, including a dewar of 5–10 L, is not portable, but only transportable (Figs. 5, 6, and 7). In the first phase, an equipment composed of radioactive sources and a Si(Li)-detector was employed to analyze several paintings under restoration at the Central Restoration Institute in Rome. An example is given by the study of the “Giudizio di Santa Lucia” by Lorenzo Lotto [7]. Using a set of two or four Am-241 point sources, many other paintings were analyzed in situ, such as La Fornarina by Raffaello in 1983 at National Museum of Palazzo Barberini in Rome (Fig. 6) [8]. This device consists in a HpGe detector equipped with confocal radioisotopic sources as shown in Fig. 5 [9]. Raffaello, Michelangelo (Tondo Doni), and Caravaggio (about twelves) paintings were analyzed in 2 to 3 years.

Fig. 5
figure 5

Image of the device that used Am-241 (emitting 59.6 keV gamma rays) sources in confocal configuration used at the beginning of the eighties in various campaigns of measurements in Italian museums [10]

Full size image
Fig. 6
figure 6

Measurement points and some results of the study of 1983 on Raphael table “La Fornarina,” using the device of the Fig. 7. It is possible to note the characteristic shape of the spectrum which presents at high energies the scatter peaks on the table by penetrating 60 keV photons [7]

Full size image
Fig. 7
figure 7

EDXRF device from the eighties of the last century; it is composed of a Gilardoni Radiolight X-ray tube, working at 80 kV and 1 mA maximum voltage and current, a CAMBERRA Si(Li) detector, and a MCA incorporated in the computer

Full size image

The second revolution happened at the beginning of the nineties with the development, around 1990–1995, of high-resolution small size, Peltier cooled HgI2 detectors (developed by Jan Iwanczyk and coworkers) (Fig. 8) [11], and, soon after, Peltier-cooled Si-PIN detectors (developed by John Pantazis and coworkers at AMPTEK) (Fig. 9) [12, 13]. These detectors had an energy resolution comparable to that of nitrogen-cooled Si detectors and allowed to assemble the first portable devices to carry out measurements in situ. However, the HgI2 detector revealed itself to be delicate versus time, and the Si-PIN was, therefore, generally preferred for portable devices, despite an important negative point: they accept only a reduced number of photons per second without losing energy resolution greatly increasing the required measuring time. Paintings and, for the first time, frescos were analyzed with the Si-PIN detector. Frescos from the church of S. Stefano Rotondo in Rome, painted by Antonio Tempesta in the year 1568, from the “Nolfi chapel” in the cathedral of Fano painted by Domenichino in 1618–1619 [14], from the church of S. Francesco in Arezzo painted by Piero della Francesca in 1452–1466 and from Giotto’s frescos in the chapel of the Scrovegni, painted in Padua in 1303–1305 (Fig. 9) [15, 16]. In frescos, the background is composed by plaster (CaCO3) that becomes CaSO4 and/or Cl (NaCl), due to pollution close to the sea. Then, alloys were analyzed such as the four panels from the Volvinius altar in the Basilica of St. Ambrose in Milan (824–859 A.C.) on gold and gilded-silver [17], the big sphere on gilded bronze located at the top of the basilica of St. Peter in Rome, and the famous marble statue of David by Michelangelo, particularly to detect and remove S due to pollution [18].

Fig. 8
figure 8

X-ray tube from EIS, working at 50 kV and 1 mA maximum voltage and current, and a room temperature (or Peltier cooled) HgI2. The MCA is incorporated in the computer

Full size image
Fig. 9
figure 9

EDXRF device for elemental analysis of frescos in the Chapel of the Scrovegni, painted by Giotto in Padua between 1303 and 1305. Two measuring heads were employed: one composed of a Hamamatsu X-ray tube working at 10 kV and 1 mA, and a Si-PIN detector, for the analysis of sulfur; the second composed of an Oxford X-ray tube with Mo-anode, for analysis of medium–high Z elements, is composed of a mini X-ray tube working at 30 kV and 500 µA and a Si-PIN detector

Full size image

The described devices allowed to analyze circular areas of the order of 0.1 mm2 corresponding to circles with diameter of about 0.5 mm. To analyze much smaller areas, the method of capillary collimators developed by Kumakhov at the end of eighties [19] was also applied to a few cases, such as the analysis of a very small gold fragment from the God’s halo in the Chapel of the Scrovegni (Fig. 10).

Fig. 10
figure 10

EDXRF spectrum of a very small fragment of God’s halo from the fresco, Chapel of the Scrovegni, painted by Giotto in Padua, analyzed with a capillary collimator located at the detector entrance

Full size image

The third revolution occurred in the first years of this century and is related to the Si-drift X-ray detector (SDD), first proposed by Gatti and coworkers in the seventies constructed by Longoni and Ketek scientists in the middle nineties [20, 21] and successively put in the market, first by Ketek and then by AMPTEK [12, 22] and by other companies.

Such a device was largely applied to study paintings [23, 24], frescos, gold, and silver, for example, pre-Columbian gold from the North of Peru [25, 26] (Fig. 11); gilded bronzes, for example a panel of the “Gate to Paradise,” a bronze door made by Lorenzo Ghiberti in Florence in 1425–1452 and brasses, such as the inscription on gilded brass at the top of the Trevi Fountain in Rome.

Fig. 11
figure 11

Portable EDXRF equipment employed between 1995 and 2010 for analysis of ancient alloys, paintings, and frescos. The equipment is characterized by an Oxford-Amptek X-ray tube with Ag-anode, working at 30 kV and 100 μA maximum voltage and current (red cylinder), and of a Si-PIN detector (rectangular box with cylinder). The MCA is incorporated in the detector box. A capillary collimator, produced by Institut für Gerätebau, was put on the detector entrance, when required. This image shows the EDXRF portable device during the analysis of a Moche pre-Columbian gold mask (Museum Royal Tombs of Sipán)

Full size image

In the meantime, almost contemporaneously to the development of X-ray detectors, also dedicated and miniaturized X-ray tubes were developed (see Figs. 7, 8, 9, 11, and 12) [12, 27, 28], and small-size pulse-height analyzers (Figs. 7, 8, 9, 11, and 12), in such a manner that, currently, a complete equipment with X-ray tube and SDD-detector (including the MCA) can be assembled, with an energy resolution of the order of 130 eV at the Fe-Kα line, a number of processed photons/s of the order of 105 and a weight of about 2 kg, including the PC or tablet (Fig. 12). Such a device allows to analyze areas of the order up to 0.01 mm2 corresponding to diameter of about 0.1 mm.

Fig. 12
figure 12

The most recent device for EDXRF-point analysis of paintings and metals. It is composed of a 40 kV, 100 µA mini-X ray tube. a Si-drift detector with 135 eV energy-resolution at 6.4 keV, and a small PC. The MCA is incorporated in the detector box (123). A typical gold EDXRF-spectrum is shown in the PC

Full size image

All the described examples are related to point EDXRF analysis, i.e., the painting under study was analyzed in many points, generally according to areas of different color or composition. Finally, around 2010, the fourth revolution: a complete scan of a painting by Macro-XRF (MA-XRF) [29, 30] or µ-XRF (MA-XRF) [31]. A mechanical system is required to this aim, to rapidly move the device in the plane of the object.

MA-XRF allows a visualization of the distribution of elements, obtained by scanning the surface of the painting with a focused or collimated X-ray beam of mm or sub-mm dimensions, according to the size of the painting. The method was first employed to test the homogeneity of a pre-Columbian nose ornament composed on gold and silver alloys (Fig. 13) and was then applied in the investigation of ancient paintings, as elemental distribution images can reveal hidden sub-surface layers, including modifications made by the artist (pentimenti) or restoration on the surface. In this way, it can provide a unique insight into the creative process of the artist and the painting’s conservation history. From that moment, scanning was largely prevailing over point analysis, in spite of the fact that, sometimes, long-time optimized point analysis can add additional information.

Fig. 13
figure 13

Nose ornament PACEB-F4-00,032 (right) on gold and silver (right) and distribution of gold, silver, copper and mercury, as deduced from EDXRF-MA scanning (figures on the left)

Full size image

In February 2020, a second “scanning equipment” developed by us (Fig. 14) scanned “La Fornarina,” a famous painting by Raphael (Fig. 15) at the Gallery of Palazzo Barberini [32] Rome, during the exhibitions of the commemoration for 500 years after Raphael’s death. The results of the scanning are shown in Fig. 16. 

Fig. 14
figure 14

Macro-scanning of the painting “La Fornarina” by Raphael; the measuring head is composed of a small-size X-ray tube working at 40 kV, 100 µA, and two Si-drift detectors

Full size image
Fig. 15
figure 15

a The painting “La Fornarina” by Raphael (painted in 1520 and located in the Galleria Nazionale di Arte antica, Palazzo Barberini, Roma); b the golden inscription RAPHAEL URBINAS on the left arm of the painting. The “gold” colored band at the border of the armband is not gold, but a Fe-based pigment, possibly yellow ochre: Fe(OH)3

Full size image
Fig. 16
figure 16

Map of elements Pb, Hg, Sr, Cu, Fe, Mn as reconstructed from EDXRF-scanning

Full size image
Fig. 17
figure 17

Sergio Mascarenhas and Roberto Cesareo hugging the bust of Albert Einstein at the “Institute for Advanced Studies” of Princeton, in the summer of 1996

Full size image

Eventually, to handle the huge number of data, new sophisticated software had to be developed since the manual inspection of millions of spectra would be impossible. This paves the way to a new and upcoming revolution. The digital revolution is where hardware has reached a miniaturization point where software advancements are more advantageous, extracting the most information possible from the data available.

2 Conclusions

During the years, energy-dispersive X-ray fluorescence (EDXRF) proved to be the ideal technique to analyze artifacts of all type, and the corresponding EDXRF equipment has had an impressive development during the last 50 years in all its instrumental components, as summarize in Table 1, i.e.,

  • X-ray sources: from radioactive sources to big X-ray tubes to air cooled, small size, dedicated X-ray tubes

  • X-ray detectors: from proportional gas counters to semiconductor Si or Ge nitrogen cooled detectors to Peltier-cooled Si-drift detectors

  • Pulse height analyzers: from big and heavy boxes to mini- electronic circuits incorporated in the detector box

  • Also the measurement modality changed versus time: from single point analysis to a complete scanning of the object to be studied

3 A look to the future

The following evolutions may be expected in the next future for EDXRF analysis in archeometry:

  • Miniaturized and dedicated X-ray tubes emitting a higher number of photons/s, allowing the scanning of a painting in µ-scanning mode and in a reasonable time

  • Si-drift detectors thicker (up to a few mm) and with larger area, to improve the efficiency of the EDXRF-device

  • A reliable software to process an enormous number of X-ray spectra, giving the relative intensity of all detected element and internal ratio (K and L-ratios)

  • A possible complementary and “portable” technique (for example X-ray diffraction or Raman spectroscopy) able to deduce the molecules associated to the elements deduced from EDXRF

The final goal would be a complete reconstruction, point by point, of the pigments employed by the artist, and, therefore, of the painting itself.

Table 1 Evolution of EDXRF-devices versus time, for analysis of paintings and metals
Full size table