Khentiamentiu: Shining light on Egyptian mirrors: New scientific research into their metallurgy

https://www.sciencedirect.com/science/article/pii/S2352409X24003729?via%3Dihub

Received 17 June 2022, Revised 19 August 2024, Accepted 26 August 2024, Available online 4 September 2024, Version of Record 4 September 2024.

https://doi.org/10.1016/j.jasrep.2024.104744

Get rights and content

Under a Creative Commons license

open access

Highlights:

•
Minimally destructive sampling and analytical methodology provide both chemical and microstructural data for metal mirrors.
•
The manufacturing process of Egyptian mirrors was consistent over time.
•
Surface enrichment: preliminary evidence on how it was produced.

Abstract

Ancient Egyptian mirrors have received a great deal of attention for their cultural significance and connections to religious beliefs. However, in-depth investigations into their manufacture have been lacking with only a handful included in wider analyses. So how were mirrors manufactured? The preliminary metallurgical analysis presented here aims to shed light on this. With the use of SEM-EDX, analyses of Egyptian mirrors from a range of UK museum collections, spanning from the Old Kingdom through to the Late Period, have revealed the chemical composition and microstructure of the metal.

1. Introduction

Egyptian mirrors were manufactured to generate desirable reflections for their users; this inevitably leads to questions concerning how the mirrors functioned and their colour. Though they were everyday objects for the elite in Egyptian society, they were religiously linked to numerous gods and their associated attributes. Much of the literature discussing these objects treat them as art, drawing upon their visible physical attributes such as their handles or focusing on their religious connections (Lilyquist, 1979 or Bènèdite, 1907 for example) in order to draw conclusions about their use.

The metal disk of the mirror has received limited attention, and exploration of the function and colour of the disks themselves is lacking as comments are often limited to 'made of bronze' or 'polished to almost gold' (Szpakowska, 2008). However, such statements can be misleading; archaeometallurgical analyses have demonstrated, for example, that Egyptian metal-workers used arsenical coppers before the introduction of tin copper alloys during the Middle Kingdom (Ben-Yosef, 2018) and that both tin and arsenical copper alloys can have a silvery appearance (Meeks, 1993). Metallurgical analyses, therefore, provide the only means with which to determine the technical choices made during the manufacturing process. Although recent papers have begun building a compositional database of Egyptian metal artefacts (Odler and Kmošek, 2016, Rademakers et al., 2017, Rademakers et al., 2018, Odler and Kmošek, 2020 etc.), these have been mainly limited to strictly utilitarian objects (i.e. bowl, razor, chisel) without the symbolic loading of mirrors.

There have been some investigations into Egyptian mirrors, for example, 40 mirrors from the Louvre Museum were analysed using Atomic Emission Spectroscopy by Michel (1972) and 6 dated between the Old Kingdom and the Middle Kingdom were included in Eaton and McKerrell (1976), which provides an overview of the alloy types used for mirrors in the Pharaonic period. A small number of mirrors have been included in wider studies (Gilmore, 1986, Masson-Berghoff et al., 2018, Rademakers et al., 2018 etc.), however, these have been limited to compositional information only. The microstructure of Egyptian mirrors has been touched upon by Odler and Kmošek (2020) and Garenne-Marot (1984) but these have not included larger numbers of mirrors and leave questions unanswered regarding surface enrichment in particular.

This investigation focuses on the analyses of Egyptian mirrors originating in three UK collections, their dates spanning the Pharaonic period from various locations throughout Egypt. The aim of this paper is to present preliminary findings into the identification of the alloy types used to produce Ancient Egyptian mirrors and to observe the microstructure of the metal in order to identify manufacturing techniques used by ancient metalworkers.

1.1. Background

Metal mirrors in ancient Egyptian society were used to view their reflection as a part of a wider assemblage of cosmetic goods including wigs, make-up and jewellery (Watterson, 2011); as exemplified by two Middle Kingdom mirrors found within cosmetic boxes containing such items from Dashur and Lahun (Lilyquist, 1979:77). Mirrors have most frequently been excavated from tombs of the elite and in particular women, although they have been discovered in association with men and children (Lilyquist, 1979:83-86). They were often located close to the deceased when deposited in the tomb (Philip, 2006:107). It is thought that the disks were wrapped in textiles to protect their reflective surface (Riggs, 2014:139) as access to a reflection was considered important, to recognise oneself for the afterlife (Szpakowska, 2008:72). Access to a reflection was viewed as a necessity to the extent that model versions of mirrors, either small metal or wooden disks, were produced specifically to be deposited in tombs as replacements of the 'real' thing, which through magic would become fully functional in the afterlife (Tooley, 1995:8). Mirrors are depicted on several mediums including tomb scenes, stele and coffins. The disks are coloured exclusively as white, yellow, orange or red (Lilyquist, 1979:94) most likely due to their religious links with the sun god Ra (Castañeda Reyes, 2010:40) as the metal disk was viewed as a depiction of the sun disk (Derriks, 2001:59) or as representations of the different metal alloys used to produce them (Radivojević et al., 2018).

Copper alloys were commonly used in ancient Egypt to create everyday objects. The development of copper smelting around the fourth millennium BC (Ogden, 2000: 150-151) allowed greater quantities of metal to be obtained that enabled the production of larger items such as mirrors, with examples coming from Abusir dated to the Early Dynastic Period (Bonnet, 1928). The disks of mirrors would have been cast separately from the handle, which was later attached by riveting (Scheel, 1989:41; Ogden, 2000:158).

Up to and during the Middle Kingdom arsenical copper was the alloy of choice (Scheel, 1989:21) with archaeological specimens ranging from 0.5 wt% through to 8 wt% (Scott, 2011:87). There is a debate surrounding the topic of deliberate alloying with arsenical coppers as arsenic naturally occurs as an impurity within certain copper ores (Moorey, 1994:242), meaning there is the possibility of ancient artisans purposely selecting ores for their arsenic content rather than intentionally adding it as a separate ingredient. However, amounts as low as 1 wt% arsenic could be an indicator of deliberate alloying where the quantities of other impurities within the metal are much lower (Ogden, 2000:153). Another observation that supports alloying control is the presence of consistent levels of arsenic within a given assemblage; as is seen in Cowell's (1987) analysis of Egyptian axes.

The alloying of tin with copper to produce bronze was first introduced in the Middle Kingdom, a transitional period in Egyptian metallurgy with arsenical copper, copper-tin alloy and their ternary alloy, arsenical copper-tin alloy, all in use. The shift from the use of arsenical copper to copper-tin alloy was gradual as tin was not readily available in Egypt and would have been imported (Scheel, 1989:18). Copper-tin alloys then became the dominant metal during the New Kingdom, possibly because tin was less toxic than arsenic. Also, smaller quantities are lost during re-melting meaning it is easier to control the amount of tin and maintain certain desirable attributes of the alloy when recycling (Scott, 2011:86-87). Levels of tin between 2-16 wt% are seen in the archaeological record (Lucas and Harris, 1962:217). Although, the use of arsenical copper continued into the New Kingdom and later, possibly due to a desire for specific characteristics of the alloy (Craddock, 1995:289-291). Leaded copper-tin alloy was not used in Egypt until the Third Intermediate Period onwards (Ogden, 2000:155).

Arsenical coppers readily segregate with as little as 2 % arsenic within the copper (Budd and Ottaway, 1995), forming an alpha (α) and gamma (γ) eutectic. These phases can be easily identified due to the specific amount of arsenic present in them, for example, the alpha phase contains up to 8 wt% arsenic and the gamma (γ) phase contains between 29–31 wt% (Scott, 2011). The segregation can be observed in the dendritic microstructures of cast-only objects; however, excess arsenic can also naturally form an arsenic-rich gamma layer on the surface of the metal (Scott, 2011:39). The gamma phase can be squeezed through interdendritic feeders towards the surface of the object, this phenomenon is known as inverse segregation (Mödlinger and Sabatini, 2016:71). There is the possibility that the casting and subsequent working conditions of the metal may have been manipulated by ancient artisans to utilise this natural phenomenon to intentionally induce an enriched arsenic surface. This kind of surface would not have required a change in the overall percentage of arsenic in the metal but would have drastically altered the appearance of the metal, making it look more silver in colour compared to that of the bulk alloy substrate.

Some scholars have attempted to recreate this enriched surface layer in arsenical coppers with varying success. Mödlinger and Sabatini's (2016:65-66) experiments demonstrated that when annealed, arsenic within the gamma phases was re-absorbed into the alpha phase and disappeared, therefore not creating an enriched layer on the surface. This leads them to suggest an alternative surface treatment was used to create enriched arsenic layers in archaeological specimens, the process of cementation, though they had no archaeological data to prove the use of cementation in Egyptian mirrors specifically (Mödlinger and Sabatini, 2016: Tab.3, pp.72). Ryndina and Kon'kova's (1982) research, however, successfully produced an arsenic-enriched surface by inverse segregation and maintained the layer through annealing, thus providing weight to the argument that ancient metalworkers could have had the knowledge and skill to do this.

2. Materials & Methods

2.1. Materials

This investigation brought together three different museum collections; the Garstang Museum of Archaeology at the University of Liverpool, The Manchester Museum and the Liverpool World Museum. In total nineteen mirror disks were analysed. Provenance information was available for 10 of the mirrors; their dates spanned from the Old Kingdom through to the Late Period and covered a large geographical area within Egypt (details can be seen in Table 1). Although not all of the mirrors have contextual information, data gathered from them can still be useful for general observations into how Egyptian mirrors were manufactured. Five of the disks had evidence of textile remains and two of the mirrors are thought to be models based on their size (E993 and 44.19.353). The artefacts were in varying states of preservation, some more heavily corroded than others with a number having previous conservation work. Images of the mirrors can be seen in Fig. 1.

Table 1. Artefact details including provenance and dimensions.

Museum	Accession Number	Provenance		Dimensions (mm)	Thickness (mm)	Weight (g)
Museum	Accession Number	Period	Location	Dimensions (mm)	Thickness (mm)	Weight (g)
Garstang Museum of Archaeology	E936	−	−	159 x 143	1.8	183
	E938	−	−	120 x 127	1.7	106
	E945	−	−	142 x 134	2.4	332
	E946	−	−	133 x 130	2.4	188
	E993	Old Kingdom / Middle Kingdom	Beni Hassan Tomb 97 (?)	45 x 43	1	9.5
	E1452	−	−	120 x 120	0.9	76
	E1519	−	−	96 x 107	1.6	95
	E1541	−	−	96 x 88	1.7	61
	E1564	−	−	130 x 120	1.5	224

Manchester Museum	1647	7th Dynasty First Intermediate Period	Diospolis Parva, Hiw Tomb Y91	135 x 131	2	110
	1648	12th Dynasty Middle Kingdom	Dendereh Tomb 480	142 x 129	2	200
	1653	18th Dynasty New Kingdom	−	101 x 121	7	400
	2859	5th/6th Dynasty Old Kingdom	Dendereh	144 x 132	3	319
	3277	26th Dynasty Late Period	Tell el-Yahudiya Tomb 320 (?)	157 x 114	4.3	28
	8744	−	−	111 x 134	6.6	57

Liverpool World Museum	4.10.97.10	5th Dynasty Old Kingdom	Deshasheh Grave 36	141 x 151	2	150
	25.9.99.131	Middle Kingdom	Diospolis Parva, Hiw Tomb Y34	104 x 99	2	94
	44.19.353	−	−	77 x 86	2	92
	1964.163	New Kingdom?	−	133 x 133	3	328

2.2. Methods

A scanning electron microscope with an energy-dispersive X-ray spectrometer (SEM-EDX) was used as the analytical technique for this investigation. At the University of Liverpool, in the Elizabeth Slater Laboratories, the Department of Archaeology, Classics and Egyptology operates a JEOL JSM-IT300 SEM with a Bruker XFlash6-30 EDX system. This specific SEM has a large chamber (up to 200 W x 200 D mm), big enough to accommodate whole specimens without the need for sub-sampling. Furthermore, SEM-EDX is non-destructive while producing high-resolution high-quality greyscale images as well as compositional information. Precise qualitative and quantitative data for both major and minor elements in the bulk metal can be routinely gathered (Goldstein et al., 2017). As the measurement of trace elements was not a concern in this investigation, the sensitivity of this analytical technique is sufficient for the level of compositional information required. The measurements were obtained under 20 kV accelerating voltage from the working distance of 10 mm with spectra collected for 60 s live time. To ensure the quality and reliability of the results two multi-element standard reference materials were used in the form of a MAC standard (11915) and MBH-32X SN6 B (bronze alloy).

The mirrors were first photographed, weighed and measured before an area for analysis was prepared. Any textile remains were recorded using the Keyence VHX-7000 4 K digital microscope before sampling occurred.

The minimally destructive sampling technique used in this methodology is called deep-filed edge abrasion, which removes a small amount of material from the edge of the disk through grinding producing a tangential taper section. This sampling technique provides a cross-section of the mirror by removing the surface corrosion layers and exposing the original bulk metal underneath. Initially, to remove the corrosion products, the edge is ground using 1200 grit corundum abrasive paper on the slowest rotational speed setting. Next, the sample is polished using 6, and then 3-µm diamond paste. It is not always possible to remove the scratches from the sample due to the extent and condition of the corrosion present. This is because the corrosion can become loose and drag across the sample. To remove the scratches, the sample would have to become deeper and larger which would become more visually apparent, the opposite of what museums had agreed upon.

After the compositional data is collected via SEM-EDX, the polished edge is etched to reveal the microstructure. For the mirrors in this investigation, the etchant Ethanolic Ferric Chloride was used as it had been shown to work well on copper alloys (Scott, 1991:72). The etchant is applied to the polished area using the 'wipe' method where a cotton bud soaked in the etchant is rolled over the surface of the metal. The etchant is then rinsed off using ultra-pure water and neutralised using a 10 % Sodium Hydrogen Carbonate (NaHCO3) solution to prohibit any further reaction with the material. The neutraliser is washed off with ultra-pure water and ethanol, and the mirror is dried. Once etched, the mirror is placed back in the SEM where images of the microstructure were taken. The microstructure is only observed via SEM, which is sufficient to observe key microstructural characteristics, as it is not possible to fit a mirror placed upright so the sample is in view under an optical microscope.

This sampling and analytical methodology was chosen as the researcher acquires information concerning composition and microstructure together, allowing research questions regarding their manufacture to be better addressed, without compromising the integrity of the whole artefact. Both the researcher and the museum gain valuable data on the artefacts without impacting the overall 'completeness' of the mirror or collection. As such, a major advantage of this technique is that museums are more likely to loan specimens out for research, compared to requesting to take a more traditional 'V' section sample.

However, there are limitations linked to the use of SEM-EDX. Due to the size of the chamber, specimens larger than 200 x 200 mm cannot be analysed in this way. If other researchers were to use this method, it would require an SEM with a large chamber that can accommodate a whole specimen and a stage that can be tilted vertically for the sample to be in view. Not all researchers would have access to such SEMs as many have small chambers that can only hold mounted cross-sections. Additionally, trace elements nor lead isotope data can be obtained this way therefore metal provenance cannot be investigated.

For a more detailed description of the methodology and a discussion of the advantages and limitations see Thomas and Gethin (2024).

3. Results

3.1. Composition

Three alloy types were identified; arsenical copper, copper-tin alloy and a ternary alloy of arsenical copper-tin. These have also been identified as alloys used for mirror production in other studies, such as by Rademakers et al. (2021). The average composition for each mirror can be seen in Table 2, these were calculated from a minimum of nine analyses collected over three different areas of the sample. The average alloyed percentage for the entire assemblage was 5 wt%, whether it be with arsenic or tin. Fifteen out of nineteen specimens were produced using arsenical copper. The lowest arsenic content observed was in 4.10.97.10 at 2.6 wt% with the highest being E1564 at 8 wt%. The average quantity of arsenic in the disks is 5 wt%. The majority of the mirrors comprise a single alpha phase whereas two arsenic phases, alpha and gamma were identified in several specimens (see Surface Treatments for further discussion). Mirrors 3277 and 8744 were the only two disks to be produced using copper-tin alloy exclusively, their tin content was 5.1 and 5.2 wt% respectively. An additional two specimens comprise a ternary alloy arsenical copper-tin alloy (1653 and 44.19.353) at 4.8 and 4.1 wt% of tin, with low percentages of arsenic at 0.3 and 1.1 wt% respectively. The average tin content for mirrors that are solely made of bronze is 5.2 wt%, and this is reduced to 4.8 wt% if the tertiary alloys are included.

Table 2. Average composition for each mirror.

Mirror	Elements (wt%)
Mirror	Cu	As	Sn
E936	94.9	5.1	∼
E938	95.4	4.6	∼
E945	95.8	4.2	∼
E946	94.3	5.7	∼
E993	97.2	2.8	∼
E1452	94.3	5.7	∼
E1519	94.5	5.5	∼
E1541	95.4	4.6	∼
E1564	92.0	8.0	∼
1647	94.8	5.2	∼
1648	95.0	5.0	∼
1653	94.9	0.3	4.8
2859	95.4	4.6	∼
3277	94.9	∼	5.1
8744	94.8	∼	5.2
4.10.97.10	97.4	2.6	∼
25.9.99.131	93.5	6.5	∼
44.19.353	94.8	1.1	4.1
1964.163	95.0	5.0	∼

3.2. Microstructure

The mirrors displayed a recrystallized granular microstructure when etched or highlighted in the corrosion. This indicates the mirrors have been through cycles of working once cast, almost certainly hammering, most likely to increase the hardness of the metal. This type of microstructure has also been identified in Egyptian mirrors by Odler and Kmošek (2020) and Garenne-Marot (1984). As extensive working can make the metal brittle and crack, the disks were annealed relieving the stress in the metal allowing for further working (Northover, 1989: 112). The process of annealing can be identified through the presence of twinning which is observable in the mirrors (Fig. 2 red arrows). Slip lines form from the stress of hammering and are removed during the annealing process (Scott, 1991: Fig. 12, 8). This characteristic was observed in the metal (Fig. 2- blue arrows) meaning extensive cold working was the final stage of manufacture. This is probably due to a mirror needing to be hard to take polish for a good reflection.

3.3. Surface Treatments

Arsenic-rich gamma metallic phases, approximately 30 wt% As were observed in five of the specimens; they were evident within E946, E1452, E1541, 1647 and 25.9.99.131 (Fig. 3 and Table 3).

Table 3. Table showing a comparison in chemical data between surface enrichment and bulk material.

Mirror	Enrichment − 1 (As wt%)	Bulk Metal − 2 (As wt%)
E946	30.0	5.4
E1452	29.6	5.9
1647	31.1	5.9
E1541	30	5.1

In mirror E946, the enriched metal appeared as a distinct band around the edge of the prepared area, which corresponds to the surface of the disk, as highlighted in the elemental maps by a purple line (Fig. 4 Left). The gamma phase metal can also be identified on the greyscale image as it appears lighter than the bulk metal (Fig. 4 Right). This layer, with such a high arsenic content, would have formed a silver-coloured metal and so the mirror surface would have produced a white clear reflection once polished (Meeks, 1993:268). From Fig. 4 Right it can also be observed that corrosion had penetrated underneath the enriched layer deeper into the bulk metal below, this is due to the gamma phase metal being more resistant to corrosion than metal with lower levels of arsenic (Northover, 1989). Unfortunately, other than the band of the enriched metal itself, no clear evidence survived to indicate how the arsenic-rich surface was formed.

The process of inverse segregation creates interdendritic feeders; these are channels of arsenic-rich metal that are forced to the surface during casting producing an enriched surface layer. The segregation occurs as the arsenic-rich metal has a reduced melting point than the surrounding lower arsenic metal, as the metal freezes there is nowhere else for the remaining liquid metal to go other than the surface, where it subsequently cools and solidifies (Meeks, 1993). Garenne-Marot (1984) states the ancient metalworkers were able to identify the natural segregating properties of arsenical coppers and manipulated the metal to produce silver surfaces on the mirrors. However, no micrographs were presented in support of this statement. The presence of interdendritic feeders in four of the mirrors, E1452, E1541, 1647 and 25.9.99.131 (Fig. 5), now provides such evidence. The fact that these mirrors also went through extensive cycles of hammering and annealing demonstrates that they were also aware of how to maintain the enrichment throughout the forging process.

It should be noted here that some mirrors had been chemically stripped since excavation or were heavily corroded meaning the original surface of the disk was most likely destroyed and no longer identifiable. The implications of this mean that surface enrichment may have been more common than has been observed in this assemblage.

4. Discussion

A standardised manufacturing technique for mirror disks in Ancient Egypt is evident from the consistency in the microstructural data in this assemblage. The procedure does not differ between shape, size or type, such as model mirrors, and was established as early as the Old Kingdom with mirror 4.10.97.10 in the 5th Dynasty through to mirror 3277 in the Late Period. The molten metal would have been poured into an open mould carved from stone or made of ceramic close to the preferred final shape (Ogden, 2000:157), once cooled the disk was forged by cold working and annealing and repeated until the desired shape, thickness and hardness was reached during the final forging. The disk would have then been polished using an abrasive material such as stone or possibly sand (Gouda et al., 2007:20), and to achieve a shiny reflection the metal may have been burnished using textiles (Scheel, 1989:40); with a harder metal the disk would have taken a better polish which in turn created a clearer reflection.

The alloy types identified in specimens with provenance correlate to the dominant alloy types used during the specific periods within Egyptian history (Ben-Yosef, 2018: 209); for example, 2859 dated to the Old Kingdom or 1648 to the Middle Kingdom were both arsenical coppers which was the main metal alloy used during those periods.

There has been an ongoing debate within the literature regarding the deliberate alloying of arsenical coppers (Moorey, 1994, Ogden, 2000 for example). As previously mentioned, an argument for the intentional addition and control of the arsenic content is the presence of a consistent level of arsenic within one artefact type (Cowell, 1987) as well as the quantities of other impurities within the metal being lower than that of the arsenic (Ogden, 2000:153). Within this assemblage, other impurities were not identified within the bulk composition and there was a consistent level of arsenic and tin observed at around 5 wt%. There will always be anomalies to a trend but fourteen out of nineteen specimens fell between 4-6 wt% arsenic or tin. With the mirrors' provenance spanning multiple periods and locations throughout Egypt, the consistency in the quantity observed suggests they were manipulating and controlling the levels of arsenic and later tin for this particular artefact type.

This leads to the question of why were they aiming for an arsenic or tin content of 4-6 wt%. These percentages may have produced certain desirable attributes in the metal which were ideal for mirrors; the physical properties may have been one of the reasons. Typically, the higher the alloy percentage the harder a metal becomes (Northover, 1989: Fig. 13.3). However, from cast there is an insignificant increase in the hardness of the metal between 3-7 wt% arsenic (Scott, 1991: App.C, pp.82), so there would have been little advantage in having 5 wt% arsenic over 3 wt% for example. Furthermore, Odler and Kmošek (2016: 248) discovered that the hardness of the Egyptian metal artefacts in their investigation were more reliant on the mechanical hardening through working than on the composition of the alloyed metal. Other studies such as Pereira et al., 2013, Lechtman, 1996 and Junk (2003) support this finding, as does the microstructural evidence of the mirrors in this investigation as they were all extensively worked. This means the ancient craftsmen were not aiming for 5 wt% for the hardness of the metal exclusively because this property of the metal would be increased through working anyway. Experiments by Budd and Ottaway (1995), however, determined that arsenical coppers containing between 2-6 wt% arsenic were capable of 60–80 % reduction without cracking by cold working, above 6 wt% the metal becomes brittle when worked and can crack, as observed with mirror E1564. This quality would be beneficial for the production of mirrors as to ensure the disk remained undamaged through the forging process. Aiming for 5 wt% would mean that the metal could be easily worked to a certain shape and thickness without breaking to achieve a hardness that would produce the desired reflection.

Another possible attribute is the colour of the metal, which is an important visual aspect of mirrors; the amount of arsenic or tin may have been manipulated to achieve a particular shade of copper, gold or silver (Scott, 1991: App.C, pp.82). Since ancient artisans could not measure the amount of arsenic or tin within their alloys as we do today, they would have relied on other characteristics such as colour. It is possible that once an arsenical copper or copper-tin alloy reaches between 4-6 wt% the colour of the metal changes, achieving a shade the ancient artisans knew would have the physical attributes discussed above and therefore suitable for a mirror. The combination of these factors, work hardening, reduction and colour provide one possible explanation as to why most of the mirrors were made of such a specific alloy.

Understanding the difference in colour is important as this would have directly impacted the reflection produced and in turn how an individual would have viewed themselves. Based on the average arsenic content of the assemblage, the metal would have been golden in colour and the mirrors with surface enrichment would have appeared silvery (Radivojević et al., 2018: Fig. 9, 116). Why produce a certain colour reflection? Firstly, in practical terms, a golden or silvery colour is much lighter than pure copper, a dark red, meaning that the reflection would be clearer. Silver in Ancient Egypt was more valuable than gold (Schorsch, 2000) so silver-appearing mirrors, produced through surface enrichment, may have been fashioned for the owner to look as if they were of high status. Actual silver mirrors would have been extremely valuable as only a small percentage of the elite, such as a member of the royal family, would have had access to the resources necessary to obtain one. Two silver Egyptian mirrors have been discovered to date (The Metropolitan Museum of Art: 26.8.97 and 26.8.98). It is thought they were owned by one of three foreign wives of Thutmose III, dated to the 18th Dynasty New Kingdom (Patch, 2005: 222).

The model mirrors E993 and 44.19.353 were manufactured using the same sequence as the full-sized mirrors, this approach of creating model versions in the same way as full-sized functional objects was also observed in axes by Cowell (1987) and general tools by Odler and Kmošek (2016). As they were produced not to be functional but only to be placed within a funerary context, the quality of the reflective surface may not have been a priority as magic in the afterlife would have ensured the mirror served its purpose. This mentality may be reflected in the properties of the metal as the model mirror E993 has lower levels of arsenic. Overall this would have produced a much softer and darker coloured metal meaning the reflection would not have been as clear as other mirrors when polished. The disk was also much thinner compared to its full-sized counterparts; this may have been intentional for the mirror to fit within a small model.

One archaeological specimen confirmed to have been produced using cementation was the Horoz Tepe bull, however, the interior bulk metal is a copper-tin alloy with no arsenic present, whereas the surface layer was arsenical copper (Smith, 1973). The difference in composition between the bulk and enrichment enabled the identification of the production method, however, the mirrors in this assemblage comprise arsenical copper with an arsenic-rich layer, meaning that the use of cementation cannot be fully supported.

The other possible method for the production of silver-coloured mirrors is inverse segregation. The lack of feeders present in E946 does not rule out the possibility that it was formed by inverse segregation. Budd (1990) demonstrated that through annealing it is possible to maintain the enriched layer on the surface while simultaneously removing the feeders. Through solid-state diffusion facilitated by heat during annealing, the high amounts of arsenic in the interdendritic feeders can be re-absorbed into the bulk metal meaning the feeders are no longer present. This is heavily dependent on the temperature and the length of time the metal is annealed. The phenomena of creating a homogenous bulk metal while maintaining surface enrichment were observed through experimentation by Ryndina and Kon'kova (1982) under reducing conditions between 400–450 °C (Ryndina and Kon'kova, 1982: Fig. 4,5–6). Furthermore, this type of surface enrichment has been observed in other artefact types in other locations around the world. For example, analyses of an early Bronze Age blade from Los Millares, Spain (Hook et al., 1991) revealed that a 5 wt% arsenical copper had been extensively cold worked and annealed in addition to having an arsenic-rich layer on the surface achieved through inverse segregation (Meeks, 1993: 269-271). The micrograph has similarities to the mirror E946, where the bulk metal has been homogenized while maintaining the enriched surface. Therefore, mirrors such as E946 with a band of surface enrichment may have originally had feeders but these were removed during the heavy working and annealing of the metal (Meeks, 1993:271). Although this example is a different type of artefact from another country it can be used as an archaeological comparison to strengthen the argument for the use of inverse segregation to produce silvery surfaces in Egyptian mirrors. Ancient craftsmen, whether from Spain or Egypt were able to manipulate the metal in such a way as to create and maintain an enriched layer from casting through hammering and annealing to the final product.

Mirrors E1452, E1541, 25.9.99.131 and 1647 provide direct evidence for the manipulation of inverse segregation to create a silvery surface with the presence of an arsenic-rich gamma surface alongside interdendritic feeders. Subsequent working and annealing were not sufficient to homogenise the material and remove the segregation. This may explain why there is no distinct band on the surface as is found on mirror E946. While surface enrichment was most likely induced through the manipulation of natural segregation of the metal, further data from additional mirrors alongside experimental work are necessary to confirm the production method.

5. Conclusion

Overall, three different alloy types were identified for the production of Ancient Egyptian mirrors; arsenical copper, copper-tin alloy and arsenical copper-tin alloy. All the mirrors demonstrate a consistent manufacturing process based on an average composition of 5 wt%, whether arsenic or tin, which would have produced a golden-coloured metal. The ancient artisans were most likely aiming for this composition for both physical attributes and the colour of the metal. Examination of the microstructure of the specimens has enabled a manufacturing operational sequence to be established; the disks were worked, with cycles of hammering and annealing that increased the hardness of the metal to produce a better polish for a clearer reflection. The identification of surface enrichment in several specimens shows they were also producing mirrors of a silver colour. Inverse segregation was the most probable method to induce the enrichment, but further analyses and experimental work will be required for confirmation.

This preliminary investigation hints at potential trends and a direction for further analyses and experimental work as additional data from a larger specimen set is required to make any comprehensive conclusions. This work formed part of the author's MSc dissertation, who is now undertaking their PhD (Thomas, in preparation) to delve deeper into the manufacture of Egyptian mirrors by expanding the size of the artefact assemblage. There, a more in-depth discussion concerning topics such as understanding alloying practices over time or how deliberate alloy selection for mirrors compares to that in other artefacts will be carried out.

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Acknowledgements

Firstly, I would like to thank the following museums that allowed the analysis to take place; Manchester Museum at the University of Manchester, the Liverpool World Museum and the Garstang Museum of Archaeology at the University of Liverpool. Secondly, I would like to thank my supervisor Dr. Matthew Ponting alongside Dr. Peter Gethin for their assistance.

Data availability

No data was used for the research described in the article.

References

Bènèdite, 1907

G. Bènèdite

Les Miroirs

Catalogue Generale des antiąuites egyptiennes du Musee du Caire, Cairo (1907)

Google Scholar

Ben-Yosef, 2018

E. Ben-Yosef

Provenancing Egyptian metals: A methodological comment

J. Archaeol. Sci., 96 (2018), pp. 208-215, 10.1016/j.jas.2018.06.001

View PDF View article View in ScopusGoogle Scholar

Bonnet, 1928

H. Bonnet

Ein frühgeschichtliches Gräberfeld bei Abusir

Hinrichs'sche Buchhandlung, Leipzig (1928)

Search This Blog

Thursday, September 12, 2024

Shining light on Egyptian mirrors: New scientific research into their metallurgy - ScienceDirect

Shining light on Egyptian mirrors: New scientific research into their metallurgy