THE ROLE OF SEPIOLITE-PALYGORSKITE IN THE DECAY OF A N C I E N T EGYPTIAN LIMESTONE SCULPTURES

Key Words--Clays, Egyptian Sculptures, ESEM, Limestone Decay, Polygorskite, Sepiolite, Swelling. I N T R O D U C T I O N D a m a g e p roduced by cycl ic swel l ing and shr ink ing of clays c o m m o n l y found as m i n o r c o m p o n e n t s in ornamen ta l and bu i ld ing s tone is a ma jo r p r o b l e m in the conse rva t ion of cul tural her i tage (Dunn and Hudec 1966; Cane r and Seeley 1978; M c G r e e v y and Smi th 1984; Wend le r et al. 1991; Mar t in -Pa t ino et al. 1993; Kuhne l et al. 1994; Ruiz et al. 1995; Brat t l i and Broch 1995). We studied one type of c lay-con ta in ing Egypt ian l imes tone s tela f rom Naga e l -Deir ( A b y d o s / T h e b e s reg ion) wh ich exhib i t s an ongo ing p r ob l em of deter iora t ion that is ev idenced by con t inuous de l amina t ion of the surface. The stela was acqui red f rom archeologica l inves t iga t ions carr ied out in a c eme te ry in N a g a elDe i r ear ly in the 20 th century. It has been no ted that this Egyp t i an sculpture, as wel l as m a n y others ca rved f rom the same l imes tone ( T h e b e s A b y d o s l imes tone) and s tored in d i f ferent m u s e u m s (Met ropol i t an Muse u m of Art, N e w York; Br i t i sh M u s e u m , London ; Phoebe Hears t M u s e u m of Anth ropo logy , Berkeley) , deve lop the same pa t te rn and degree of decay, even though they were not fully exposed to an ou tdoor env i r o n m e n t (Charo la et al. 1982; Brad ley and Middle ton 1988; Rodr iguez -Nava r ro et al. 1996). In fact, during the t ime these l imes tone sculptures were s tored in the m u s e u m env i ronm en t , they exper i enced a comple te loss of the ca rved surface relief, whi le be ing subj ec ted only to the re la t ive humid i ty and t empera tu re changes that occur red in the s torage areas. Accord ing to prev ious studies carr ied out on T h e b e s A b y d o s l imes tone sculptures (Oddy et al. 1976; H e l m s 1977; Charo la et al. 1982; H a n n a 1984; Brad ley and Midd le ton 1988; Midd le ton and Bradley 1989; Mi l le r 1992), it was though t that mos t of the obse rved damage was due to cyclic c rys ta l l iza t ion o f smal l quant i ties of salt ( sod ium chlor ide and sod ium nitrate). In fact, such sculptures had been t reated and desa l ina ted by aqueous poul t ic ing, bu t the decay process was not arrested, as ev idenced by ma jo r loss of surface s tone after s torage for a per iod of years. Brad ley and Midd le ton (1988), s tudying var ious types of Egyp t i an l imes tone sculptures in the col lec t ion of the Br i t i sh M u s e u m that exhib i t s imi lar deter iorat ion, po in ted out that the minera logy , and especia l ly the clay content , should play a role in the suscept ibi l i ty of this s tone to decay. In a p rev ious work (Rodr iguez -Navar ro et al. 1996) some p re l imina ry resul ts on the analys is of the decay m e c h a n i s m of this mater ia l were presented. In this paper, new analyses demons t r a t e that the or ig inal hypothesis , that is, the swel l ing of the clay minera l s in this l imestone, is the pr incipal m e c h a n i s m respons ib le for the obse rved damage . The a im of this work is to ident i fy and charac ter ize the clay wi th in this s tone and demons t ra t e its role as the ma in cont r ibu tor for its decay. In this paper we present and discuss data showing that no rmal ly nonexpandab le clays (sepiol i te and palygorski te) can promote s ignif icant d a m a g e due to swel l ing (osmot ic Copyright 9 1998, The Clay Minerals Society 414 Vol. 46, No. 4, 1998 Clays and the decay of Egyptian sculptures 415 swelling), a process that has been extensively reported in the case of expansive clays (smectites), but not in the case of sepiolite and palygorskite. MATERIALS AND METHODS Samples from an Egyptian limestone stela dated to the II to IV Dynasty (2720-2150 BC) from the Phoebe Hearst Museum (Berkeley, California) were studied by means of optical microscopy, XRD, ESEM and transmission electron microscopy (TEM) with energy dispersive spectrometry (EDS) microanalysis. The clay fraction was extracted using the following experimental protocol: 50 g of crushed limestone were treated with 1 M acetic acid and the acid-insoluble residue was washed repeatedly with distilled water. The silt (2-20 ixm particle size) and clay (<2 txm particle size) fractions were separated by centrifugation. Oriented clay aggregates (air-dried, EGand DMSOsolvated, and after heating to 550 ~ were prepared. XRD analysis of the oriented clay aggregates was performed with a Philips PW-1710 diffractometer, equipped with a graphite monochromator and automatic slit, using CuKo~ radiation. Silicon powder was used as an internal standard. The clay content was calculated from the weight of the clay fraction following drying at 90 ~ for 24 h. The clay fraction morphology and chemical composition were studied by means of TEM with coupled EDS (TEM-EDS, Philips model CM20). The high magnification study of the in situ dynamics of clay swelling was performed using an environmental scanning electron microscope (ESEM, ElectroScan, Model E-3) by treating limestone samples (2 x 4 x 4 mm) with distilled water (using a syringe) after they were placed in the ESEM chamber at low vacuum (2 torr). Additionally, repeated condensation-evaporation cycles were performed by increasing (up to 5.5 torr) and reducing (to 2.0 torr) the ESEM chamber pressure while maintaining the sample at constant temperature (1.5 ~ using a cooling stage (Messier and Vitale 1993). The expansion of the pure clay (extracted from the acid-insoluble residue and airdried for 48 h) when in contact with water (following condensation of water in the ESEM chamber) was also observed by means of the ESEM. The same experiment was performed after adding a 1% by weight NaC1 solution to the pure clay (1 cc of saline solution was added to 1 g of pure clay), following drying at room temperature and before introducing the sample into the ESEM chamber. Numerous still and video images were recorded before, during and after water treatment. To observe stone damage due to contact with water, wetting/drying cycles were performed by immersing 5 samples (2 x 3 x 4 cm) in distilled water for 4 h, then drying for 20 h in an oven at 90 ~ Most of the samples collapsed after the first cycle, but some withstood 3 cycles. In order to reproduce the damage mechanism of the limestone under conditions similar to those in the museum environment, 4 Egyptian limestone blocks (2 x 3 x 5 cm) were placed in a chamber where they were submitted to cyclic relative humidity changes (RH ranging from 40% up to 90%; each cycle lasting 8 h; total number of cycles 100). Damage was evaluated by weighing the material that fell off of the stone surfaces. Total weight loss was expressed as a percentage of the original weight of the samples. Total expansion due to water absorption of the stone was estimated using TMA. Four limestone samples (5 x 5 x 5 mm) were tested. Each sample was placed in the TMA sample holder with the bedding planes parallel to the bottom and was allowed to equilibrate in the laboratory environment (30 ~ and 55% RH) for 2 h. Then, distilled water was added and the isothermal linear expansion of the sample was recorded using a Mettler TA300 TMA instrument. When maximum expansion was reached, the sample was dried and contraction of the sample was recorded. Two wetting/drying cycles were performed for each sample and the results were averaged. The stone compressive strength (for both dry and wet samples) was evaluated using an Instron Model 4201. Ten samples were cut (20 X 20 X 40 mm) and tested after drying in an oven at 80 ~ overnight. Half of the samples were tested after immersion for 30 min in distilled water.