PART II
Art Conservation: Culture Under Analysis BEN B. JOHNSON and THOMAS CAIRNS Conservation Center Los Angeles County Museum of Art Los Angeles, Calif. 90036
the advent of modern meth' » ods of chemical analysis, such as polarography, atomic absorption, neutron activation analysis, energydispersive X-ray emission spectros copy, laser microprobe analyzer, and spark source mass spectrome try, a great deal of attention has been focused recently on detailed analysis of art objects (1) with two objectives in mind. Firstly, there are the systematic programs of analyses of a large number of closely related objects in an attempt to fingerprint them by a good sta tistical profile, i.e., trace elements in cast bronzes. Such programs deal with research on the history of technology answering such ques tions as source of raw materials and techniques in fabrication. Sec ondly, there are specific analyses for practical purposes. This often in volves the analysis of a single sam ple or even a small group of samples to answer specific questions usually relating to treatment. In this respect the attention of the chemist is drawn to relate chemistry to authenticity. Until recently anachronisms in the use of mate rials were a good guideline to fol low, but latter-day forgers try hard to mimic the original artist and/or technique. An outstanding exam ple in the use of materials lies in the chronology of white pigment —lead white (PbO) was used since classical times; zinc white (ZnO) made its appearance around 1810; titanium white (TiO) was commer cially available around 1920. Evi dence of zinc white or titanium white in a 17th century style paint TTTITH
30 A
•
ing would certainly raise questions as to its authenticity. At this point it must be emphasized that the sam ple taken for analysis by the chem ist in the museum laboratory is from an original area and not from a recently restored area where later materials might well have been em ployed. Without a prior physical examination of a painting, for in stance, by uv, ir, or X-rays, mis takes might easily be made by a novice to the conservation—chemis try field. Frankel (2) in a recent article demonstrated the potential use of X-ray fluorescence (3) in the so-called detection of art forgeries via pigment identification. Such publications can, in this instance, give false impressions that chemis try alone tackles the job whereas interdisciplinary approaches are really necessary. In matters of au thenticity, therefore, three points finally decide the outcome: scien tific data, stylistic criteria, and con servator's experience and availabil ity of related comparative material. Instrumentation
Conservation—chemistry covers an extremely wide spectrum of ma terials for analysis. In the field of inorganic chemistry one encounters materials from metals, pigments, and stones to ceramics. On the other hand, organic materials en countered are mainly natural prod ucts, i.e., gums, glues, resins, oils. This conglomeration prompts the utilization of sophisticated instru mentation capable of handling such diverse materials. Chemists have long been aware
ANALYTICAL CHEMISTRY, VOL. 44, NO. 2, FEBRUARY 1972
of the restrictions imposed on their studies by the limits of sensitivity of many of those currently avail able analytical methods which per mit a number of elements to be de termined simultaneously in one sample. This has led to the adop tion of more sensitive methods for a restricted number of elements, e.g., neutron activation analysis of paper (4). Recently, spark source mass spectrometry has provided the chemist with the ability to cover the full range of elements in any sample in a single determination and the ability to detect those ele ments down to very low concentra tions, i.e., parts in 10e. The ΑΕΙ MS702 spark source mass spec trometer in the Conservation Center of the Los Angeles County Museum of Art is used to tackle a variety of problems encountered in the conser vation-chemistry field. Today, spark source mass spectrometry has found its way into even more new avenues of research. I t has been used alongside atomic absorption and neutron activation analysis by Morrison and Kashuba (5) in the analysis of returned lunar samples (6). Harrison et al. (7) have re viewed the forensic application of spark -source mass spectrometry in analyzing such products as hair and glass. Recently, the FBI labora tory in Washington, D.C., has ac quired a spark source mass spec trometer. The technique owes its widespread success to its unique capability of multielement determi nations at low concentrations. With such diverse materials as gums and oils, separation into ma-
REPORT FOR ANALYTICAL CHEMISTS
Conservation-chemistry represents a new and fascinating applied scientific discipline harnessing chemical knowledge to unlock the secrets of the history of technology. Particular attention is paid to ancient metallurgy and pigment and media studies
jor and minor components is of pri mary importance, followed by iden tification. To achieve this ultimate goal on small samples, a combined gas chromatograph—mass spectrom eter system must be employed. The ΑΕΙ MS902 high resolution mass spectrometer at the Conservation Center is used in conjunction with a Perkin-Elmer F - l l capillary gas chromatograph. Reasons for Analysis
Concentration on chemical analy sis of art objects is generally prompted by the curator, collector, and art historian. He appreciates the beauty and craftsmanship of an ancient object. However, he is un able to look into the historical study of technology since there are very few texts on the subject (8). Each object can be considered unique and has locked into its structure and fabrication some knowledge of the properties of the materials used. From the standpoint of his mate rials, the artist was rarely concerned about any scientific investigation into their nature—his concern was solely directed to physical proper ties. The aesthetic enjoyment of an object is greatly enhanced when its original creation is fully understood. In the case of metallic objects, tech nical analyses can often answer questions as to the source of raw materials (mines, trade), processing (smelting, cupellation), models, casting (molds, alloy composition), finishing (cold working, techniques such as chasing, incising, engrav ing, joining), and how these various stages are interrelated, if at all.
Were the processes involved indig enous achievements or results of cultural exchange? Ancient Metallurgy
Sampling. Ancient metals and their alloys are sometimes either homogeneous or heterogeneous, and truly representative sampling is dif ficult to obtain (9). In particular, the existence of more than 4 wt % of lead (Pb) in a cast bronze (Cu/ Sn) is clearly demonstrated (pol ished cross section) by the appear ance of globules of Pb of widely varying size irregularly distributed throughout the Cu/Sn matrix (10), i.e., Pb is insoluble in the Cu/Sn al loy. I t is strongly recommended, therefore, that both X-ray examina tion and, if possible, cross sections be prepared before sampling a structure. Knowledge that such a structure is under investigation aids interpretation of analyses. One of the greatest problems to day is comparing results from two different sources by two different techniques. At the moment a com parative study organized by the In ternational Council of Museums (ICOM) between a number of mu seum laboratories is underway (11). The same samples are being ana lyzed by different techniques to document standard deviations and accuracy of results by the various techniques available in the conser vation field throughout the world. For instance, Caley (12) has pointed out that the determination of tin by wet chemical methods is known to yield high values by about 10%. On the other hand, spark
source mass spectrometry boasts of ± 1 0 % accuracy for "trace ele ments" (13). With spark source mass spec trometry the question of homoge neity is a critical one since the ac tual sample size consumed during a typical analysis can be of the order of a few milligrams. The introduc tion of the ion beam chopper (se lective sampling of the positive ion beam) has greatly improved the re producibility (14) of analysis for both homogeneous and heterogene ous materials by use of the inte grating properties of the Q2 photoplate. However, spark source mass spectrometry relies heavily on the use of standard reference materials, i.e., Cu doped with various trace ele ments at known concentration levels. The few existing bronze standards are somewhat inadequate and usually too heterogeneous for use in spark source mass spectrome try. In particular, standards pre pared via a dilution technique are extremely heterogeneous (15). Trace element analysis must be treated with some caution when in terpretation is attempted. Two main sources of trace elements are the ores from which the metals were smelted and the smelting technique. To date, detailed trace element analysis by spark source mass spec trometry and neutron activation analysis is fragmentary, and any definitive conclusions are only ten tative. Historical. In an attempt to un derstand the development of metal lurgy in ancient times, it is neces sary to outline the chronological or-
ANALYTICAL CHEMISTRY, VOL. 44, NO. 2, FEBRUARY 1972
•
31 A
Report for Analytical Chemists
Figure 1. Typical Luristan bronze bar bit depicting a wild m o u n t a i n goat in profile with its head seen frontally w i t h a bird on its r u m p ( L . 2 5 6 7 . 6 7 - 2 4 1 , 12.5 χ 11.5 c m , bit 2 1 χ 1 c m , w t 6 1 3 grams)
der of events t h a t led m a n to casting and alloy formation. Admittedly, native gold and silver were among the first metals to be used, but na tive copper is reported to have been first discovered about 5000 B.C. in the Sinai Peninsula. The impor tance of copper (and of bronze) lies in the fact t h a t it was extensively used, and its corrosion in soil gen erally forms a stable patina. Per sia, rich in ores of all kinds, was logically a center of early ancient metallurgy (16). B y 4000 B.C. the discovery of the reduction of oxide ores b y smelting with charcoal took place. Copper ores employed were probably the two oxides, cuprite ( C u 2 0 ) and tenorite (CuO) and the two car bonates, malachite [CuCOvCu ( O H ) 2 ] and azurite [ 2 C u C 0 3 ' C u ( O H ) 2 ] . Carbon, coal, or coke was heated in the presence of oxygen to form carbon monoxide which in turn reacted with the ore to give molten metal and carbon dioxide. This discovery enabled man to de velop the art of smelting and cast ing. To cast a pure copper artifact presented some difficulties. Pure copper melts a t 1083°C. I t is a very sluggish viscous liquid. Such high temperatures would certainly have taxed the furnace capabilities of the artisans at this early date. However, a number of pure copper castings do exist, mainly from Northern India and Tibet but usu ally of a much later date. 32 A
•
T h e restrictions imposed by at tempting to use pure copper for di rect casting resulted in the dis covery of adding tin ore to the cop per ores during smelting. Tin was utilized from around 1800—1600 B.C. in North West Persia in the form of cassiterite or stannic oxide (Sn02). Bronze was therefore born. The addition of tin to copper in the ratio 25:75 reduces the melt ing point from 1083°to 795°C—a considerable drop. Such an alloy is more fluid and presents fewer prob lems in casting. Pure copper has the tendency to shrink greatly on cooling and thereby lose a n y fine detail of the original mold. Bronze, on the other hand, shrinks less and is more malleable and easier to cold work. Techniques of metal working in the F a r E a s t also developed as the Chinese during the Shang dy n a s t y (ca. 1523-1028 B.C.) and the Chou d y n a s t y (ca. 1027-222 B.C.) achieved the zenith of practical per fection in their bronze castings (10). Another important factor in bronze technology is the use of lead in the C u / S n alloy especially where the casting was to be extensively cold worked. T h e principal lead ore is galena, lead sulfide ( P b S ) , and might well have been known
before cassiterite and bronze since smelting of this particular ore was primarily undertaken to separate out the silver content b y a process commonly known as cupellation. T h e dross t h a t forms on the surface of the crucibles containing the melt is continually removed until a shin ing surface is obtained. This dross contained all the base metal im purities. Separation of gold and sil ver is achieved b y a modification of this process whereby salt is added to the melt to remove the silver as chloride. This so-called chlorination modification has been in prac tice since the second millennium B.C. I n this w a y both gold and silver were refined to a fairly high degree of purity. Rarely, however, was cupellation carried to the ulti mate extreme. The addition of lead, therefore, to bronze was a practical suggestion and found widespread usage in casting impart ing to the alloy desirable extra qual ities. T h e entrance of zinc (Zn) into bronze technology might well have been extremely early especially if either the copper ore or the tin ore contained a natural amount of Zn as an impurity. During the smelt ing of such ores, however, the metal lic Zn vaporizes extremely easily
Figure 2. Krishna Rajamannar Bronzes, South Indian, early 12th century. left t o right, R u k m i n i , Krishna, Satyabhama, and Garuda (M70.69.1;2;3;4.)
ANALYTICAL CHEMISTRY, VOL. 4 4 , NO. 2, FEBRUARY 1972
From
Report for Analytical Chemists
and would have been lost to the atmosphere save trace amounts. I t was not until around 800 B.C. t h a t brass ( C u / Z n ) started to occasionally appear (17). Finely ground calamine [Zn4Si207(OH)2-2H20], charcoal, and granulated copper are placed in sealed crucibles and heated. T h e reduced metallic Zn vaporizes and then alloys with the copper since the reaction takes place in a sealed system. Hence the a p pearance of substantial quantities of Zn in ancient cast bronzes would suggest questionable authenticity. Luristan Bronze Study. A carefully selected group of Luristan bronze horse bite from the Foroughi collection (18) are a t the moment under investigation by spark source mass spectrometry. Such horse trappings are in a very distinctive decorative style (Figure 1) and belong to the period during the 7th and 8th centuries B.C. T h e y have no counterparts outside their point of origin—the mountainous province of W e s t Persia. Beginning early in this century, excavations have revealed more and more of these beautiful ancient bronzes belonging to the Luristan civilization. Such prerequisites made them very suitable candidates for a study of the potential of trace element profiles in characterizing artifacts of an ancient culture (19). Drillings were t a k e n from both the cheek plates and bit of one such trapping. I n essence, to date, the results m a y be summarized briefly since the study has not yet been fully completed. T h e major constituents ( C u / S n / P b ) v a r y from piece to piece, b u t the presence of the same trace elements (Bi, Sb, Ag, Se, As, Ni, Fe, Co) were noted in each analysis within certain concentration levels. A number of the bits, however, were manufactured from native copper and are excluded from the above generalization. W o r k is continuing to achieve the number of analyses needed to draw reliable conclusions. T h e noticeable absence of zinc is worthy of mention. A number of forgeries, declared such on stylistic grounds, were also examined. T h e presence of high percentages of zinc (above 2 wt % ) together with other trace elements certainly damned them, scientifically speaking. Such
a combination of science and stylistic criteria is the most effective method in the detection of fakes and forgeries. South Indian Bronzes. Quite recently an interesting bronze group of the early 12th century A.D. depicting Krishna and his two wives, Rukmini and S a t y a b h a m a , together with his messenger, Garuda (Figure 2 ) , were acquired by the Los Angeles County Museum of Art through the generosity of Mr. and M r s . H a l Wallis. T h e occurrence of such a group of high quality and importance is unique. On religious grounds they were probably cast and fabricated (cold worked after casting the rough shape) as a group with Sutras dictating proportions, attributions, etc. This prerequisite, therefore, permitted a detailed study (20) of the chemical composition from piece to piece in an attempt to ascertain if an exact science was operative or was each composition by chance. I t turned out t h a t all four pieces resembled each other fairly closely. Drillings were taken from various locations in each piece and analyzed several times, then averaged. Table I illustrates the typical results obtained for S a t y a b h a m a (Figure 3 ) . Many other South Indian bronzes are now under study to determine whether any further correlations can be established.
Table 1. Analysis Results for Satyabhama (M70.69.3) Ht, 28 i n . ; base d i a m , 8.5 i n . ; w t , 58 1b wt
%
Hip
Neck
Heel
Av
Base
Cu
93.05
92.65
92.45
92.71
93.31
Sn
2.25
2.21
2.66
2.37
1.62
Pb
3.50
3.96
3.53
3.66
4.33
Bi
0.01
0.01
0.04
0.02
0.03
Sb
0.09
0.09
0.13
0.10
0.09
Ag As
0.08
0.08
0.11
0.09
0.09
0.022
0.026
0.04
0.02
0.032
Zn
0.08
0.11
0.09
0.09
0.05
Ni
0.50
0.45
0.55
0.50
0.40
Co
0.02
0.02
0.02
0.02
0.01
Fe
0.39
0.39
0.38
0.39
0.03
Mughal Indian Miniature Painting
Although there are several excellent studies of the technique of I n dian miniature painting which are based purely on literary sources and tradition passed from generation to generation, until present day very little has been published which deals with the technical examination of actual paintings. Chandra's treatise (21) is a wealth of information but leaves m a n y basic chemical questions unanswered, e.g., whether lead white, which is poisonous and can t u r n black in the presence of sulfur, was used by Mughal painters in the 16th and 17th centuries or whether zinc white, which was preferred more recently, was adopted. T h e common appearance of ultramarine in Mughal paintings raises a n a t u r a l doubt as to whether lapis lazuli (3Na 2 0-3Al 2 O s '6Si0 2 -2NaS), a costly stone mineral, yielded the
Figure 3. Satyabhama, one of Krishna's t w o wives (M70.69.3.)
ANALYTICAL CHEMISTRY, VOL. 4 4 , NO. 2, FEBRUARY 1972
•
33 A
Report for Analytical Chemists
Figure 4. Page f r o m a Ragamala: Todi Ragini series. Mid-18th century (M71.1.42). Wom an w i t h vina stands in grove of trees (9.5" χ 6.25")
blue or whether it was obtained fromazuritc [ 2 C u C 0 8 - C u ( O H ) 2 ] , a much cheaper material. Agrawal's paper (22) is the most comprehensive report to date on materials of Indian painting. I t is mainly a survey of historical refer ences but includes scientific exam ination of manuscripts as well. To answer some of the questions regarding pigments, a study was un dertaken (23) of the miniature paintings (Figure 4) from the col lection of Nasli and Alice Heeramanaeck (24). Pigment Identification. T h e lim ited number of pigment analyses has precluded any final conclusions at this stage, but some interesting patterns have evolved. The analy sis of the blues during the early Mughal period (mid-15th to midloth century) has revealed t h a t genuine lapis lazuli was the most preferred and most frequently used blue, i.e., lack of evidence for the suggested use of azurite as a cheap substitute. I n addition Mughal artists preferred to use their pig ments in pure form in overlapping 34 A
•
layers or in mixtures with a second pigment such as lead white. F o r white pigments Mughal art ists of the first half of the 17th century preferred lead white whereas in earlier paintings be longing to the 15th and 16th cen turies kaolin (A1 2 0 3 · 2 S i 0 2 · 2 H 2 0 ) was dominant. Vermilion (HgS) was the most widely used red pigment. Second to vermilion, minium ( P b 3 0 4 ) was em ployed, especially in paintings of the 16th and 17th century. Orpi ment (As 2 S 3 ) and realgar (As 3 S 2 ), yellow and orange, respectively, were identified on 16th and 17th century paintings but not in signifi cant patterns to be of diagnostic value. In a study of the green pig ments, both malachite and copper resinate (25) were found. A number of organic pigments were encountered and research is in progress via G C M S to elucidate their molecular structures. Media Studies
P a i n t is the mixture of a suitable medium plus finely ground pigment.
ANALYTICAL CHEMISTRY, VOL. 4 4 , NO. 2, FEBRUARY 1972
Medium is the descriptive terminol ogy applied to the binding agent or vehicle for the pigment particles. Traditionally three major classes exist which can be chemically differ entiated: plant gums (gum arabic) which are polysaccharides, glues of animal origin (size) and egg-tem pera which are proteins, and drying oils (linseed) which are mixtures of various triglycerides. Historically both gums and glues (including egg products) were used in earliest times. I t was not until the 15th century t h a t drying oils such as lin seed prevailed and quickly replaced the use of egg-tempera. Gums, however, are still used extensively t o d a y as the medium for commer cially available water colors. Gums. Starch and cellulose are the most ubiquitous of the plant polysaccharides known. Plant gums also belong to this general class as demonstrated as early as 1929 by Butler and Cretcher (26) who identified various hexoses (C 0 H 1 2 0e) as products from acid hydrolysis of gum arabic. I n es sence, plant gums are high-molec ular-weight polysaccharides built up by repeated condensation of various monosaccharides (both hexoses and pentoses). The hydrol ysis products of such gums from various botanical sources (27) are related to their taxonomic origin (Table I I ) . Besides the pigment identification of the Mughal Indian miniatures described previously, media samples were also taken (about 0.5 mg) and hydrolyzed with 3 % HC1 under vacuum a t 105°C for 24 hr (28). The hydrolysis products were then neutralized with 200 mg Amberlite I R A 68, filtered and evaporated (12 hr at 45°C) to dryness. T h e resi due was then examined by thinlayer chromatography according to Stahl (29) and later by gas chroma tography to obtain quantitative re sults. I n almost every case studied, gum arabic was the preferred medium. One or two cases, how ever, were of protein origin. Flieder (30) has successfully used this technique in the identification of both sugars (from polysaccha rides) and amino acids (from pro teins) in a large number of illus trated manuscripts. Glues.
Before the advent of dry-
Report for Analytical Chemists
Table II. Common name
Source
Composition of Various Plant Gums D-glucuronic acid
D-galactose
52
Gum arabic
Acacia senega/
16
Cherry gum
Prunus cerasus
12
21
Peach gum
Prunus persia
7
36
D-mannose
L-arabinose
Rhamnose
19
14
10
Xylose
55 43
14
Table III. Composition of Various Proteins Glues Amino acid
Casein,
%
%
Gelatin
Elastin
Ovalbumin, 64.9%
Conalbumin, 13.8%
Egg-white, % Ovomucoid, β. 2%
Lysozyme, 3.4%
Avidine, 0.1%
Glycine
2
27.25
26.7
3.05
5.7
3.8
5.7
Alanine
3.2
11.23
21.3
6.72
4.4
2.3
5.8
Valine
7.2
2.78
17.7
7.05
8.2
6.0
4.8
4.2
Leucine
9.2
3.45
9.0
9.2
8.8
5.1
6.9
4.9
Isoleucine
6.1
1.53
3.8
7.0
5.0
1.43
5.2
5.5
8.15
6.3
4.5 10.5
Serine
6.3
3.73
0.85
4.2
6.7
Threonine
4.9
2.36
1.12
4.03
5.9
5.5
5.5
4.6
Phenylalanine
5.0
2.5
6.2
7.66
5.7
2.91
3.12
5.9
Tyrosine
6.3
0.24
1.5
3.68
4.6
3.18
3.58
0.88
Tryptophane
1.2
1.2
3.0
0.3
3.6
4.9
0.51
3.8
Proline
11.3
Hydroxyproline Cystine
15.47
13.5
13.24
1.6 0.35
0.34
10.6
5.4
2.72
1.4
1.64
6.7
6.8
0.47
2.06
1.41
1.35
Cysteine Methionine
2.8
0.63
Aspartic acid
7.1
6.7
1.1
5.2 9.3
13.3
2.03
13.0
0.95
Glutamic acid
22.4
11.56
2.4
16.5
11.9
6.5
18.2
9.7
4.32
6.6
1.04
0.96
Histidine
3.1
0.7
2.35
2.57
2.15
Arginine
4.1
9.04
1.3
5.72
7.6
3.7
12.7
6.5
Lysine
8.2
4.37
0.5
6.3
10.0
6.0
5.7
6.2
Hydroxylysine
0.76
ing oils in the 15th century, glues and egg-tempera were widely used, although isolated examples exist of the use of oils as early as the 13th century. This takes the conserva tor-chemist into the realms of pro tein and amino acid chemistry. Here again the hydrolysis products (amino acids) have a direct bearing on the origin of the protein used (Table I I I ) . One can easily dis tinguish (31) casein (from milk), from animal glues (gelatin and elastin), and from egg-white. The two observed cases of protein as media in Mughal Indian miniature painting were both of animal glue origin. Drying Oils. Vegetable drying oils consist largely of triglycerides (glycerol ester of fatty acids) of five fatty acids: palmitic, stearic,
oleic, linoleic, and linolenic (Table IV). The analysis of triglycerides by gas chromatography is still somewhat in the development stage (32). To date, analysis of such molecules (mol wt approx. 950) has been achieved by transesterification of the triglyceride entity into the methyl esters of the three parent fatty acids which are then volatile enough to undergo facile gas chro matography on nonpolar columns. Table V lists the composition of a number of the more important vege table oils in terms of fatty acid con tent. The iodine value is the per centage of iodine chloride, calcu lated as iodine, which is capable of being absorbed by the oil and is a direct measure of the total amount of unsaturation in the oil. Gunstone and Padley (S3) have
recently proved that the distribu tion of the fatty acids as esters on the glycerol backbone conforms to a modified random distribution (Table VI). A novel new way of rapid and sensitive analysis of triglycerides in oils has recently been developed by
Table IV. Principal Fatty Acids in Drying Oils Name
Double Molecular bonds, formula no. Location
Palmitic
C16H32O2
0
Stearic
Ci 8 H 3 6 0 2
0
Oleic
C, 8 H 3 1 0 2
1
9
Linoleic
C I8 H 3 20 2
2
9, 12
Linolenic
CisH3(,02
3
9, 12, 15
ANALYTICAL CHEMISTRY, VOL. 44, NO. 2, FEBRUARY 1972
•
35 A
Report for Analytical Chemists
Table V. Average Component Characteristics of Some Drying Oils
Source
Oil
Perilla
Perilla
Linseed
Linum
ocimoides
Candlenut
Aleurites
Soyabean
Glycine
Sunflower
Helianthus
Linoleic
Linolenic
198
7
20
5
68
10
20
16
53
moluccana
164
13
10
49
28
132
14
23
55
8
136
11
16
74
0
hispidu annus
Wt%
333
22
332
15
331
18
330
10
322
4
321
8
320
5
310
6
Others
% Oleic
180
Table VI. Triglyceride Composition of Linseed Oil
300
Unsaturated entities palmitic + stearic, %
usitatissmum
Hites (34) on the basis that a molecular-weight distribution can simply be measured by the resultant mass spectrum of the oil, i.e., measurements on the various M+ and (M — 18) + ions. Positional isomers, however, cannot be distinguished. In spite of this, the rapidity of such measurements overcomes the other conventional and tedious methods of esterification followed by chromatography. A method to locate ultimate double bonds in triglycerides has been developed by Serck-Hanssen (35) using a rapid isothermal gas chromatographic determination of the main and end carbon chains split off as monocarboxylic acids by permanganate oxidation of the oils in acetone during a few minutes at room temperature. Autoxidation of Drying Oils. The mechanism by which conjugated species polymerize in an autoxidation process is not yet fully understood, but it is certain that a system involving the reaction of free radicals is responsible for the construction of a three-dimensional macromolecular structure (36).
Triglyceride11
Iodine value
1 11
• 3, 2, 1, and 0 refe r to the acids linolenic, linoleic, oleic, and saturated species, respectively.
Farmer (37) was the first to demonstrate that oxidation may occur at reactive methylene groups in the fatty acid entities with the formation of a hydroperoxide. In the past, classical theory had demanded that oxidation involve the production of a cyclic peroxide via the double bonds themselves. Oxidation is now regarded as a free radical chain reaction as follows involving only the reactive allylic methylene groups in both linoleate and linoleneate chains: Initiation:
R H —» R -f- H
Propagation:
R -)- 0 2 -» R 0 2 R 0 2 + RH -» ROOH + R
Termination:
2R0 2 -» R—O—O—R R + R -» R—R R 0 2 + R -> R_0—O—R
Dimerization can therefore occur between unsaturated fatty acid chains on separate triglycerides (interpolymerization) or between fatty acid chains on the same glycerol backbone (intrapolymerization). The initial number of unsaturated sites available within these triglycerides is the key to the drying process. Polymerization proceeds at room temperature via an autoxidation process involving uptake of atmospheric oxygen. Via such a mechanism a network of bonded triglycerides is formed retaining the pigment particles within the polymeric framework. In addition to this polymeric structure, some triglycerides may remain unchanged owing to lack of sufficient unsaturation to participate in the autoxidation process, i.e., palmitic and stearic. Decomposition may also occur
complete easy to use liquid chromatograph from less than $2,000
Now you can get a complete high performance liquid chromatograph at a rock bottom price. The LC4000-1 has no frills, no fancy packaging—only top quality components that give you top quality performance. It provides pulseless solvent delivery at pressures to 1000 psi. And the detector provides the highest possible sensitivities. The LC4000-1 is simple, functional, and very easy to use. It is a complete, high quality instrument that can be used in a wide variety of applications, from basic research to quality control. What's more, the modular design enables you to upgrade easily by simply adding the LC-4100 solvent delivery system, or other components from our complete line of products for liquid chromatography.
varian aerograph
2700 M I T C H E L L D R I V E WALNUT CREEK/CALIFORNIA/94598 CIRCLE 189 ON READER SERVICE CARD A N A L Y T I C A L CHEMISTRY, VOL. 4 4 , NO. 2 , FEBRUARY 1 9 7 2 • 3 7 A
Report for Analytical Chemists
(possibly with the hydroperoxide formed) with the production of a number of smaller molecules. Stolow (38) has already undertaken a detailed study of the mechanism of this polymerization process and a number of the controlling factors affecting the rate of polymerization in an attempt to relate the chemistry of such a polymerized system to real time. I t has been demonstrated successfully that such tentative correlations do in fact exist, but much more experimental data must be collected to evaluate the program statistically. Stolow (38) encountered a number of compounds by gas chromatography that he was unable to identify by normal procedures. I t is hoped that the use of the combined gas chromatograph-mass spectrometer system at the Conservation Center will help solve these problems. At the moment, however, research is underway at the Conservation Center to further the basic work on the gas chromatography of triglycerides. New Materials
It is often necessary for the conservator-chemist to study the properties of new materials not only from the point of view of their application to conservation practices but also as a potential material for use by contemporary creative artists. Such a case is polyurethane elastomers which have already been used by prominent artists such as Claes Oldenberg in his composite molded relief of the Chrysler airflow car (39). After the first limited edition was published, the polymeric material discolored badly, and studies were initiated at the Conservation Center to determine causes. Research into the basic components of the resin used revealed that an aromatic diisocyanate had been employed together with a suitable polyol. Polyurethane elastomers based on conventional aromatic diisocyanates are prone to yellow (i.e., oxidation) and lose gloss on exposure to sunlight. Replacement of the aromatic component by an aliphatic saturated hydrocarbon molecule such as hexamethylene diisocyanate removed the ability of the elastomer to undergo further re-
action after casting (Jfi). Tests on this new suggested formulation (i.e., prolonged exposure to ultraviolet) indicated that no such discoloration was likely to occur in the future. These discoveries necessitated a republication of the art work with new castings made of the more stable formulation. Role of Analytical Chemistry
Today's public consciousness of cultural heritage has elevated Conservation to a new significance in the museum world. However, lack of suitably qualified analytical chemists has somewhat hindered rapid advancement in the field. The marriage of the two disciplines, chemistry and art, has not yet been formally conceived academically although it has existed to some extent in certain talented individuals throughout the world. Rapid development of more scientific laboratories in the U.S.A. devoted to this topic is under discussion by Congress and will necessitate qualified staff which do not formally exist. Fresh manpower oriented to this new discipline and committed to the principles of Conservation will be in demand in the very near future. References (1) Symposium on "Application of Spect r o g r a p h s Techniques in the Museum Laboratory," 10th National Meeting of Society of Applied Spectroscopy, St. Louis, Mo., 1971. (2) R. Frankel, Isotop. Radiât. Technol., 8, 1 (1970). (3) R. S. Frankel and D . W. Aitken, Appl. Spectrosc, 24, 557 (1970). (4) R. L. Brunelle, W. D . Washington, C. M . Hoffman, and M . J. Pro, J. AOAC, 54, 920 (1971). (5) G. H. Morrison and A. T. Kashuba, ANAL. C H E M . , 41, 1842 (1969).
(6) K. M. Reese, ibid., 42, 26Α (1970). (7) W. W. Harrison, G. G. Clemena, and C. W. Magee, J. AOAC, 54, 929 (1970). (8) "Art and Technology—A Symposium on Classical Bronzes," S. Doeringer, D . G. Mitten, and A. Steinberg, Eds., M I T Press, Cambridge, Mass., 1970. (9) E . R. Caley, "Analysis of Ancient Metals," Pergamon Press, New York, N . Y , 1964. (10) R. J. Gettens, "The Freer Chinese Bronzes—Technical Studies. Vol I I , " Smithsonian Institution, Washington, D.C., 1969, ρ 124. (11) R. M. Organ, Archaeometry, 13, 27 (1971). (12) E . R. Caley, "Critical Evaluation of Published Analytical Data on the Com parison of Ancient Metals" in "Applica tion of Science in the Examination of Works of Art," Boston Museum of Fine Arts, 1967.
38 A • ANALYTICAL CHEMISTRY, VOL. 44, NO. 2, FEBRUARY 1972
(13) R. M . Elliott and P . Swift, Spectrosc, 21, 312 (1967).
Appl.
(14) P . G. T. Vossen, ANAL. C H E M . , 40,
632 (1968). (15) R. Brown, MS702 Users Meeting, St. Louis, Mo., 1971. (16) H . E . Wulft, " T h e Traditional Crafts of Persia," M I T Press, Cam bridge, Mass., 1966. (17) R. J . Forbes, "Extracting, Smelting and Alloying" in "History of Technol ogy," C. Singer, A. R. Hall, and E. J. Holmyard, Eds., Oxford, England, 1954, ρ 572. (18) "7000 Years of Iranian Art," Smith sonian Institution, Washington, D.C.. 1964. (19) P . R. S. Moorey, Archaeometry, 7, 72 (1964). (20) Β . Β . Johnson, "Krishna Rajamannar Bronzes : An Examination and Treatment Report" in "Krishna: T h e Cowherd King," P . Pal, Los Angeles County Museum of Art, to be pub lished, 1972. (21) M . Chandra, "The Technique of Mughal Painting," The U. P . Historical Society, Lucknow, India, 1949. (22) O. P . Agrawal, "A Study in the Technique and Materials of Indian I l lustrated Manuscripts," paper pre sented at ICOM Symposium, Amster dam, Holland, 1969. (23) Β. Β. Johnson, "The Technique of Indian Miniature Painting," paper pre sented at Symposium in Indian Art, Los Angeles County Museum of Art, to be published, 1972. (24) "The Arts of India and Nepal: The Nasli and Alice Heeramaneck Col lection," Boston Museum of Fine Arts, 1966. (25) R. D . Harley, "Artists' Pigments c. 1600-1835," Butterworths, London, England, 1970. (26) C. L. Butler and L. H. Cretcher, J. Amer. Chem. Soc, 51, 1519 (1929). (27) F . Smith and R. Montgomery, "Chemistry of Plant Gums and M u cilages," New York, N.Y., 1959, ρ 106. (28) L. Masschelein-Kleiner and F . T r i cot-Marckx, Bulletin Institut Royal du Patrimoine Artistique, Brussels, 8, 180 (1965). (29) Ε . Stahl, "Thin-Layer Chromatogra phy," Academic Press, New York, N.Y., 1965. (30) F . Flieder, Stud. Conserv., 13, 49 (1968). (31) "Traite de Biochemie Générale," Masson, Paris, France, 1952. (32) R. Watts and R. Dils, / . Lipid Res., 9,40 (1968). (33) F . D . Gunstone and F . B. Padley, / . Amer. Oil Chem. Soc, 42, 957 (1965). (34) R. A. Hites, ANAL. C H E M . , 42, 1736
(1970). (35) K. Serek-Hanssen, Acta Chem. Scand., 21, 305 (1967). (36) G. H . Hutchinson, J. Oil Color Chem. Ass., 41, 474 (1958). (37) Ε . Η . Farmer, Trans. Faraday. Soc, 38, 340 (1942). (38) N. Stolow in "Application of Science in Examination of Works of Art," Bos ton Museum of Fine Arts, 1967. (39) "Profile Airflow," Gemini G. E . L„ Los Angeles, Calif., 1969. (40) Modern Plastics Encyclopedia, 1970-71,ρ 224.
