JOURNAL OF ELECTRON MICROSCOPY TECHNIQUE 14~1-5 (1990)

Preparing and Analyzing Fractured Archaeological Fibers ALLE V ANGEL AND KATHRYN A. JAKES Center fo.r Advanced Clothing, Ohio State

Ultrastructural Research, University of Georgia, University, Columbus, Ohio 43210 (K.A.J.)

Athens, Georgia 30602 (A .A .) ; Department of Textiles and

KEY WORDS

ABSTRACT A technique was developed to prepare archaeological fiber cross sections for electron microscopic examination and x-ray analysis. Use of this new method allows chemical and morphological information to be obtained from the interior of a single fiber or yarn. Fibers are fractured while frozen and then freeze dried. Following mounting and carbon coating, fibers are examined by scanning and backscatter electron microscopy and then analyzed by using energy- dispersive spectrometry. Elemental distribution is mapped by using image-processing software. In this report, the described technique is employed in the examination of ancient fibers from three different long-term storage environments (moist buried, dry buried, museum stored). Data obtained by examining the interior of fibers such as these provide insight into the conditions of a fiber’s growth, the treatments applied during the fiber’s processing and use, and the conditions in which the fiber was stored.

Scanning electron microscopy, Microanalysis, Elemental mapping

INTRODUCTION In the past, information describing archaeological

materials and their sources was gained by physical examination. Fibers were identified by microscopic appearance; yarn and fabric construction were used to interpret technological advancement of the fabric man- ufactui-ers; and the design on the textile was used to infer its cultural significance to the textile producers. But fibers contain much more information than that which a gross physical examination can provide, and in addition. such examination often is limited in value due to changes in morphology caused by degradation. Archaeological and historic textiles contain a record of the subtle interactions which have occurred between them and their long-term storage environments, as well E s containing information concerning their growth conditions and any treatments applied throughout their lifetimes. Application of chemical and physical methods of analysis can contribute informa- tion concerning the biologic (i.e., growth conditions), systemic (i.e., treatments and use by people), and diagenetic contexts ke . , burial conditions) of materi- als. Thus, Koestler et al. (1985a,b) and Indictor et al. (1985) have employed energy-dispersive x-ray spec- tromet-y to determine which mordants were used on modern and historic textiles, and Jakes and Sibley (1984, 1985, in press) have reported x-ray microanaly- sis data in their examination of pseudomorphs and “mineralized or partially mineralized” textiles. Hoke and Pe:rascheck-Heim (1977) and Hardin and Duffield (1986) report x-ray microanalysis data in their exami- nation of metal wrappings of yarn from historic tex- tiles. Other chromatographic and spectroscopic tech- niques have been used in the identification of dyes on ancient textiles (Saltzman, 1978; Schweppe, 1986; Walker and Needles, 1986).

This report describes a technique which applies scanning and backscatter electron microscopy in com- bination with energy-dispersive spectrometry (EDS) to

characterize the physical structure and elemental dis- tribution in cross sections of ancient fibers. Small, fragile fibers are prepared for analysis by freezing, fracturing, and then freeze drying. Fracturing the fiber exposes a face which is examined for its elemental distribution. By the use of this technique, data are obtained from the interior of a single fiber. Thus information is learned about the composition of the whole fiber rather than of its surface alone, as would be the case in analysis of longitudinal mounts of fibers.

MATERIALS AND METHODS Samples from three environments were prepared for

analysis:

1. Bast fibers buried in a moist, oxidizing, claylike environment obtained from mound C, Etowah, Geor- gia, dated approximately 1200 A.D.;

2. Camelid hair fibers buried in an extremely dry, sandy environment obtained from a Paracas, Peru, mummy bundle dated 150-500 B.C.;

3. Metal-wrapped silk yarns from Turkish velvets dated approximately 15th century. These samples were museum stored and not buried.

Fibers were rapidly frozen on a precooled brass block maintained a t liquid nitrogen temperature by immer- sion in liquid nitrogen. Fibers were cross fractured several times on the brass block with a precooled razor blade. The block and fibers were then quickly placed in

Received March 16, 1988; accepted in revised form February 2, 1989. Address reprint requests to Kathryn A. Jakes, Ph.D., Department of Textiles

and Clothing, 1787 Neil Avenue, Ohio State University, Columbus, OH 43210- 1295.

Allen Angel’s current address is Cavin Analytical Consultants, Inc., 2165 West Park Court, Stone Mountain, GA 30087.

(c) 1990 /.LAN R. LISS. INC

2 A. ANGEL AND K.A. JAKES

a high vacuum evaporator where the fibers were al- lowed to freeze dry overnight. The fractured dried fibers were mounted with pure carbon paste in 1 mm holes drilled into carbon stubs or planchets. Each sample was oriented in an upright position with at least 0.5 mm of the fiber extending out of the hole. The fractured plane was positioned parallel to the stub surface. Samples were then coated with approximately 45 nm of carbon. A Philips 505 SEM equipped with four scintillator-type backscatter electron detectors and a Tracor Northern energy-dispersive spectrometer were used for the analysis. Samples were initially examined in the secondary electron mode and then in backscatter mode. If considerable z-contrast was noted, several 50 nm spot analyses were made from points a t the periph- ery and the center of the fractured fiber face. Analysis time varied from 50 to 500 seconds with most spectra being collected under the following conditions: 20 kv, 80 pA emission, 50-100 nm spot size and a 0" stage tilt which provided an x-ray take-off angle of 25.5'. Follow- ing the collection of spectra, x-ray elemental maps were obtained by using Tracor Northern "IPP' soft- ware with dwell times of 0.02-0.04 seconds and a map resolution of 256 by 256 pixels.

RESULTS AND DISCUSSION Individual fibers were easily isolated from yarns

with fine forceps while observing them under a dissec- tion scope. Single metal-wrapped yarns were sectioned as units. Fracturing the fibers and yarns after contact freezing on a liquid-nitrogen-cooled block provided many nearly flat, cross-sectioned surfaces. Although not all fiber types became brittle when frozen, compres- sion or other mechanical damage typically caused by sectioning fibers a t room temperature was minimized. In addition, freeze drying reduces surface condensation and possible elemental translocation during the drying process. To provide optimal geometry for the collection of x-ray photons and for accurate elemental mapping, only those fibers which revealed fairly flat cross- sectioned surfaces were selected for analysis. At least five fibers of each type were examined and analyzed. Not all fibers fractured smoothly, but it was assumed that those fibers selected for analysis are representa- tive of all of the fibers in a yarn bundle and that all of the fibers experienced the same conditions throughout their lifetimes except for some possible variations in growth conditions.

Scanning electron imaging of fractured bast fibers (Fig. 1) demonstrated many nearly flat, porous faces. X-ray analysis revealed Al, Si, P, S, K, Ca, Fe, and Cu. Some fibers contained C1 and Ti. With backscatter electron imaging (Fig. 21, compositional differences were evident as demonstrated by high signal intensity a t some peripheral areas of the fiber. This increased z-contrast is due to the concentration of elements of higher atomic number at or near the surface of the fiber, a consequence of the moist oxidizing environmen- tal conditions of the fiber burial site. Under these circumstances, soil elements such as Fe and K will not dissolve in the water permeating the site; rather, these elements will be transported as microparticulate mat- ter and will be deposited on the fiber surface (Jakes and

Fig. 1. Scanning electron image of the fractured face of an Etowah mound, Georgia, bast fiber, dated approximately 1200 A D Bar = 100 I*m.

Backscatter electron image of fractured bast fiber face displayed in Figure 1. Note increased signal intensity at periphery of fiber face. Bar = 100 wm.

Fig. 2.

Angel, 1989). The x-ray maps (Fig. 3B,D,E,G,H) of these cross sections also show clumping of Fe, K, P, Si, and Al, respectively, on the surface of the fiber due to this soil deposition. Penetration of the fiber by copper, on the other hand (Fig. 3A), is indicative of the inter- action of the fiber with copper metal in the burial site. In a suitable environment, fibers can be mineralized, forming a fossillike pseudomorph (Jakes and Sibley, 1985). Calcium in the fiber (Fig. 3C) reflects the calcium oxalate crystals which formed during fiber growth. The source of the sulfur found throughout the fiber (Fig. 3F) is unknown. It is unlikely that the sulfur signal originated from the carbon paste, but it is remotely possible. Spectra collected from carbon paste on a graphite planchet displayed appreciable contin- uum and only a small chlorine peak. To examine the source of the x-ray signal in elemental maps, the entire energy spectrum (0-10 keV) was defined as a region or window for elemental mapping. X-ray analysis of the soil and copper samples from the same site reveal no sulfur content, but the corrosion layers of the copper metal objects do contain copper sulfide corrosion prod- ucts. The sulfur may be derived from other hair or feather fiber in the burial, or from some sort of textile treatment, but many more samples need to be exam- ined before a determination of source can be made.

Figure 4 displays a scanning electron micrograph of a camelid fiber from a Paracas, Peruvian mummy bundle dated 150-500 B.C In contrast to the bast fibers buried in a wet, oxidizing environment, this fiber was buried in a very dry limestone-type soil. Corresponding spectra collected by analyzing the fiber edge and its center indicated that calcium, potassium, and silicon were concentrated at the edges of the fiber. Elemental

PREPARING AND ANALYZING FIBERS 3

Fig. 3. X-ray maps of the bast fiber in Figure 1 (bar under each map = 100 pml: A, copper; B, iron; C, calcium; D, potassium; E, phosphorus; F, sulfur; G, silicon; H, aluminum.

distritution maps in Figure 5B-D verify a clumped distribution of K, Si, and Ca, respectively, on the surfact? of the fiber. No transport of these elements to the inside of the fiber has occurred because of the environmental conditions of the burial. In this case,

Fig. 4. Scanning electron image of fractured camelid fiber ob- tained from a Paracas, Peru, mummy bundle dated 150-500 B c Bar = 10 pm.

Fig. 5. X-ray maps of the fracture fiber face displayed in Figure 4 demonstrate sulfur (A) throughout the fiber and potassium (B), silicon (0, and calcium (D) at the periphery of the face. Bar = 10 pm.

sulfur was found throughout the fiber (Fig. 5A), which can be attributed to the sulfur-containing protein structure of the hair fiber.

Figure 6 displays a scanning electron micrograph of a metal-wrapped yarn from a 15th-century Turkish velvet which was stored rather than buried. Backscat- ter electron imaging of this fiber demonstrates the intensity difference between the core yarn and the wrapping (Fig. 7) as well as some surface impurities.

X-ray spectra collected via spot analyses of both the metal wrapping and the yarn (Fig. 8) revealed an in- tense Ag peak and s, Cu, and Au. Spectra (Fig. 9) col- lected from the yarn matrix display a reduced Ag peak and the presence of Si, S, and Ca. Elemental analysis

4 A. ANGEL AND K.A. JAKES

Fig. 6.

Fig. 7.

Fig. 6.

Scanning electron micrograph of a metal wrapped yarn

Backscatter electron image of the yarn in Figure 6. The

X-ray spectrum collected by spot analysis of the metal

Fig. 9. In contrast to Figure 8, few x-ray spectra collected from the central area of the Turkish yarn displayed intense silver peaks.

Fig. 10. X-ray maps collected from metal wrapped yarn face displayed in Figure 6 and 7. Bar under maps = 100 Fm. Iron (A), silver (B), and gold (C) are confined to the periphery of the yarn whereas calcium (D) is distributed throughout the entire fiber.

from a 15th-century Turkish velvet. Bar = 100 Fm.

metal wrapping is clearly identified by z-contrast.

wrapping of the Turkish yarn demonstrates an intense silver peak.

and mapping calls into question the exact source of this textile because the metal wrapping is gold-coated silver- copper alloy, a metal compositon typical of European metal threads of the period rather than those of Asia Minor (Reath, 1927). The energy spectra for both the wrapping and core yarn indicate the presence of silver, but there is no evidence in these particular x-ray maps (Fig. 10B) that silver has been transferred from the wrapping to the yarn. In other examples, however, the transference ofsilver into the core yarnsofsimilar metal- wrapped Turkish yarns was observed. Because of the

large spot sizes (50-100 nm) utilized during most anal- yses, it is possible that some signal may originate from beneath the area analyzed, especially in the case of fibers where the yarns are wrapped in metal foil. Smaller spot sizes and decreased accelerating voltage can provide an additional means of verification of ele- mental transference by reducing excitation volume but have the disadvantage of increased collection time and contamination. Calcium throughout the fibers (Fig. 10D) may be due to some textile treatment whether in fabric production or use.

PREPARING AND ANALYZING FIBERS 5

CONCLUSIONS Electron microscopy combined with x-ray analysis

and elemental mapping is a useful tool for the identi- ficatio n and characterization of archaeological and historic fibers and for the reconstruction of their his- tory. Fibers can be identified by their physical mor- phologies, and support for that identification is pro- vided by the elemental composition. Through the examination of the interior of fractured fibers and yarns clues to textile treatments such as dyes or mordants and the conditions of long-term storage can be attained. This technical application uses readily availa.de technology to provide additional information that i:: not addressed by analyzing the exterior of fibers. After examining fibers for which much informa- tion is known and accumulating a database of informa- tion concerning the factors of the environment which persist through alteration or degradation after burial, it will be possible in the future to examine fibers of unknown origin and infer some of the history of those materials.

ACKNOWLEDGMENTS The authors wish to thank Cathy Kelloes and Maya

Bouloss for technical assistance. The research was supported in part by state and federal funds appropri- ated to the Ohio Agricultural Research and Develop- ment Center, The Ohio State University.

REFERENCES Hardin, I.. and Duffield, F. (1986) Characterization of metallic yarns

in historic Persian textiles by microanalysis. In: Historic Textile and Paper Materials, Conservation and Characterization. Advances in Chemistry, Series No. 212. H. Needles and H. Zeronian, eds. Amer can Chemical Society, Washington, D.C., pp. 76-91.

Hoke, E . . and Petrascheck-Heim, I. (1977) Microprobe analysis of

gilded silver threads from medieval textiles. Stud. Conservation, 22:49-62.

Indictor, N., Koestler, R., and Sheryll, R. (19851 The detection of metallic mordants by energy dispersive x-ray spectrometry, part I. Dyed woolen textile fibers. J. Am. Inst. Conservation, 24:104-109.

Jakes, K.A., and Sibley, L.R. (in press) Evaluation of a partially mineralized fabric from Etowah. In: Proceedings of 25th Interna- tional Symp. Archaeometry, 24-29 May 1986, Athens, Greece.

Jakes, K.A., and Sibley, L.R. (1985) Chemical and physical analysis of textile fabric pseudomorphs. In: Proceedings of Application of Science in the Examination of Works of Art. P. England and L. VanZelst, eds. Boston Museum of Fine Arts, 7-9 September, 1983. pp. 221-224.

Jakes, K.A., and Sibley, L.R. (19841 An examination of the phenom- enon of textile fabric pseudomorphism. In: Archaeological Chemis- try 111. Advances in Chemistry, No. 205. J.B. Lambert, ed. Ameri- can Chemical Society, Washington, D.C., pp. 403-424.

Jakes, K., and Angel, A. (1989) The determination of elemental distribution in ancient fibers. In: Archaeological Chemistry IV. Advances in Chemistry. No. 220. R.O. Allen, ed. American Chem- ical Society, Washington, D.C., pp. 451-464.

Koestler, R., Sheryll, R., and Indictor, N. (1985a) Identification of dyeing mordants and related substances on textile fibers: A prelim- inary study using energy dispersive x-ray spectrometry. Stud. Conservation, 305-62 .

Koestler, R., Indictor, N., and Sheryll, R. (1985b) The detection of metallic mordants by energy dispersive x-ray spectrometry, part 11. Historical silk textiles. J. Am. Inst. Conservation, 24:llO-115.

Reath, N.A. (1927) Velvets of the Renaissance from Europe and Asia Minor. The Burlington, 50:258-304.

Saltzman, M. (1978) The identification of dyes in archaeological and ethnographic textiles. In: Archaeological Chemistry 11. Advances in Chemistry, No. 171. G.F. Carter, ed. American Chemical Society, Washington, D.C., pp. 172-185.

Schweppe, H. (1986) Identification of dyes in historic textile materi- als. In: Historic Textile and Paper Materials, Conservation and Characterization. Advances in Chemistry, Series No. 212. H. Nee- dles and H. Zeronian, eds. American Chemical Society, Washing- ton, D.C., pp. 153-174.

Walker, C., and Needles, H. (1986) Analysis of natural dyes on wool substrates using reverse-phase high-performance liquid chromatog- raphy. In: Historic Textile and Paper Materials, Conservation and Characterization. Advances in Chemistry, No. 212. H. Needles and H. Zeronian, eds. American Chemical Society, Washington, D.C., pp. 175-186.