Folder 45: Kusko, B.H. Analysis of Works of Art and Archaeology using Particle Induced X-Ray Emission with the AGLAE Accelerator at the LRMF, 1988-1989

This folder contains two copies of a typewritten final report: Kusko, B.H. "Analysis of Works of Art and Archaeology using Particle Induced X-Ray Emission with the AGLAE Accelerator at the Laboratoire de Recherche des Musees de France." Only one copy has been digitized and is presented her...

Full description

Saved in:
Bibliographic Details
Main Author: Kusko, B. H.
Format: Online
Language:eng
Created: Saint Louis University Libraries Digitization Center 1988
Online Access:http://cdm17321.contentdm.oclc.org/cdm/ref/collection/speccoll/id/1475
Tags: Add Tag
No Tags, Be the first to tag this record!
id sluoai_speccoll-1475
record_format ojs
institution Saint Louis University
collection OJS
language eng
format Online
author Kusko, B. H.
spellingShingle Kusko, B. H.
Folder 45: Kusko, B.H. Analysis of Works of Art and Archaeology using Particle Induced X-Ray Emission with the AGLAE Accelerator at the LRMF, 1988-1989
author_facet Kusko, B. H.
author_sort Kusko, B. H.
title Folder 45: Kusko, B.H. Analysis of Works of Art and Archaeology using Particle Induced X-Ray Emission with the AGLAE Accelerator at the LRMF, 1988-1989
title_short Folder 45: Kusko, B.H. Analysis of Works of Art and Archaeology using Particle Induced X-Ray Emission with the AGLAE Accelerator at the LRMF, 1988-1989
title_full Folder 45: Kusko, B.H. Analysis of Works of Art and Archaeology using Particle Induced X-Ray Emission with the AGLAE Accelerator at the LRMF, 1988-1989
title_fullStr Folder 45: Kusko, B.H. Analysis of Works of Art and Archaeology using Particle Induced X-Ray Emission with the AGLAE Accelerator at the LRMF, 1988-1989
title_full_unstemmed Folder 45: Kusko, B.H. Analysis of Works of Art and Archaeology using Particle Induced X-Ray Emission with the AGLAE Accelerator at the LRMF, 1988-1989
title_sort folder 45: kusko, b.h. analysis of works of art and archaeology using particle induced x-ray emission with the aglae accelerator at the lrmf, 1988-1989
description This folder contains two copies of a typewritten final report: Kusko, B.H. "Analysis of Works of Art and Archaeology using Particle Induced X-Ray Emission with the AGLAE Accelerator at the Laboratoire de Recherche des Musees de France." Only one copy has been digitized and is presented here.
publisher Saint Louis University Libraries Digitization Center
publishDate 1988
url http://cdm17321.contentdm.oclc.org/cdm/ref/collection/speccoll/id/1475
_version_ 1797134584977031168
spelling sluoai_speccoll-1475 Folder 45: Kusko, B.H. Analysis of Works of Art and Archaeology using Particle Induced X-Ray Emission with the AGLAE Accelerator at the LRMF, 1988-1989 Kusko, B. H. Proton-induced X-ray emission; Imaging systems in archaeology; Archaeometry; Antiquities -- Analysis This folder contains two copies of a typewritten final report: Kusko, B.H. "Analysis of Works of Art and Archaeology using Particle Induced X-Ray Emission with the AGLAE Accelerator at the Laboratoire de Recherche des Musees de France." Only one copy has been digitized and is presented here. 1988; 1989 2011 image/pdf 67 1 45_Item 0002.pdf Thomas A. Cahill Papers--Crocker Historical and Archaeological Project, 1981-2009 The items in this folder are part of the Thomas A. Cahill Papers--Crocker Historical and Archaeological Project, 1981-2009. They are from Series 1: Thomas A. Cahill Research Papers, 1981-1994. This series consists of various research papers and published articles based upon Dr. Cahill's research using Particle Induced X-ray Emission (PIXE) techniques in analyzing inks and papers. 67 1 45 Permission to copy or publish must be obtained from the Saint Louis University, Pius XII Memorial Library, Special Collections Department Saint Louis University Libraries Digitization Center text/image eng Saint Louis University Libraries Special Collections, Archives & Manuscripts / ___ L ~ 1=== " The Analysis of Works of Art and Archaeology using Particle Induced X-Ray Emission with the AGLAE Accelerator at the Laboratoire de Recherche des Musees de France B. H. Kusko Fulbright Research Scholar September 1988 - June 1989 AGLAE Laboratoire de Recherche des Musees de France Palais du Louvre 34 Quai des Tuilleries 75041 Paris, Cedex 01 The following is a melange of 1) a fmal report of my activities to J. Ligot on the progress of PIXE at the LRMF, 2) a paper for the Amsterdam PIXE V Conference, 3) a talk for the PIXE V Conference, 4) a talk at CNL for Cahill et al. about AGLAE, 5) a short editorial comment, and 6) short notes on PIXE results suitable for publication. During my trip to California (2-18 August) I will untangle the melange and present a separate report and article. I did not have enough time before I left. The article for the PIXE conference will contain an introduction, a little bit about the accelerator, measurin~ electronics, and data reduction, but mostly the results on the pl.grnents"A la Mamie" and the false Sienna. I want to show how PIXE can be useful to museum scientists, and to emphasize that PIXE complements other techniques used at the LRMF. The Analysis of Works of Art and Archaeology using Particle Induced x- Ray Emission with the AGLAE Accelerator at the Laboratoire de Recherche des Musees de France 1. Introduction During the last four years the Louvre Museum has been undergoing a major renovation. Not only is the main entrance now a giant glass pyramid, but three stories under the pyramid a team of scientists are bombarding precious works of art with a particle accelerator. The Accelerateur au Grand Louvre d'Analyse Elementaire (AGLAE) is a part of the Laboratoire de Recherche des Musees de France (LRMF), which has been given more space and equipment in a new underground laboratory. It is only recently that methods of analysis using high energy ion-beams (protons, alphas, 15N, etc) have been applied to works of art and archaeology, and usually by physicists working in nuclear laboratories in their spare time (refs 1). The Louvre Museum is thus the first museum to have an accelerator to be used exclusively for the study and analysis of works of art and archaeology. The LRMF has taken a bold leap into the future with this powerful "high-tech" approach to the conservation of our cultural heritage. The LRMF was founded in 1931 with the puropse of studying the materials and techniques used in the production of art and archaeological objects. The laboratory has always had sophisticated analytical tools and expertise in physics, chemistry, metallurgy, geology, the material sciences, as well as archaeology and art history. Presently the LRMF is located at the top of the Pavillion de Flore. There are magnificent views of Notre Dame, the Arc de Triomphe, the Pyramid, and all of Paris. In one or two years they will all move underground in a new, more spacious laboratory. (Needless to say ...) The accelerator is already working in the underground laboratory. The AGLAE team includes four physicists (Menu, Salomon, Calligaro, and Kusko) and two technicians (Simon and de Vauxmoret) from the LRMF, and is guided scientifically by G. Amsel, head of the Solid State Physics Group at the University of Paris VII. Amsel and his staff, including E. D'Artemare, E. Girard, and especially J. Moulin, have given us valuable technical assistance. Ch. Heitz and his staff at the Center for Nuclear Research in Strasbourg designed and built the multi-puropse scattering chamber. Eric Clayton wrote the PIXE reduction code we are using and has assisted us on a number of occasions. Tom Cahill of the CNL, U.C. Davis, and J.-P. Quisefit of the University of Paris gave us help for the air sampling program at the Louvre. We are collaborating with G. Grime and F. Watt of the Oxford Microprobe facility, for the p-probe analysis of painting samples and for the installation of a p-probe at AGLAE. The students we have had over the past year (doctoral and masters level) have contributed to the perfecting different aspects of the AGLAE facility (N. Brun, P. Walter, E. Le Bourhis, E. Bartel, V. Rouchon). Of course AGLAE is not separate from the LRMF, but an integral part of it. We have therefore benefitted greatly from the expertise of many at the LRMF. J. Ligot is the director of the LRMF and the driving force behind the whole AGLAE project. Our efforts over the past year have been mostly to increase our understanding of the new laboratory and to gain experience in running the accelerator, the target chambers, the radiation and charged particle detectors, the acquisition and pulse processing electronics, the computers and data analysis programs, and all the other pieces of equipment it takes to have a state-of-the-art ion-beam analysis facility. [[ Editorial sidetrack by BHK. We are often asked about the "results" we have obtained so far. Usually the questioner is expecting us to announce a startling discovery relating to some great masterpiece we have analyzed. We then have to explain that we are still in the process of setting up the AGLAE facility, and that we have preliminary results from several works of art, but that most of the time we have been analyzing well characterized "standards". We have also analyzed several samples with PIXE that were previously anal¥zed at the LRMF using such tools as scanning electron microscopy, lnfrared absorption spectrometry, x-ray diffraction, and ultraviolet emission spectrometry. The few preliminary results on works of art, including pigments, glass, metals, and textiles will be presented below. Between 1984 and 1987, the LRMF performed 47,899 analyses on 6068 objects ref 2). Of all these analyses, how many led to truly remarkable discoveries? (How many does Jack Lang know about?) Many valuable results were obtained in the field of museum conservation, including dating certain objects, discovering provenance by their trace element composition, or finding fakes--results that are important for those desiring to know more about the composition of a work of art. Now the accelerator is just another tool for the physicists at the conservation laboratory, and not a magic machine to produce dazzling results. One might then ask, is the accelerator really worth all of the time and money? In my opinion, yes. It 2 provides museum scientists with a number of IBA techniques such as PIXE, PIGE, RBS, NRA, and AMS, all of which have been shown to go a long way in characterizing works of art and archaeolo9Y. What is important is the close collaboration among the physiclsts, the chemists, the archaeologists, the curators, and the art historians--in order to pose important questions, to provide quantitative data, and to formulate correct interpretations. A real team effort is needed for the laboratory to be successful. ]] 2. PIXE at AGLAE We have chosen PIXE as the ion-beam technique with which to start. It is non-destructive, sensitive to many elements, and relatively simple to perform. Between 15 September and 30 June we acquired 386 PIXE spectra on various artifacts, on our way to developing a reliable and accurate PIXE facility. Although more work must still be done, we can now perform PIXE analyses under vacuum or with an extracted beam. Results will be presented for pigment samples, glass standards, Gallo-Roman mosaics, gilded picture frames, gold jewelry, a false Italian primitive painting, an inscribed (cotton?) Egyptian shroud, and air pollution samples collected inside the Louvre Museum. 2.1. Accelerator AGLAE is based around a 6SDH-2 2.0 Mev tandem pelletron purchased from NEe in Middleton Wisconsin. The reasons for installing an accelerator at the Louvre Museum and the special features of this machine have been described elsewhere (ref 3). The machine was designed from the point of view of using many different ion beam techniques on works of art and archaeology (ref 4). In brief, it is capable of accelerating protons from 300 kev to 4.0 Mev, alpha particles from 300 kev to 6.0 Mev, and 15N ions to 8.0 Mev. It is also capable of accelerating deuterons and 3He ions. It was necessary to include to possibility of doing proton microprobe work and accelerator mass spectroscopy (AMS). Guaranteed beam current for protons is 5 pa through a 1 mm2 collimator, and it is also capable of delivering low currents (100 pa) for PIXE type work. It was installed in the winter and spring of 1988, and the first beam was realized in June 1988. It is presently equipped with two beam lines. 3 I must add a remark about the accelerator and its operation over the past year. Unfortunately, the beams we had to work with were not very stable. It took all that J. Salomon could do to keep our count rate within reasonable limits. The problems I think are twofold: first, we do not have enough experience running this kind of complicated machine and second, the machine itself has several problems that are the fault of the manufacturer. In addition, the beam transport system of the machine was never aligned by a NEC engineer, and in March 1989 it was discovered that there is a strong coupling between the RF field of the source and the injection magnet power supply. This results in uncontrollable steering of the beam and thus unstable beam current at the target. In July an engineer from NEC will be coming to Paris for one week to see the problems for himself. 2.2. Target area A new target chamber has been fabricated in Strasbourg and was delivered to the Louvre on 3 March. This new chamber is very versatile and will be used for PIXE, PIGE, RBS, and NRA. Up to the present time the chamber has not been used, due to several problems with the electronics controlling target positioning and movement. C. Heitz, who worked for two years on the design and fabrication of this chamber will be coming to Paris in the fall, 1989 to finish the installation and perform tests. The target chamber we used this past year is an older one that was also built a Strasbourg. It has been adequate to perform tests of PIXE on small samples under vacuum. The target ladder can hold up to six samples and is moved manually. The sample surface is perpendicular to the axis of the incident beam. X-rays are detected at a backward angle of 135 degrees. The detector is collimated, and subtends a solid angle of 28 millisteradians. We tried putting the x-ray detector at 90° to the beam line, with the target at 45°. The gain in solid angle (28 to 88 msr) more than offset the loss due to an increase (by 40%) of the bremsstrahlung background. A larger solid angle reduces the incident beam current 4 needed to give 2000 counts per second, the optimum counting rate for our system. It was not until June that we became aware of the fact that the detector collimator was too large, and permitted a large background "shelf" to exist on the low energy side of characteristic x-ray peaks. This shelf is due to x-rays that are not totally converted (into an electrical signal), corning from an interaction near the edge of the silicon crystal. The DEA training "stage" of E. Barthel showed that the problem becomes very important in the analysis of metals, where there is usually one dominant peak in the x-ray spectrum at energies between 5 and 10 kev, well above the usual SEB background (up to 5 kev for 2.5 Mev protons) always present in a PIXE spectrum. Good collimation reduces this shelf by a factor of ten, resulting in better sensitivity (see figure ). (It is important not to overcollimate the detector since the solid angle is already small (- 50 msr), and a slightly smaller collimator than is required results in a much smaller solid angle.) Good collimation also reduces the FWHM resolution by 5 - 10 ev. A second Si(Li) x-ray detector from EGG-HNU (Oak Ridge, Tennessee, USA) was delivered to the laboratory in July. This will allow us to dedicate a detector for both the beam-in-air system and the vacuum system. It also gives us the possibility of using two detectors simultaneously for high-sensitivity measurements. It arrived with a small vacuum leak and will be brought to EGG-Ortec in Paris for repairs. The resolution met specifications (155 ev at 5.895 kev) , but the background below the Mn peaks was 10 - 100 times worse than for our other Si(Li) detector, which is essentially identical. A beam-chopper for measuring the incident beam current has been built at Jussieu. It was brought to the Louvre at the end of April and is in the process of being installed. Presently the beam current is determined by measuring the charge acquired by the target and target holder, which are electrically isolated from the target chamber and the rest of the accelerator. This method is fairly reliable, as long as electron supression is used to keep secondary electrons from leaving the target. Beam current can also be determined by measuring 5 the protons backscattered from a thin mylar foil placed in front of the target. This thin foil monitoring technique has the advantages of eliminating charge buildup on insulating samples analyzed in vacuum, and it can be used to monitor the beam current when analyzing samples in air with an extracted beam. 2.3. Electronics The experimental setup for the x-ray electronics during PIXE analyse is shown in figure 1. The detector is an EG&G Ortec 7900 Si(Li), with 30 mm2 area, 8.0 pm Be window, and a FWHM resolution of 147 ev at 5.895 kev. The high voltage bias is supplied by an Ortec 459 power supply. An Ortec 972 spectroscopy amplifier and an Ortec 444 biased amplifier are used together for pulse-processing and dead-time corrections. A Seiko EGG 7800 multi-channel analyzer with a Seiko 1820 ADC interface is used to collect the spectra. An Enertec 7143 linear ratemeter is used to monitor the x-ray count rate. The charge induced by the beam current is measured by a Brookhaven Instruments Corporation 1000a current integrator, and stored in an Ortec 996 counter and timer. A sophisticated circuit for determining and correcting for electronic dead-time losses according to the princiales described in (ref 5) is in the process of being installed. A special VME based module carries out in real-time the task of adding an extra beam dose to compensate for the lost pulses. We are presently working on automatic control of the accelerator and experimental systems (ref 6). 2.4. Data reduction (ref 6) (for Pixan code see mid-term report) 3. RESULTS 6 3.1. PIXE in vacuum PIXE results in vacuum will be presented on glass standards and pigments samples. Spectra were also acquired on Gallo-Roman mosaics by N. Brun, geological standards by E. Le Bourhis, and metal standards by E. Barthel, but will not be presented here. These three students have also analyzed their samples with other instruments, including the scanning electron microscope, the electron microprobe, and the ultra-violet spectrometer. A compilation and comparison of their data are presented in "Test Comparitif des Resultats Obtenus par MEB, PIXE et Spectro UV", included here as appendix 1. They show that PIXE provides better results for more elements than the SEM or electron microprobe. Their results will be discussed in section 3 below. 3.1.1. Glass standards see mid-term report 3.1.2. Pigments For a test of the feasability of the PIXE technique on painting pigments, we started with pure pigments that were formed into small pellets and thin layers. We did this to make it as easy as possible at first. If we can show that these tests are successful, we can then try analyzing pigments in paintings with the extracted beam. 3.1.2.1. Pigment introduction In the last 30 years there have been a number of studies concerning a single pigment, or several pigments used by one or a group of artists. There have also been many publications on pigments used over the centuries, based on literary sources. Very little, however, has been written about historical pigment collections, comong either from an artist of the past or a pigment dealer who left behind enough material for analysis. The Museum of Popular and Traditional Art in Paris was lucky to have acquired a collection of products for painters from the boutique 7 "A la Momie". This boutique was founded in Paris in 1712, and sold ingredients for fabricating colors and varnishes for painters of all kinds (from artists to building painters). "A la Momie" specialized in exotic materials coming from allover the world, including incense, and translucent and resinous gums (like myrrh). The boutique takes its name from "la Momie", a chemical product taken from Egyptian "mummies" and used to make brown and black pigments. In 1798 a real embalmed corpse from Egypt took its place in the storefront window, and remained there almost until the closing of the establishment. Activities stopped in 1983 on the retirement of the last owner, M. Trouveron, who thereupon gave the remaining jars of materials to the MPTA. This collection is a major sampling of materials available to artists up to 1983, and contains many old, but unfortunately undated samples. The inventory lists 132 items in labeled jars, and included pigments, colorants, gums, and resins. Among the pigments were synthesized ones, including Prussian blue, Naples (chrome) yellow, and cadmium red. The large number of samples meant a systematic study could be made. Although limited in usefulness because of the lack of dates, a study of the pigment collection "A la Momie" was undertaken, for a number of reasons. These were enumerated by S. Colinart (ref): 1. To determine the exactness of the denominations, i.e. is there agreement between the label and the contents of the jar? 2. To determine the exact composition of pigments with imprecise names like green lake, or unknown names like Basel blue? 3. Do colors correspond to pure pigments or are they mixtures? 4. Are there any additives such as mineral charges or organics? 5. Are there differences between different versions of the same pigment (Vert Milori 1,2,3,and 4) or the same pigment prepared "a l'huile", "a l'eau", or "a l'essence", to be used with oil, water, or essence of turpentine? Samples of pigments were generously given to the LRMF by Madame Jaoul, curator at the MPTA. They came in powdered or granular form, as well as "trochiques", which look just like chocolate chips. 8 3.1.2.2. Pigment results I will present the results of the PIXE analyses here. Earlier results were obtained at the LRMF by S. Colinart, J.-P. Rioux, A. Duval, and N. Bucsek, using X-ray fluorescence, X-ray diffraction, infrared absorption spectroscopy, and scanning electron microscopy. Their report gave the chemical composition and qualitative concentrations of the compounds found in the pigments. The samples were prepared for PIXE analysis in two ways, one by forming pellets (diameter 15 mm, thickness 1-3 mm), the other by applying a thin layer of powder to apiezon coated Kapton (8 pm thick) . The latter resulted in samples having areal densities between 0.5 and 2.0 mq/cm". PIXE results have been obtained for 18 blue and green pigments. An inventory is given in table 4. The results were normalized to Fe, an element which was always found in major or minor quantities, and whose x-ray yield is high. The elements were then summed to 100%, after estimating the C, N, and 0 contents, assuming the chemical composition given by the earlier study. The results on the pellets are given in tables 5 to 8, and are discussed in more detail below. 3.1.2.2.1. Blues The large variety of blue pigments in the "A la Momie" collection has allowed a study in detail of a specific pigment type. The earlier study by LRMF showed that many of the blue pigments were based on the pigment ferric ferrocyanide, Fe4(Fe(CN)6)3. Ferric ferrocyanide was first fabricated in 1710 and is known as Prussian blue. A variety of pigments similar to Prussian blue have evolved over the years, by the addition of other colorants and charges. These include Paris blue, Berlin blue, Mineral blue, French blue, and others. In addition, labels on the pigment bottles from "A la Momie" indicate that the same pigment was prepared "a l'essence", "a l'eau", and "a l'huile", depending on its intended use with essence of turpentine, water, or oil. 9 If we use the facts discovered in the earlier study, that the pigment contains ferric ferrocyanide (bleu de Prusse) and potassium ferrocyanide [K4Fe(CN)6' jaune de Prusse], and we assume that the potassium comes only from the jaune de Prusse, then we can reconstruct the exact composition of the blue pigments. The results are given in table 5. Samples 22 and 30 (bleu de Prusse and bleu de Prusse pure) contained 74% ferric ferrocyanide and 25% potassium ferrocyanide. Trace amounts (less than 1%) of Si, S, Ca, Cr, Mn, Co, Cu, and Zn were also detected. Samples 01 and 17 (bleu de Berlin and bleu mineral en grains) contained 70% ferric ferrocyanide and 29% potassium ferrocyanide. (Perhaps pigment 17 was misslabeled). Trace amounts of Ai, Si, S, Cl, Ca, Cr, Cu, Zn, Ba, and Pb were also detected. Sample 28 (bleu mineral) contained 40% ferric ferrocyanide and 5.3% potassium ferrocyanide, as well as 46% aluminum oxide (A1203) and 5.6% sulfur (calcium sulfate?). Minor amounts of Si, K, and Ca were found and trace amounts of Cl, Mn, Co, Cu, Zn, Sr, Ba, and Pb were also detected. Sample bmI is a bleu mineral (le Franc) made at a different workshop. It contained 55% ferric ferrocyanide and 6.8% potassium ferrocyanide, as well as 28% aluminum oxide and 13% phosphorus. Minor amounts of Si, K, and Ca were found and trace amounts of Cl, Mn, Cu, Zn, Sr, and Ba were detected. 3.1.2.2.2. Greens The greens proved to be very interesting pigments, shown in the earlier LRMF study to be in most cases a mixture of lead chromate (PbCr04) and ferric ferrocyanide, with a charge of barium sulfate (BaS04)' Thus they were mixtures of yellow and blue pigments, and were not derived from the more common copper-based green pigments. 3.1.2.2.2.1. Vert Milori series: 1,2,3, and 4. 10 This pigment was known to contain the same basic ingredients, but in slightly different proportions. We have been able to determine quantitatively the precise elemental composition and therefore see the chemical variations between the four "shades" of the pigment Vert Milori. The results are given in table 6. We have confirmed the use of BaS04' since the Ba/S ratio was very close to 4.28. However, the Pb/Cr ratio was high by a factor of two if one assumed the Pb and Cr were only present as PbCr04. [Perhaps the Pb is also due to lead white, 2PbC03 + Pb(OH)3]· One can almost see how one pigment evolved into another pigment. Assuming we start with the formula for Vert Milori 1, we get Vert Milori 2 by using more lead chromate and less ferric ferrocyanide. We get Vert Milori 3 from Vert Milori 2 by using more barium sulfate and less ferric ferrocyanide. Then Vert Milori 4 is obtained from Vert Milori 3 by using more lead chromate and less barium sulfate. 3.1.2.2.2.2. Other greens Laque vert (no. 16) was known from the earlier LRMF study to contain a mixture of Prussian blue, lead chromate, barium sulfate, and starch, with traces of Cu and As. We see in table 7 that it contains around :17% PbCr04 :9% Fe (FeCN6) 3 :20% Al :14% Si :19% Ca :minor amounts of S (3%), K (3%), Cu (2%), and As (2%) :trace amounts of P, Cl, Ti, Zn, Sn, and Ba. (Normally lakes are composed of Al203 + NaC04 + KAlS04 + dye.) Vert prussique Milori (no. 23) is around 55% BaS04 (Ba/S = 4.38), with 2% PbCr04 (Pb/Cr = 14.6 so the Pb must also come from another pigment-probably lead white), and 3.3% Fe (5% ferric ferrocyanide ?), as well as 5% Si, and 1% of K, Ca, Ti, and Sr. Trace elements that were found included AI, P, Cl, V, Mn, Cu, and Zn. Pigment no. 46 (zinc green, not previously analyzed at LRMF) was shown to be a mixture of around 70% zinc yellow (4ZnO·4Cr03·K20·3H20) 11 and 1% ferric ferocyanide. The ratios of Cr/Zn = 0.78 and K/Zn = 0.29 are very close to the ratios for pure zinc yellow, 0.80 and 0.30. A large amount of Na was observed (30%). (Not from NaCl). Trace elements that were found included Cl, Ca, V, Mn, Ni, Cu, Ba, and Pb. Pigment no. 51 (chrome green, not previously analyzed at LRMF) was shown to be lead based (probably PbCr04 and lead white), with 4% Ba, 3.3% S, (BaS04?)' 2.3% Fe, (ferric ferrocyanide ?), and trace amounts of Ca, Ti, Ni, Cu, Zn, and As. Pigment no. 78 (laque vert, not previously analyzed at LRMF) was a Sn based green, with PbCr04' BaS04 and perhaps Fe(FeCN6)3. There is too much Sn (Sn/Pb = 1.25) to be from lead-tin yellow (PbSn03 Sn/Pb = 0.57 or Pb2Sn04 Sn/Pb = 0.29). Trace elements included P, Cl, Ca, Ti, Co, Ni, Cu, Zn, As, and Sr. 3.1.2.2.2.3. Cinabre Vert-Jaunei Leg. Etienne The usefulness of PIXE for determining the small variations between shades of the same pigment was demonstrated on three shades of a yellow-green pigment called cinabre vert-jaune, from the workshop (Leg. Etienne). They appear to be a mixture of around 40% lead chromate and less than 0.5% ferric ferrocyanide. There is additional Pb (45%) coming from perhaps lead white, and 2% AI, coming from perhaps aluminum oxide. Traces of Cl, Ca, Mn, Ni, Cu, Zn, and As were also detected. Looking at the results in table 8 we see that the differences in shade comes from differences in Fe, K, (and Mn) content, or the amount of ferric ferrocyanide in the formula. All other major, minor and trace elements remained at the same level. 3.1.2.2.3. Pigment Conclusion The preliminary results on 18 blue and green samples suggest it would be useful to continue this study. We have seen how PIXE can be used to quantitatively determine the major, minor, and trace elemental composition of paint pigments. Subtle differences were detected between similar pigments, and pigments with different labels were shown to have the same chemical composition. We should continue until 12 we have analyzed all of the pigments and colorants we have from the collection "A la Momie". Unfortunately, the date of manufacture of the pigments is not known, so the study will be of limited usefulness. The next step is to use the beam-in-air PIXE system to measure the composition of pigments in real paintings. While this is not as easy as analyzing prepared pure pigments, PIXE has the potential to be used as a sensitive (and non-destructive) probe for the elemental analysis of paintings. This potential will be explored over the next several months. 3.1.3. Summary of student results (see appendix, "50 Millions d'Analyseurs) 3.2. PIXE in air We at AGLAE have constructed a simple beam-in-air system, based on the one used at the Crocker Nuclear Laboratory, U.C. Davis. It consists of a snout on the end of the beam line that is covered with a thin plastic (8 pm Kapton) window, through which the beam passes into the air. The target and detector geometry is the same as in the vacuum system. The target is placed 1-2 cm from the beam exit window, and the x-rays travel through 3.0 cm of air before reaching the detector. Since air attenuates the low energy x-rays, the elements Na, Mg, or Al cannot be detected. It is possible, however, to detect Mg and Al if the target and detector can be placed in a helium atmosphere. We have tried the beam-in-air system on several works of art, including a false Italian primitive painting made in the 19th century, gilded picture frames from the 18th and 19th centuries, and several ancient gold artifacts. 3.2.1. False Italian Primitive As a test of the extracted beam PIXE system, we attempted to analyze the pigments in the false Italian primitive painting, La 13 Vierge et L'enfant (a XIXth century rendition of a XIVth century Sienna), in a completely non-destructive manner. The experiments took place on 14 June, 1989, and were assisted by J.-P. Rioux, L. Faillant- Dumas, and A. Jouan. The AGLAE crew consisted of J. Salomon, E. Barthel, V. Rouchon, and B. Kusko. Great care was taken to insure total safety for the painting, by using an extremely low beam current (200-300 picoamperes), cooling the impact point with a jet of air, and limiting the analysis times to less than 5 minutes. [Accumulated charge was therefore only 30 nanocoulombs.] The results were successful. Absolutely no damage was observed and good spectra were acquired for six colors and the bole. Unfortunately the bole spectra was lost while transferring it to the SUN workstation, so we have no quantitative results for its composition. (As I recall from the raw spectrum it contained Zn, Fe, and Ca, perhaps Ba, as well other trace elements.) This makes it difficult to determine exactly the composition of the pigments, since the proton beam usually passes through the pigment and excites x-rays to be emitted from the bole and the pigment simultaneously. A preliminary pigment composition was determined by the computer program PIXAN. The beam current was not measured, so the premiminary results were summed to 100%, after including reasonable amounts of carbon and oxygen (elements not seen by PIXE yet present in major amounts). The final quantitative results are presented in tables 1 and 2. Approximate limits of detection are included, and experimental uncertainties are on the order of 10-20%. 3.2.1.1. Discussion of false Sienna results The following were some notes and preliminary comments about the results obtained. All I knew of the previous anaysis when I wrote this discussion was that it was determined to be a fake made in the 19th century. This was confirmed immediately by the presence of large amounts of barium, probably in the form of barium sulfate, which was not used as a pigment before the 1780's. [However, a natural form exists that 14 was described as early as the 16th century.] R.L. Feller, in his article on barium sulfate, in Artists' Pigments-A Handbook of their History and Characteristics, states that BaS04 can be found in three forms of a white pigment in a painting; as a natural mineral, as a synthetic pure substance, or as a synthetic "lithopone" mixture--(ZnS + BaS04) . Since zinc was also found in every analysis it is very likely that it comes from a lithopone layer underlying the paint layer. We found a form of lead antimonate yellow. Although not used as a pigment in paints until around 1625 (known then as Naples yellow), it had been known since 1500 BC, when it was used in Egypt to color glass. Its composition can be Pb3(Sb04)2 or PbO'Sb20S or even Pb2Sb207' Since we found a Pb/Sb ratio of 2 + 0.2, perhaps the yellow was a misture of the first and the third. The red was shown to be mercury based, perhaps cinnabar or vermillion (HgS). Since we do not have an analysis of the bole it is hard to determine the sulfur content of the pigment. It is interesting to note that iodine was found in the three red analyses, as well as the blue. [There is a red pigment, iodine scarlet, HgI2 (Hg/I = 0.79) available since the 18th century, but the Hg/I ratio for the red analyses ranged from 3.4 to 12.4. Perhaps the iodine comes from a glue component of the varnish made from fish?] The red and yellow pigments were analyzed in two places, to test both the precision of PIXE and the homogeneity of the pigment composition. The results show that the bole has a varying influence on the results (thinner pigment layer -- more bole x-rays) . One previous analysis of the red pigment was made--during the press conference of 31 January 1989, (when the painting was damaged) . It was damaged because of carelessness on our part, since the beam of several nanoamperes was striking the same spot for over 30 minutes. (The paint and varnish layer was burned by the heat generated by the 15 proton beam. If we work with beam currents and analysis times well below (xl00) the threshold for damage, PIXE can be a totally non-destructive method of analysis.) The spectrum we acquired at that time, however, contained many more peaks from the trace elements than the spectra acquired on 14 June, because the increased charge resulted in a 10 to 20 times increase in sensitivity. The results are included in table 2. 3.2.1.2. Comparison with previous analyses After the PIXE results were written up I found a summary of the previous results of analyses of the False Sienna performed by the LRMF (La Vie Mysterieuse ... pg 69 - 71.) In general the chemical results are in good agreement, except for the composition of the red pigment. The earlier analysis found it to be cadmium based, while the PIXE analysis found it to be mercury based. 3.2.1.3.False Sienna Conclusion It is necessary to reanalyse this false Sienna, in order to obtain quantitative information on the bole before more definitive results can be obtained. We can try to increase sensitivity. We should also try pigment homogeneity tests, by analyzing one or two pigments in four or five places. (We should try repeating several analyses to test precision of the PIXE method on paintings.) V. Rouchon has suggested we take micro-photographs (x 5-10) of the point of beam impact both before and after a PIXE analysis. We can even take photographs again after one week, one month, three months, etc, in order to see if any long term damage occurs. 3.1.2.1.4. False Sienna Appendix A quick analysis of the false Sienna was performed on 28 July, 1989, as part of the preparation for making a film about AGLAE by the CNRS. We were able to obtain another spectrum of the bole, and also one of the gesso. In addition, the red was analyzed with the gain set to see the K lines of cadmium. 16 An analysis of the bole (or the thin layer below the pigment layer) showed it to contain 30% barium, 17% zinc, 9% iron, 9% calcium, and 7% sulfur. It may be a mixture of the synthetic lithopone mentioned by Feller (with equal amounts of ZnS and BaS04), and bole made from natural iron earths. The analysis of the gesso (?), a brown substance sticking out from the left side of the painting showed it to contain 40% calcium, 20% barium, 19% sulfur, and 8% zinc. The red was again found to be mercury based (Hg time cadmium was detected at the 2% level. 26%), but this 3.2.2. Egyptian Shroud An Egyptian shroud inscribed with writing was analyzed on 26 June. Textiles are fairly easy to analyze with the external beam PIXE system, and fortunately, many minor and trace elements are found that can be used in a study of provenance, methods of production, dyes and pigments, etc. It is hard, however, to get good quantitative results on the pigments and dyes used to color or inscribe the textiles. This is because x-rays are detected simultaneously from the colorant as well as the textile substrate. One has to make estimates of the amount (areal density) of colorant on the textile in order to derive quantitative numbers for its concentration. We made four analyses of the shroud, two of the inscription (brown) and two of the adjacent blank textile. Figure shows the x-ray spectrum of the inscription (and cloth). The results are given in table ,using calcium as an internal standard (Ca=100). [The assumed composition of the textile was 50% carbon, 20% oxygen, and 0.5% calcium.] The results are plotted in figure . PIXE was able to detect minor and trace amounts of silicon, sulfur, chlorine, potassium, calcium, manganese, iron, copper, and zinc. The writing clearly contained iron, and it may contain sulfur and manganese. It is not possible to age-date the writing with this information, it is probably mineral in nature (and could be very old.) Since we do not know how much ink is on the cloth, it is impossible to get absolute 17 numbers for the amount of iron in the ink. There may, of course, also be organic elements in the ink that can not be determined by PIXE. Although we used low beam currents (- 1 na), we observed that the proton beam induced changes in the inscription, turning the ink from brown to black at the point of impact. Since the cloth was undamaged under similar irradiation conditions, the changes are probably due to chemical changes in the ink and not damage due to sample heating. It was hoped that the change would be reversible, having only a temporary effect, but this was not observed. Although they faded a little, the marks were still visible several days after the analysis. 3.2.3. Gilded picture frames See report by V. Rouchon. The results by PIXE have been hard to interprete, since the recuction code PIXAN assumes the sample is homogeneous, and gilded frames are made of up of thin layers (gold, bole, gesso, substrate) as shown in figure 3.2.4. Gold jewelery using a selective zinc filter Results to date have been incorrect. For some reason I am having problems running PIXAN with a zinc detector filter. 4. Air Pollution at the Louvre Museum The first and most extensive application of the PIXE technique has been the analysis of atmospheric aerosol samples. The reason is based on the capabilities of PIXE as well as the nature of fine particles in the air. Information is required not only about their chemistry, but also about their size. Particle size governs transport and removal mechanisms, lung capture, and light scattering, as well as being an important way to pinpoint the sources of the pollution. The simultaneous measurement of composition and size requires the collection of aerosols by impaction or filtration methods. Such 18 devices can only collect a few monolayers of particles before slzlng errors occur due to bounce-off or clogging. This results in a sample of around 10 pg of total mass per size range, which poses serious problems for most chemical analysis methods. These samples, however, are ideally suited for analysis by ion-beam techniques, notably PIXE and forward alpha scattering techniques (FAST). PIXE is used to detect the elements from sodium through the end of the periodic table, and FAST is used to detect the elements fron hydrogen to fluorine. Therefore all the elements from hydrogen to uranium can be detected on a single filter substrate by using these two methods (ref 7)). For three weeks during June 1989, we collected aerosol samples in the grand gallery of the Louvre Museum. We used a (FSU) streaker sampler, by which air is pulled through a nuclepore filter with a nozzle (lmm x 4mm) that moves slowly across the filter at a rate of 1 mm per hour. The resulting band was 4 mm wide and 168 mm long, giving us time resolved air samples. During a PIXE analysis we collimate the beam to 2 mm, which gives us a time resolution of 2 hours. The most striking immediate results were obtained by just looking at the bands. They were almost black for the entire week, except for a nearly clean stretch which went from Saturday (varying times) to Monday morning at 7 am. This puzzled us at first, because Sunday is usually the most crowded day in the Museum. A likely explanation is the fact that the black color is due to carbon on the bands, which comes from the large number of trucks in Paris. On Sunday the truck traffic is drastically reduced, resulting in much less carbon in the air. Apparently the air inside the Grand Gallery of the Museum mixes almost immediately with the outside air. The filters have not yet been analyzed by PIXE, it is necessary to first build a target collimator and easier way to move the target holder. Each strip will be analyzed in 84 places, glvlng us PIXE results for every two hours over a period of seven days. 5. Conclusions 19 We have seen that PIXE can provide quantitative information the major, minor, and trace elements for a wide range of materials. We have made a good start, but much work still remains to be done. The most important task right now is to install the beam-chopper and beam monitoring foils. These methods for measuring the beam current are essential if we are to get absolute measurements with PIXE. The beam chopper will introduce some complications in the measurement of electronic dead-time, and it will take some time to work them out. The new target chamber needs to be tested. The beam-in-air system, which has already been tested on paintings and gilded frames, must be perfected before reliable quantitative results can be obtained. A fast valve needs to be installed, and a laser alignment system must be designed and built. Tests should also be made in a helium atmosphere. We must also carry out tests with selective filters, in order to improve the minimum detectable limits for samples composed of one dominant element. A two-detector system should be set up for further improvements in minimum detectabme limits. PIXE should be coupled to PIGE and RBS measurements for a more complete analysis. In my view the best place for PIXE in the LRMF is at the beginning of any analytical approach. Since PIXE is non-destructive, sensitive, and multi-elemental, it can be used to determine the major, minor, and trace elemental composition (Na through U) of a work of art under study. This can lead the wa y to a more refined analysis of compounds and organics. These can be performed with a good knowledge of the sample, and can avoid redundant or unnecessary testing. It is then unlikely that an analysis will miss a major inorganic component, as in the chemical analysis of the red pigment in the false Sienna. Cadmium red was discovered but the larger amount of mercuric sulfide (vermillion) went unnoticed. 6. Acknowledgements (See Introduction, paragraph 3.) 7. References 20 1. Proceedings of the International Workshop on Ion Beam Analysis in the Arts and Archaeology, Nuclear Instruments and Methods, B14, (1986), 1-167. 2. Rapport d'Activities, 1987, LRMF. 3. M. Menu; IBA in the Museum: Why AGLAE; invited talk, IX International Conference on Ion Beam Analysis, June 1989, Queens University, Kingston, Canada. 4. G. Amsel, Menu, M., Moulin, J., and Salomon, J.; The 2MV tandem pel letron accelerator of the Louvre Museum, presented at the IX International Conference on Ion Beam Analysis, June 1989, Queens University, Kingston, Canada. 5. Amsel, and E. Girard; A simple automatic deadtime loss correction system for particle counting processes induced by ion beams with fluctuating intensities; , 218 (1983) 16-20. 6. M. Menu, Calligaro, T., Salomon, J., Amsel, G., and Moulin, J.; The dedicated accelerator based IBA facility AGLAE at the Louvre; presented at the IX International Conference on Ion Beam Anaysis, June 1989, Queens University, Kingston, Canada. 7. Cahill, T.A., R.A. Eldred, D. Shadoan, P.J. Feeney, B.H. Kusko, and Y. Matsuda. Complete elemental analysis of aerosols: PIXE, FAST, LIPM, and mass. Nuclear Instruments and Methods, B3 (1984) 291-295. 21 nOli juillet 1989 Comite de Redaction: Nathalie BRUN Veronique ROUCHON Etienne BARTHEL Eric LE BOURHIS 50 Millions d'Analyseurs Test comparatif des resultats obtenus par M.E.B., P.I.X.E. et spectro. U.V. Nous presentons ici des resultats d'analyses obtenus en [uin 1989 (sauf mention contraire) sur differents types de materiaux par les trois methodes: M.E.B., P.I.X.E. et V.V. prix: 1 rouble 30 kopecks 1) Standards geologiques On dispose ici : - des concentrations annoncees par Geostandard Newsletter. Ce sont les compilations de mesures obtenues par differentes techniques. - des resultats des mesures par P.I.X.E. effectuees par Eric LE BOURHIS pendant son stage de D.E.A. - des resultats de mesures M.E.B. realisees par Veronique ROUCHON pendant son stage d'option. Granite MAN: ------------------------------------------------ Ele Na Al Si P S Cl K Ca Mn Fe tab. 4.33% 9.33% 31.1% .606% 100 140 2.63% .421% 310 .332% MEB. 1.48% 9.76% 32.8% .510% 4.59% .870% .92% PIXE 1MeV 7.63% 11.7% 32.9% .710% 65 650 eta. .509% 405 .334% eta. designe l'etalon interne choisi. tab. designe les va leurs tabulees. PIXE 1.5 MeV 3.14% 10.5% 30.6% .550% 89.3 491 eta . .380% 273 .308% Diorite DRN: Ele ------------------------------------ Na Mg Al Si K Ca Ti Mn Fe Feldspath FK-N: Ele Na Al Si K Glauconite GL-O: Ele Mg Al Si K Ca Fe tab. 2.21% 2.65% 9.27% 24.6% 1.41% 5.03% .653% .170% 6.78% tab. 1.91 % 9.85% 30.3% 10.6% tab. 2.68% 3.99% 32.9% 6.59% .641% 13.7% MEB. 1.20% 1.70% 9.55% 25.1% 1.81% 5.99% .810% .250% 8.07% MER .160% 9.17% 31.4% 12.7% MEB. 1.69% 3.42% 25.3% 7.64% .880% 17.8% PIXE 1 MeV 4.62% 11.9% 28.4% eta. 3.62% .720% .190% 7.11% PIXE 1MeV 2.86% 12.4% 35.5% eta. PIXE 1MeV <3.89% 5.08% 26.9% eta. .680% 14.2% PIXE 1.5 MeV .319% 3.89% 11.3% 26.5% eta. 4.50% .629% .175% 6.68% PIXE 1.5 MeV .94% 13.4% 35.7% eta. PIXE 1.5 MeV 3.72% 5.81% 28.4% eta. .600% 14.3% Biotite MICA-FE: Ele -------------------------------------------------- Mg Al Si K Ti Fe tab. MEB. PIXE 1 MeV 4,49% 12.3% 18.1% eta. 1.54% 18.8% PlXE 1.5 MeV 5.62% 13.3% 18.8% eta. 1.54% 18.7% Le PIXE, comme le MEB, a des problernes de detection des elements legers ( Z inferieur a 13 ). Pour les elements plus lourds, on constate un meilleur seuil de detection et une meilleure precision aux faibles concentrations avec Ie PIXE. 2.74% 10.3% 16.0% 7.26% 1,49% 17.9% 1.77% 9.31% 16.9% 8.03% 1.69% 21.6% On a ici: 2) Echantillons base-cuivre: - les analyses cffectuees antcricurernent par Loic HURTEL par spectrometric U.V. - les analyses faites au MEB par Veronique ROUCHON au cours de son stage d'option, - les analyses par PIXE realisces par Etienne BARTHEL au cours de son stage d'option. echantillon BV: Al Mn Fe Ni Cu Zn As Sn Sb Pb UV MEB .125 .28 .375 .26 83.6 87.3 3,48 4.28 9.73 8.41 1.33 1.01 PIXE .44 ,47 .23 81.8 3.4 .11 10.6 1.13 .88 avec correction de tailing .41 ,43 .21 83.9 3.1 9.88 1.05 .82 echanrillon CA UV MEB PIXE Al .058 Mn .17 .25 Fe .03 .14 Ni .335 .51 Cu 62.2 63.4 63.2 Zn 34.26 35.2 33.8 Sn .67 .45 .98 Sb .11 Pb .96 .54 .97 echantillon 40 UV Fe .03 Ni .10 Cu 86.3 Zn .40 As .056 Sn 12.8 Sb .14 Pb .14 MEB PIXE avec tailing .028 .026 .054 .050 84.6 85.8 .47 .43 .043 .040 14.6 13.5 87.6 12.4 .21 .20 Le MEB est moins sensible que le PIXE et la spectro UV: son seuil de detection est plus eleve. 3)Standard de verre. ? No-.-i ,5; 2. On dispose: 0 ~ cJ: jVAil -c"" J' -'l-~\L-t 1',+ v - ., 0 -des mesures par PlXE obtenues par Bruce KUSKO. L- \ -I to -des analyses au MEB realisees par Alain DUVAL en 191)7. -des valeurs annoncees par Ie fournisseur CORNING qui sont des compilations de mesures de differents laboratoires. -des analyses avec la microsonde du laboratoire de petrographic de Jussieu realisees par Nathalie BRUN. Brill A: Na Mg Al Si PS CI K Ca Ti Mn Fe Cu Zn Sn Sb composition du standard 1%( 10.7 1.66 .60 31 .06 .06~ .10 2.39 3.78 .48 .79 .75 .97 .03 .16 ;.~ c~ Co u;:. 240 PIXE 12 1.5 .53 29 .31 .19 .10 (~)/C) 3.7 .53 .85 .80 1 .04 .33 .07 ?~oo MEB 6.5 1.1 1.4 32 .13 .16 2.9 5 .9 1.1 1.05 1.4 Microsonde 10.3 1.57 .42 32 2.4 3.9 .46 .70 .95 1.05 4) Echantillons de verre opaque. On dispose des mesures par PIXE, par microsonde et MEB toutes effectuees par Nathalie BRUN. echantillon 18444. PIXE MEB Microsonde Na 16.7 7.0 7.25 Mg .25 .31 Al 1.28 1.14 1.22 Si 27.3 31 33.8 P .39 5 .39 .35 .15 Cl .79 .6 K .35 .52 .55 Ca 3.95 5.15 3.95 Ti .036 Mn .095 .15 Fe .41 .65 .35 Cu 3.17 4.52 3.66 Zn .006 Sn <.07 .18 Sb 3.59 4.6 2.85 Pb .2 .19 echantillon 19028 PIXE MEB Microsonde Na 6.1 10.1 Mg .8 1.02 Al 1.2 1.12 Si 24.7 24.5 P <.3 5 < .1 .11 .30 Cl .95 .85 K 1.3 1.35 1.4 Ca eta. 5.7 5.0 Ti .10 .17 .07 Mn .90 .27 .18 Fe .92 1.11 .85 Cu 9.25 12.3 10.7 Zn .055 Sn 1.15 1.53 .69 Sb .5 .4 Pb 5.4 6.6 6.3 , Certains ecarts pourraient etre dus a l'heterogeneite des verres opaques, la microsonde ne prenant en compte qu'un volume limite. L'analyse de l'etain et de l'antimoine est difficile a cause de la presence d'une grande quantite de calcium. http://cdm17321.contentdm.oclc.org/cdm/ref/collection/speccoll/id/1475

Similar Items