Inscriptions on A Roman Gilded Silver Helmet and A Twentieth Century Tinplate Can: Authenticity Examination and Surface Characterization

(1)

Conservation and Restoration Program, Master's, Metals Specialization

Inscriptions on A Roman Gilded Silver Helmet and A Twentieth Century

Tinplate Can: Authenticity Examination and Surface Characterization.

Student:

Ke-Shiuan Han

Student number:

11353864

Supervisor:

Tonny Beentjes, UvA

External advisor:

Ineke Joosten, RCE

Arie Pappot, RMA

Second reader:

Herman den Otter, UvA

(2)

Acknowledgments

I would like to thank my supervisor Tonny Beentjes (UvA) for the unlimited support to my research, guiding me to explore different possibilities and being encouraging all the time. I am grateful to Ineke Joosten (RCE), for introducing me different analysis techniques and kind supports. I thank Arie Pappot (RMA) for his help and sharing his insights on both study cases. I would like to thank Caecilia Thoen (Anne Frank Museum) for letting me study the tin can. I thank Ruurd Halbertsma (RMO) for sharing his knowledge of the Peelhelm, and helping me getting access to the museum archives.

I thank Maartje Stols-Witlox (UvA) for coordinating this module and providing aids to my questions. Rene Peschar (UvA), I thank him for sharing his knowledge and advising on the artificial aging test in the research. I am grateful to Maarten van Bommel (UvA) for his critical reflection and advice, and Ella Hendriks (UvA) for her consultations in the process.

I am grateful to Ellen van Bork (UvA/RMA) for her advices and instructions of Hirox and other equipment in the metal studio, and Tamar Davidowitz (UvA/RMA) for her help. I would like to thank Martin Jürgens (RMA) for sharing his knowledge about the RTI technique and the instruction of it. I thank Lambert van Eijck (TUD) for giving me the amazing opportunity to conduct the neutron tomography for the research. Zhou Zhou (TUD), I am grateful for his suggestion and explanation for the NT analysis, and processing the images. I thank Hans van-der-Weijde (TATA Steel Europe) for sharing his knowledge about steel and tinplate, and letting me visit the tinning production line. I would like to thank Ronny Meijers (The Valkhof Museum) for not only sharing the knowledge of archaeological metal objects, but also kindly providing his personal gilding collection for my research.

I am grateful to Magdalena Pilko (UvA) for sharing her study with me, and the encouragement she gives. I would like to thank Roy van der Wielen (UvA) for helping me solve the software problem of RTI. Furthermore, I would like to express my gratitude to the metals PI and MA students, they are the strongest backup for my study here.

Finally, I would like to thank my family in Taiwan for letting me to fulfill my dreams. I thank Fei for her great support professionally and personally. Ling-Ling Kuo (CM Museum) I thank her for encouraging me to advancing myself. Special thanks to Nikki den Boon for the support on my study and life here in Amsterdam.

(3)

Abstract (English)

Inscriptions on A Roman Gilded Silver Helmet and A Twentieth Century Tinplate Can: Authenticity Examination and Surface Characterization.

This thesis for the Master of the Arts in Conservation and Restoration of Cultural Heritage at the University of Amsterdam in 2018 investigates the inscription authenticity on two study cases. First is the Roman gilded silver helmet from the collection of Rijksmuseum van Oudheden in Leiden; second one is a twentieth century tinplate can from a private collector.

The “Peelhelm” was excavated in 1910 in the Netherlands, the inscription “STABLESIA VI ” on the cap indicates the helmet belongs to the sixth equites stablesiani in the late Roman period. However, the inscription was not recorded in its earliest documentation; therefore, there is some suspicion that it might have been added later to increase the historic value. The “Geconfyte gember” tinplate can is an ordinary industrial product; however, the inscription “ANNE MARGOT 1944 FRANK” found on the lid could greatly enhance the value if it is authentic. Anne Frank is the well-known Jewish author of “The Diary of a Young Girl”.

The goal of this study was to determine the authenticity of inscriptions on these two historic metal objects. The inscription characterization by optical examination indicates the order of sequence for the inscription and other incidents that took place on the metal surface. In the case of the Peelhelm, the order was established based on the association between crinkles/cracks and inscribed strokes. In the case of the “Geconfyte gember” can, the order was based on the relation between the corrosion layer and punch marks. The results suggest that the inscription “STABLESIA VI ” was added before the damage formed on the gilded silver helmet, and it was unlikely to be a later application after excavation; while “ANNE MARGOT 1944 FRANK” was applied after the tinplate can corroded. Since the can is an unprovenanced object, the inscription examination itself could not draw a conclusion. Reconstruction of the punched mark and historical background investigation on the can have been carried out, and the result suggests the inscription is possibly a recent application. Comparable surface features produced in the reconstruction may be considered as reference material for similar studies in the future.

Key words: inscription, authenticity, metal object, surface characterization, gilded silver, tinplate.

(4)

Abstract (Dutch)

Inscripties op een Romeinse vergulde zilveren helm en een twintigste-eeuws blik: authenticiteitsonderzoek en oppervlakte karakterisering.

Deze scriptie voor de Master of Arts in conservatie en restauratie van cultureel erfgoed aan de Universiteit van Amsterdam in 2018 onderzoekt de authenticiteit van inscripties op twee studieobjecten. De eerste is een Romeinse, vergulde zilveren helm uit de collectie van het Rijksmuseum van Oudheden in Leiden; de tweede is een twintigste eeuws tinnen blik van een particuliere verzamelaar.

De “Peelhelm” werd in 1910 in Nederland opgegraven. De inscriptie “STABLESIA VI ” op de kap geeft aan dat de helm behoorde tot de zesde cavalerie-eenheid stablesiani in de late Romeinse tijd. De inscriptie werd echter niet opgenomen in de vroegste documentatie van de helm; daarom bestaat er een vermoeden dat dit later is toegevoegd om de historische waarde te vergroten. Het “Geconfyte gember” tinnen blik is een industrieel product, maar het opschrift “ANNE MARGOT 1944 FRANK” op de deksel kan de waarde enorm vergroten als het authentiek is. Anne Frank is de bekende Joodse auteur van dagboek ‘Het Achterhuis’.

Het doel van deze studie was om de authenticiteit van de inscripties op deze twee historische metalen voorwerpen vast te stellen. Inscriptiekarakterisering door middel van optisch onderzoek, geeft de volgorde van applicatie aan van de inscriptie en ander incidenten die hebben plaatsgevonden op de metalen oppervlakte. In het geval van de Peelhelm werd de volgorde vastgesteld op basis van het verband tussen kreukels / scheuren en de inscriptiemarkeringen. In het geval van de "Geconfyte gember" kon de volgorde gebaseerd worden op de relatie tussen de corrosie laag en de inscriptiemarkeringen. De resultaten suggereren dat de inscriptie “STABLESIA VI

” was toegevoegd voordat de schade werd gevormd op de vergulde zilveren helm, en dat het onwaarschijnlijk is dat de inscriptie is toegevoegd na de opgraving, omdat “ANNE MARGOT 1944 FRANK” werd aangebracht nadat het blik al was gecorrodeerd. Omdat de herkomst van het blik onbekend is, kon het inscriptieonderzoek zelf geen conclusie geven. Reconstructie van de afdrukken en een historisch achtergrondonderzoek is daarom uitgevoerd, en het resultaat suggereert dat de inscriptie mogelijk recent is toegevoegd. Vergelijkbare oppervlaktekenmerken die tijdens de reconstructie zijn geproduceerd, kunnen in de toekomst als referentiemateriaal voor soortgelijke studies worden beschouwd.

Steekwoorden: inscriptie, authenticiteit, metalen voorwerp, oppervlaktekarakterisering, verguld zilver, blik.

(5)

Table of figures

All figures and drawings have been made by Ke-Shiuan Han, unless stated otherwise below.

FIG 1.1 THE PEELHELM ... 2

FIG 1.2THE GECONFYTE GEMBER CAN ... 2

FIG 1.3THE FLOW CHART OF THE RESEARCH ... 5

FIG 2.1AN OVERVIEW OF THE PEELHELM AND THE INSCRIPTIONS ... 6

FIG 2.2THE INSCRIBED SURFACE “STABLESIA VI ” ON THE RIGHT SIDE OF THE CAP. ... 7

FIG 2.3THE EARLIEST SKETCH OF THE HELMET.IMAGE IS TAKEN FROM THE RMO WEBSITE. ... 7

FIG 2.4THE HELMET AFTER RESTORATION IN 1970S.IMAGE PROVIDED BY RMO. ... 8

FIG 2.5THE DAMAGE AREA WITHOUT FILLING,1970S.IMAGES PROVIDED BY RMO. ... 8

FIG 2.6WHEN THE HELMET AFTER THE LATEST CONSERVATION.IMAGE PROVIDED BY RMO. ... 8

FIG 2.7XRF ANALYSIS ON THE PEELHELM ... 11

FIG 2.8THE PEELHELM WAS EXAMINED UNDER THE MICROSCOPE ... 11

FIG 2.9THE LOST AREA GLUED TO FILLING MATERIAL ... 12

FIG 2.10SUSPICIOUS SCRATCHES ON THE LEFT FLAP ... 12

FIG 2.11THE TOOL USED ON INSCRIPTION 1 HAD RECTANGULAR SHAPE ... 12

FIG 2.12THE TOOL USED ON INSCRIPTION 2 HAD RECTANGULAR SHAPE ... 12

FIG 2.13THE WIDTH OF STROKES IN INSCRIPTION 1 ... 13

FIG 2.14THE WIDTH OF STROKES IN INSCRIPTION 2 ... 13

FIG 2.15FLOWING LINE ON THE INSCRIPTION 1 ... 13

FIG 2.16HEAVY PAUSED LINE ON INSCRIPTION 2 ... 13

FIG 2.17THE STROKES ENDS BECAME WIDER ON INSCRIPTION 2 ... 13

FIG 2.18THE NECK PROTECTOR.IMAGE PROVIDED BY RMO. ... 14

FIG 2.19THE EDGE ON THE CAP.IMAGE PROVIDED BY RMO. ... 14

FIG 2.20 THE LEFT END OF THE DECORATION PATTERN ... 14

FIG 2.21 THE RIGHT END OF THE DECORATION PATTERN ... 14

FIG 2.22THE CRACK BREAKING THROUGH “S” ... 15

FIG 2.23THE STROKE FOLLOWING THE CURVE ON “ ” ... 15

FIG 2.24“S” IN 1911 PHOTOGRAPH ... 17

FIG 2.25THE CURRENT STATE OF “S” ... 17

FIG 2.26THE SECOND “ ” IN 1911 PHOTOGRAPH ... 17

FIG 2.27THE CURRENT STATE, CRINKLE OVER “ ” ... 17

FIG 2.28THE CRINKLES FOUND IN THE 1911 PHOTOGRAPH WENT OVER “L” AND “I” ... 17

FIG 3.1AN OVERVIEW OF THE “GECONFYTE GEMBER” TINPLATE CAN ... 19

FIG 3.2THE INSCRIBED SURFACE “ANNEMARGOT1944FRANK” ON THE LID ... 20

FIG 3.3THE BUMPY SURFACE OF THE BOTTOM END CAPTRUED BY HIROX ... 21

(8)

FIG 3.5STITCHED RTI IMAGES OF THE PUNCH MARKS ON THE BOTTOM END ... 21

FIG 3.6THE DAMASCENE PATTERN OBSERVED ON THE TIN CAN ... 22

FIG 3.7THE “MOTTLE” PATTERN RECORDED IN THE LITERATURE ... 22

FIG 3.11THE CAN WAS EXAMINED WITH A HIROX KH770 ... 28

FIG 3.12THE CAN WAS EXAMINED WITH A LEICA DM2500M ... 28

FIG 3.13 LEFT:UNCORRODED TINPLATE; MIDDLE: PARTIALLY OXIDIZED/CORRODED; RIGHT: COMPLETELY CORRODED ... 29

FIG.3.14A CLEAR TIN LAYER ON THE SURFACE WITH AN AVERAGE THICKNESS OF 3.29ΜM ... 31

FIG.3.15A DISCONTINUOUS LAYER OF TIN DID NOT GIVE A PROMISING THICKNESS MEASUREMENT ... 31

FIG 3.16THE PUNCHING DEVICE ... 32

FIG 3.17PUNCHING WITH THE HAMMER ... 32

FIG 3.18SPECIMENS IMMERSED IN ACID SOLUTION ... 35

FIG 3.19ACCELERATED AGING IN PROGRESS ... 35

FIG 3.20THE EDGE AROUND THE HOLE ON THE TOP LID ... 36

FIG 3.21THE SURFACE OF THE TOP LID ... 36

FIG 3.23TYPE II SURFACE FEATURE ... 36

FIG 3.24TYPE III SURFACE FEATURE ... 36

FIG.3.25TYPE III SURFACE FEATURE ... 36

FIG 3.26INSCRIBED SURFACE FEATURE MAPPING ON THE LID ... 37

FIG 3.27THE SURFACE OF THE BOTTOME END ... 37

FIG 3.28TYPE IV SURFACE FEATURE ... 37

FIG 3.29TYPE V SURFACE FEATURE ... 38

FIG 3.30TYPE VI SURFACE FEATURE ... 38

FIG 3.31SURFACE PUNCHED BY LIGHTER FORCE ... 38

FIG 3.32SURFACE PUNCHED BY HEAVIER FORCE ... 38

FIG.3.33SURFACE PUNCHED BY BAMBOO STICK ON NATURALLY CORRDED TINPLATE SURFACE ... 39

FIG 3.34SURFACE PUNCHED BY IRON TOOL ON ARTIFICIAL CORRODED TINPLATE SURFACE ... 39

FIG 3.35PUNCHED SURAFCE BEFORE BRUSH CLEANING ... 39

FIG 3.36PUNCHED SURFACE AFTER BRUSH CLEANING ... 39

FIG 3.37THE MARK HAD AN IMPRINT FROM THE TOOL SURFACE ON NON-CORRODED SURFACE,S6-1 ... 40

FIG 3.38THE SURFACE OF THE TOOL USED IN RECONSTRUCTION ... 40

FIG 3.39THE CORROSION WAS PUNCHED INTO THE CAVITY ... 40

FIG 3.40MARK ON OXIDIZED BUT STILL METALLIC SURFACE,S2-1(POL0。 ) ... 40

FIG 3.41THE RIM OF THE PUNCHED AREA TURNED INTO LIGHTER COLOR, ON CORRODED SURFACE S3-2 ... 40

FIG 3.42THE PATTERN CAUSED BY THE IMPACT ON CORRODED SURFACE,S4-2 ... 40

FIG 3.43THE NET PATTERN AROUND THE IMPACTED AREA, ON CORRODED SURFACE S1-1 ... 41

FIG 3.44THE RADIAL PATTERN AROUND THE IMPACT AREA, ON CORRODED SURFACE S1-2 ... 41

FIG 3.45S1-1 AFTER 144 HOURS ... 41

(9)

FIG 3.47S6-1 AFTER 288 HOURS. ... 41

FIG 3.48BLISTERING AND FLACKING OFF ON THE SURFACE ... 41

FIG 3.49 ... 42

PUNCH MARK ON NATURALLY CORRODED TINPLATE,S1-2 ... 42

FIG 3.50SEM-EDX,S1-2 ... 43

FIG 3.51THE ELEMENTAL MAPPING OF CU,ZN CONCENTRATED ONLY IN THE PUNCHED AREA ... 43

FIG 3.52 ... 43

PUNCH MARK ON NATURALLY CORRODED TINPLATE THEN BEING ARTIFICIALLY CORRODED,S1-1 ... 43

FIG 3.53SEM-EDX,S1-1 ... 44

AREA NO.2 WAS COMPOSED OF SN,O,FE,S; THE COMPOSITION OF AREA NO.3 WAS THE SAME AS FOR NO.2. ... 44

FIG 3.54 ... 44

NATURALLY CORRODED TINPLATE THEN BEING ARTIFICIALLY CORRODED,S5-1 ... 44

FIG 3.55SEM-EDX,S5-1 ... 45

FIG 3.56SEM-EDX,S5-1 ... 45

FIG 3.57SEM-EDX,S5-1 ... 45

FIG 3.58 ... 46

PUNCHED MARK ON UNCORRODED TINPLATE THEN ACCELERATED CORRODED,S6-1 ... 46

FIG 3.59SEM-EDX,S6-1 ... 46

FIG 3.60THE SCHEMATIC DIAGRAM OF LAYER STRUTURE OF NEWLY PRODUCED TINPLATE ... 50

FIG 3.61THE SCHEMATIC DIAGRAM OF LAYER STRUCTURE OF CORRODED TINPLATE ... 50

FIG 3.62THE TRADE MARK UNDER NORMAL LIGHT ... 52

FIG 3.63THE TRADE MARK UNDER UV ... 52

FIG 3.64THE EARLIEST RECORDS OF THE “PIDO” ... 53

FIG 3.65THE ADVERTISEMENT FOR “PIDO” ... 53

FIG.3.66THE TIN CAN IN POSITION FOR NT EXAMINATION ... 53

FIG 3.67THE TIN CAN IN POSITION FOR X-RAY EXAMINATION ... 53

FIG 3.68NT IMAGE OF THE TIN CAN ... 55

FIG 3.69THE DISC-SHAPED SHEET WITH FOLDS ON IT ... 55

FIG 3.70MICROSCOPIC IMAGE THROUGH THE HOLE ON THE LID ... 55

FIG 3.71THE NT IMAGE OF DRY CANDIED GINGER ... 55

FIG 3.72THE NT IMAGE OF PICKLED CANDIED GINGER ... 55

FIG 3.73X-RADIOGRAPHY OF THE TIN CAN,75KV,5.8 MA ... 56

FIG 3.74A CROSS-SECTION OF DOUBLE SEAMING IN A CONTEMPORARY TIN CAN ... 57

FIG 3.75A DARKER RING IN THE MIDDLE OF THE JOINTING AREA OF BOTTOM LID UNDER X-RAY ... 57

FIG 3.76A PICKLED CANDIED GINGER ... 58

(10)

Table of tables

TABLE 1.1.THE CONDITIONS REGARDING THE AUTHENTICITY OF INSCRIPTIONS ON METAL OBJECTS ... 1

TABLE 3.1THE TIN LAYER THICKNESS MEASURED ON THE CAN SURFACE ... 23

TABLE 3.2 THE CORROSION PENETRATION RATE ON PURE TIN ... 27

TABLE 3.3POSSIBLE SCENARIOS OF INSCRIPTION APPLICATION AND MATERIALS NEEDED ... 28

TABLE 3.4COLLECTED TINPLATE ... 30

TABLE 3.5THICKNESS OF TINPLATE AND TIN LAYER MEASURED ON COLLECTED SAMPLES ... 31

TABLE 3.6 THE CORROSIVE AGENTS TEST FOR TINPLATE, LASTING 144 HOURS ... 34

TABLE 3.7SURFACE FEATURE TYPES OBSERVED FROM THE OBJECT ... 47

TABLE 3.8 SCHEMATIC DIAGRAM OF SURFACE FEATURES ON THE OBJECT AND THE STRAIGRAPHY ... 48

TABLE 3.9 THE REPRODUCED SURFACE FEATURES ... 49

TABLE 3.10 SCHEMATIC DIAGRAM OF REPRODUCED SURFACE FEATURES, AND CROSS-SECTION ... 49

TABLE 3.11SURFACE FEATURES REPRODUCED IN THE RECONSTRUCTION EXPERIMENT ... 51

TABLE 3.12.THE PARAMETER OF REPRODUCTION OF PUNCHED SURFACE ... 60

TABLE 3.13 PROBABILITIES OF DIFFERENT CONDITIONS ON THE INSCRIPTION AND THE CAN BASED ON THE RESEARCH RESULTS ... 63

(11)

1 Introduction to the research project

1.1 Research background

There are many sophisticated examination methods to identify whether a metal object is authentic or fake, either by composition analysis or comparison of the techniques used with the historical context. Even though the object itself is genuine, there is a possibility that an inscription on it might be fabricated later to increase its historical value. Identifying inscribed surface features on metal objects provides another approach in addition to style analysis or vocabulary study. This is especially suitable for those inscriptions that are not made by a metal worker or cannot be attributed to a style in a certain period.

When discussing the authenticity of an inscription, there are four conditions that can be encountered. In condition 1, both the object and the inscription are genuine; in condition 2, the object is authentic, but the inscription is suspicious; in condition 3, the object is suspicious while the inscription is authentic; in condition 4, both the object and the inscription are suspicious. Table 1.1. explains each circumstance regarding the authenticity of inscriptions. As the inscription could not have existed without its carrier, both the object and inscribed marks must be consistent with their context. Therefore, this is the definition for being an authentic inscription in this research project.

Table 1.1. The conditions regarding the authenticity of inscriptions on metal objects

1 Object ○ Both object and inscription are genuine and correct in regard to their material, history, and context. Inscription ○

2 Object ○ The genuine object with clear provenance, but the inscription is added to increase the value.

Inscription ?

3 Object ? An inscribed area from a genuine fragment piece is cut off and then fabricated onto a different object. Inscription ○

4 Object ? Neither object and inscription are correct to their material, history, and context. A pure forgery. Inscription ?

Two study cases in this research offer a rare opportunity to practice approaches to studying the inscriptions on historic metal objects. The first object, the “Peelhelm”, from the collection of the Rijksmuseum van Oudheden in Leiden, was excavated in 1910 in the Netherlands (Fig 1.1). The helmet is a ritual object that belonged to the sixth equites stablesiani in the late Roman period. It is one of the most important and complete gilded silver helmets from the Roman Empire. The inscription

(12)

‘STABLESIA VI ’ was not recorded in its earliest documentation; therefore, there is some suspicion that it might have been added later to increase the historic value.

The second object, the “Geconfyte gember” tinplate can, from a private collection, was purchased from a local flea market several years ago (Fig 1.2). ‘ANNE MARGOT 1944 FRANK’ is inscribed on top of the lid. Anne is the well-known Jewish author of “The Diary of a Young Girl” who, together with her older sister Margot, was arrested by the Nazis in the August of 1944 after years of hiding. After being found, the sisters were transported to a concentration camp in 1944, where they died seven months later. The inscription would greatly enhance the value of this ordinary industrial product, if it can be confirmed that the inscriptions on it were made by the famous indicated owner.

Fig 1.1 the Peelhelm Fig 1.2 The Geconfyte gember can

The authenticity study of the inscription is essential for the historical research in both cases. It also offers a chance to broaden the knowledge of authentic inscription identification.

1.2 Literature review

Ever since the 1970s, a number of studies have scientifically investigated tool marks on various inorganic objects. The two major interests of these studies are: a) finding out what type of working tool was employed to shape or finish the object, and b) evaluating the applicability of different identification techniques. In terms of the inscriptions, the methods can be generally divided into visualizing the inscribed surfaces, and comparing the composition of genuine and fake inscriptions.

Casson (1935) was the first to discuss different visual effects caused by specific tools used to make the inscriptions on Greek metal objects. Fifty years later Larsen (1987) used SEM and silicon rubber imprints to identify marks caused by pressing tools on archaeological metal objects. He advocates the idea of systematically identifying and

(13)

documenting tool marks as a reference data bank for future comparative analysis. This ambition is being realized in an ongoing project of the Metropolitan Museum of Art which aims to document and characterize tool marks on Egyptian hard stone sculptures as a reference for unprovenanced objects.1 However, the focus of this project is on identifying the tools that could only be present at a certain period and area. Also, the vacuum environment of the Scanning Electron Microscope (SEM) and silicon rubber are often limited to be usable on porous or fragile objects.

The order of application sequence of tool marks could be established based on the mark left on the corrosion layer. Certain sources mention that designs carved into bronze corrosion layers indicate they were added recently (Beale, 1990); or that finding marks of grinding underneath the inscription suggests that the inscription was added after the grinding process (Li et al., 2011). A correlation between corrosion layers and tool marks can be determined by surface examination, as long as there is a corrosion layer on the object and the evidence has been preserved well.

To summarize, scientific research into the authenticity of inscriptions has three major approaches. The study of Meijer et al (2005) on three pieces of archaeological objects is a great example to illustrate how these approaches can debate the authenticity issue on metal objects. These three approaches are as follows: to date the compositional material of the object by using the isotope technique; to analyze the chemical composition of the inscription, because some forgery procedures involve adding foreign material onto the original surface; and to compare the association between the tool marks and the corrosion layers. This example reflects the issue addressed in the research background section, as both object and inscription are required to be correct to their context regarding the authenticity of the inscription. The examination methods are divided into invasive and non-destructive types, with or without taking samples from the study object. In most cases, sample taking is not allowed, nor is destructive examination of any kind that would influence the surface texture or material nature. The literature survey reveals that the issue of authenticity is most often encountered in connection with archaeological findings. Authenticity issues for inscriptions on industrial objects have been little discussed.

1.3 Research objectives

The objective of this research is to identify the inscribed surface of the study cases by non-destructive methods, and to discuss the implementation of the methods used in the study.

(14)

1.4 Research questions

The research question of this project is to investigate whether the inscriptions were added in the correct context. To be more specific, for the Peelhelm, whether the inscription “STABLESIA VI ” was applied before burial or after excavation. In the case of the “Geconfyte gember” tinplate can, the question is whether the inscription was punched as indicated in 1944, or later.

Theoretically, marks inscribed on the metal surface in different stages will leave different surface features, based on the relation to the surrounding metal surface or corrosion products. What are the authentic inscription features on the two objects are the sub-questions. By successfully identifying the surface features, the order of sequence of inscription application might be established. This, together with the other evidence found in the study object, may serve to answer the main research questions.

1.5 Methodology

Due to time constraints, only surface examination was carried out on the Peelhelm. For the can, the complete method, consisting of all three procedures, was conducted. To understand the study objects better, the knowledge about materials from literature will be confirmed by the initial object investigation and examination, followed by the reconstruction as comparison reference regarding the inscription’s authenticity. Technical literature research on the two study cases was conducted, including the historic background, manufacturing techniques, and material degradation. This provided a basis not only for the examination but also for the reconstruction experiments. For the object with clear provenance, in this case the Peelhelm, the collection history and conservation records were investigated, along with the material analysis. For the tinplate can, the main purpose of the investigation was to find out approximately when it was produced. Due to the limitations of analysis techniques and condition of the objects, only non-destructive techniques were employed to characterize and document the surface textures of the inscriptions.

The purpose of reproducing inscriptions on metal sheets is to represent the surface features under controlled conditions; this explains the surface features observed on the object. Theoretically, inscriptions applied onto the metal surfaces a long time ago differ from freshly made ones. This will be confirmed by the reconstruction phase of the research. Because the tin can is unprovenanced an investigation on the can itself was carried out after the inscription study to support the authenticity of the inscription.

(15)

1.6 Overview of the research structure

Although the investigation of the inscription is the main topic shared by both study cases, the two objects differ in every aspect. The investigations follow similar steps, but for the can an extended reconstruction of the inscription and background study of the object were executed. Figure 1.3 outlines the research structure. For the purpose of readability, the research performed on each object will be described and discussed in separate chapters. A comprehensive conclusion and discussion will be given in at end.

Fig 1.3 The flow chart of the research

Examination of the authenticity of inscriptions on historical metal objects

Literature research

Scientific examination

Discussion Conclusion

Reconstruction of inscriptions

The Peelhelm _{The tinplate can}

(16)

2 Study case 1: “Peelhelm”, the gilded silver helmet

The Peelhelm falls into condition 2 for the authentic examination of inscriptions (see table 1.1). The object is an archaeological find with a clear provenance, but one of the inscriptions is suspicious. The main goal in this chapter is to discover evidence that supports one of two assumptions: that the inscription existed before the excavation, or that it was fabricated later. An overview of the object description and gilding techniques from the period will be given first in order to clarify the research question and to provide a fundamental knowledge of the material. The second part will focus on the examination of the inscriptions and results. Figure 2.1 is an overview of the helmet and the location of the inscriptions.

Fig 2.1 An overview of the Peelhelm and the inscriptions

2.1 Object description and current knowledge

According to the literature (Evelein, 1911; Pouls, 2006), the “Peelhelm” helmet was excavated by a local peat cutting worker on June 17th, 1910 near the town Deurne, the Netherlands. The helmet was found together with other finds such as bronze coins from the Constantine period, bronze horse bells, and some organic leather and fabric objects. Soon after it was dug out from the ground, it was acquired for the RMO collection in September 1910. In December 1910, the fragment piece that was used to attach the right cheek flap to the cap was sent to the RMO by another peat cutter (indicated in fig 2.2).2 A sketch of the helmet and a telegram made by the amateur art historian Van Beurden on June 20th, 1910 are the earliest record of the finding.3 In the sketch of the helmet (fig 2.3), only one of the inscription “M TITVS LVNAMIS LIBR

2_{Pouls, 2006: 12} 3_Ibid

(17)

I-L” on the neck protector was illustrated. This document is the origin of the suspicions regarding the inscription’s authenticity, as it is unclear why the inscription “STABLESIA VI ” on the right cheek flap was not recorded in this sketch.

Fig 2.2 The inscribed surface “STABLESIA VI ” on the right side of the cap; the arrow indicates the fragment acquired separately in

December 1910.

Fig 2.3 The earliest sketch of the helmet. Image is taken from

the RMO website.4

The helmet is in Constantine style; the metal workers were of non-Roman origin and the decoration is eastern style.5 The reason why the helmet was buried in Deurne remains mysterious,6 but the period of the helmet can be dated in 4th century according to the date on the bronze coins near it.7 The helmet is made of gilded silver sheets affixed by rivets to an original iron inner cap, which has almost completely rusted away; only the gilded silver exterior of the helmet has been preserved.8 In the earliest academic paper on the Peelhelm which was published in the journal

Oudheidkundige Mededeelingen van het Rijksmuseum van Oudheden te Leiden in

1911, the interpretation of both inscriptions was discussed. The inscription “STABLESIA VI ” indicates the ownership belonged to the 6th division of the equites stablesiani in the late Roman period; while “M TITVS LVNAMIS LIBR I-L” suggests that the weight of the gilded silver exterior was equal to 368 grams (LIBR I-L) and was probably inscribed by the certifier M Titus Lunamis (M TITVS

4 _{http://www.rmo.nl/tentoonstellingen/archief/de-gouden-peelhelm/een-telegram-en-een-tekening}_,

accessed 03-06-2018

5_{Pouls, 2006: 27}

6_{Some believe the helmet was given a ritual burial after the owner had completed his military service,}

whereas others believe the owner left the whole case of ritual objects by accident. However, no human remains were found at the site.

7_{the bronze coins are datable between 315-319. Evelein, 1911: 133; Johnson, 1980:311.} 8_Ibid.

(18)

LVNAMIS).9

The historic value of the helmet is enhanced by the meaning carried by the inscription “STABLESIA VI ”. Evelein (1911) points that out the equites stablesiani were cavalry corps and have been recorded in the late Roman Empire document Notitia

Dignitatum that details the administrative organization of the eastern and western

Empires. Since these cavalry troops were scattered during the late Roman Empire in several provinces, it is agreed that the Peelhelm contributes to the history of this corps in northern regions.

When the helmet was found, the iron interior was rusted away and only the leather inside kept the helmet together. The peat cutter removed the leather and caused damage to the helmet.10 According to the archives provided by the RMO and literature, the helmet has gone through several cleaning process since it was excavated. It was first washed in a nearby canal by peat cutters right after it was found, and a second wash and polishing were given by the wife of the finder. Before 1911 the helmet was restored by the goldsmith L. Verkuil, working at the studio of van Rossum du Chattel in Leiden, and the reconstructed right cheek flap was added in this restoration. In the 1970s, a large-scale restoration was conducted. On the basis of archival picture, it is assumed that the current filling of the missing area was done in that period (fig 2.4 and 2.5). The latest conservation was carried out around 10 years ago (fig 2.6). Since then, the gilded silver surface has become darkened and tarnished. It is not documented what kind of treatment was carried out and what kind of materials were used during each conservation or restoration. However, it is certain that the original surface after excavation has been altered by the cleaning and/or polishing process over the past 100 years.

Fig 2.4 The helmet after restoration in 1970s. Image

provided by RMO.

Fig 2.5 The damage area without filling, 1970s. Images provided by

RMO.

Fig 2.6 When the helmet after the latest conservation.

Image provided by RMO.

9_{Evelein, 1911: 133-151} 10_{Pouls, 2006: 11}

(19)

2.2 Preliminary examination and gilding techniques in late Roman period

For the purpose of gaining basic knowledge of the gilded silver surface, literature research on the gilding techniques and non-destructive elemental analysis were carried out.

Gilding is the application of a layer of gold onto the surface of a less precious metal for decoration. The gold adheres to the metallic substrate by either mechanical or physical methods.11 Generally, the gilding methods in the ancient worlds varied in procedural details but can be summarized into three major types (Lechtman, 1971; Mango, 1986; Oddy, 1988; Oddy, 1993). First type involves using adhesives to attach the gold alloy leaf to the substrate. Normally, the leaf gilding is applied after the object is fabricated. The second type is diffusion gilding, performed by laying the gold leaf on the sheet surface and gently heating the sheet afterwards. The interdiffusion of gold and the base metal takes place during the heating process to create the bonding between two metals. The gold surface may then be burnished, and the gilded metal can be embossed or worked in other ways. The last type is the well-known fire gilding technique by using the gold amalgam process. This involves spreading a mixture of gold and mercury onto a scrupulously cleaned metal surface. Heating then causes the mercury to evaporate, leaving behind a well-bonded layer of gold.

Since the helmet was made around the 5th century and has an eastern origin, the focus on gilding techniques of that period and area gives a precise direction. Dandridge (2000) points out that studies of early Byzantine silver objects suggest that four different techniques were employed for their gilding. The techniques include amalgam of gold and mercury applied over a cleaned substrate followed by gentle warming, spreading amalgam of gold and mercury over a surface which has been pre-amalgamated with mercury followed by gentle warming, gold leaf applied over a surface pre-amalgamated with mercury followed by gentle warming, and diffusion gilding.12 With Byzantine silver, there seems to be have been a preference for mercury gilding.

The straightforward way to determine the presence of fire-gilding is mercury detection by X-ray fluorescence spectrometry since the mercury is never driven off completely in the heating process.13 X-ray fluorescence analysis is a qualitative and semi-quantitative surface elemental detection technique, although it is sensitive to low

11_{Oddy, 1993:171} 12_{Dandridge, 2000: 123} 13_{Andrew Oddy, 1993: 177}

(20)

concentrations of elements that could be useful to determine trace elements. The XRF analysis was conducted on the helmet and results were interpreted by Arie Pappot (RMA) (fig 2.7).14 The results showed there was no mercury detected on the helmet, which excluded the possibility of fire-gilding techniques. It is possible that the gilding technique used on the helmet is diffusion gilding. Furthermore, the reconstructed right cheek flap was made from a copper alloy instead of a silver alloy.

Selwyn (2000) declares that the stability of gilded objects depends on the thickness, adherence and porosity of the gold layer, all of which depend upon the manufacturing techniques.15 Both diffusion gilding and fire gilding create a strong secured gold layer; the only difference is that fire gilding leaves a porous surface. In comparison, the diffusion-bonded gilded surface will become less corroded as long as the gold layer is continuous and well secured.16 On the basis of the gilding technique used and the conservation history of the helmet, a surface with less amount of corrosion product in the inscription would be expected.

2.3 Inscription examination methods

Due to the short stay of the Peelhelm at Ateliergebouw (April 9and 10, 2018) and efforts to avoid transporting the object from one examination area to another, only visual examination of the inscription was carried out on the Peelhelm. The aim of examination techniques used to document the surface features will be explained in the following subchapter.

Ultraviolet illumination

UV fluorescence serves as a quick and useful indicator for certain materials or additional foreign materials on the surface, based on the different florescence reaction of different material. In some forgeries, pigments mixed with organic adhesives are used to create the patinated appearance on the newly made inscription surface. Many types of organic adhesive have a distinguishable UV fluorescence: for example, the natural adhesive tends to have a warm tone fluorescence, whereas the synthetic one shows a cold tone fluorescence. Sometimes the polishing material residues left in the grooves or surface coatings also have distinguishing fluorescence reactions which differ from metal or corrosion products. A handheld UV lamp with a wavelength of 360 µm from the Rijksmuseum Amsterdam (RMA) metal conservation department

14_{The Olympus Delta handheld X-ray fluorescence spectrometer was operated by Arie Pappot (RMA)}

under the condition of 40kV, 35 mA and 13 seconds live time, spot size app. 8 mm2_. 15_{Selwyn, 2000: 26}

(21)

was used on the helmet to check for any added material on the inscriptions.17

Optical microscopy

The aim of inscription inspection is to document the surface features and the tool marks left on the material. The reactions of the gilded silver surface to the applied, force can be identified by optical microscopy examination. A digital microscope Hirox KH7700 with a MX(G)2016Z lens, UvA and RMA, was used for surface examination of the helmet (fig 2.8).

Fig 2.7 XRF analysis on the Peelhelm Fig 2.8 The Peelhelm was examined under the microscope

2.4 Results

Visual examination and ultraviolet illumination of the helmet in general

There was no foreign material added to the helmet surface or the inscription, based on the UV examination. The warm tone of UV fluorescence indicated a natural adhesive was used to fill the lost areas (fig 2.9) during the previous restoration during 1970s. The visual examination, stain-like material and scratches (fig 2.10) suggest there might be a layer of coating on the gilded silver surface. However, there was no significant fluorescence to support that assumption under the UV examination. A plastic thread that was used to fix the right cheek flap in the past restoration showed a cold tone fluorescence.

17_{The fluorescent handlamp produced by Bresciani s.r.l is equipped with a Florescent UV lamp 9W}

(22)

Fig 2.9 The lost area glued to filling material Fig 2.10 Suspicious scratches on the left flap

Visual examination of the inscriptions

The inscribed mark

The tools used to inscribe “MTITVS LVNAMIS LIBR1-L” (inscription 1) and “STABLESIA VI ” (inscription 2) both had a rectangular tip, judging from the shift of strokes (fig 2.11 and 2.12).

Fig 2. 11 The tool used on inscription 1 had rectangular shape

Fig 2.12 The tool used on inscription 2 had rectangular shape

However, the width of each stroke in the former inscription is slightly smaller than the latter (fig 2.13, 2.14). This suggests that the size of the tools used on the two inscriptions was slightly different. The punch tool for inscription 1 was sharper than the one used for inscription 2.

(23)

Fig 2. 13 The width of strokes in inscription 1 Fig 2. 14 The width of strokes in inscription 2

A comparison of letters that are present on both inscriptions was performed, and a different inscribing style could be ascertained apart from the writing style. The letters on “MTITVS LVNAMIS LIBR1-L” were composed of lighter, smoother and flowing inscribing lines (fig 2.15); the letters on “STABLESIA VI ” were punched relatively heavily, and the strokes paused or appeared to be less fluent (fig 2.16). The other difference in terms of the punching method can be noticed on inscription 2: the end of strokes occasionally became wider, which is probably due to the tilting angle of the punch (fig 2.17).

Fig 2. 15 Flowing line on the inscription 1 Fig 2. 16 Heavy paused line on inscription 2

(24)

The gilded metal surfaces

The gilded silver substrate where “MTITVS LVNAMIS LIBR1-L” was inscribed on the neck protector was intact, and less distorted or damaged (fig 2.18, 2.19).

fig 2.18 The neck protector. Image provided by RMO.

fig 2. 19 The edge on the cap. Image provided by RMO.

In contrast, the silver substrate baring “STABLESIA VI ” has suffered from great physical distortion. An interesting discovery was the space of the inscription on the cap, which was different at the opposite side of the helmet. The embossed patterns are not continuous through the inscribed area, and there were no signs of being flattened by later alteration. Two ends of the decoration patterns next to the inscribed area were designed to stop at the borders (Fig 2.20, 2.21). The inscribed area seems to have been left blank on purpose after the helmet was fabricated.

Fig 2.20 the left end of the decoration pattern Fig 2. 21 the right end of the decoration pattern

The inscription features of “STABLESIA VI ” in current state

During the examination, a crack was found that went through the left side of the first letter ‘S’ (fig 2.22). The stroke through the letter was continuous. At the edge of the crack, the gilded sheet was distorted by the inscribing tool. Additionally, the corroded interior wall of the crack suggested that the damage was not recent. Another feature

(25)

was noticed in the last symbol “ ” of the inscription. When strokes meet the crinkles in the sheet material, the inscribed lines follow the surface curves instead of being stopped or discontinuous (fig 2.23).

(26)

2.5 Discussion

Under microscopic examination the suspicious inscription “STABLESIA VI ” on the right rim of the cap did not demonstrate signs of recent application. However, it could not be ignored that, if it is forgery, it would then have been applied between 1910 to 1911. The evidence the author was searching for must have survived in the past 100 years, considering the fact that several cleaning or polishing treatments have been carried out on the helmet in the past. The corrosion product or the encrustation from the burial soil on the metal surface probably has been removed almost completely. The authentication examination of the inscription could not rest on the observation of the corrosion layer. Accordingly, the relation between the punch strokes and the surrounding metal play the essential role in indicating the order of sequence. It is also necessary to compare the current state to that of the historical images.

Two significant indications of order were found in the observation of the inscriptions’ current state. First was the crack in the gilded silver breaking through the letter “S”, suggesting that the crack was formed after the inscription was applied. The second feature was the inscribed line of strokes stuck to the curve of the wrinkled surface instead of being distorted or discontinuous on the last mark “ ”. It would be difficult to keep the designated path and direction when inscribing a crumpled surface. As a result, this featured punched line represents the application before the metal substrate became wrinkled.

A comparison with the earliest photography was carried out to trace the existence of the above evidence as found in its current state. The crack was documented in the 1911 photographs (fig 2.24, 2.25), whereas the crinkle did not appear in 1911(fig 2.26, 2.27). The crinkle was proved to be a post-alteration by comparison with the archival image. Verifying this feature could be a sufficient indication for the order of the application sequence. Therefore, crinkles were examined thoroughly in the 1911 photographs (detail images can be found in theAppendix I). This similar feature was found in the letter “L” and “I” (fig 2.28), which implies that the inscription was applied prior to the surface distortion. Both features represent post-application damages, thus two possible surmises could be drawn concerning the inscription. First, if it was a forgery, the inscription was added to the helmet in 1910 or 1911. The crack and crinkles were caused by the early treatment or from the handling after the inscription application and were preserved during the first restoration. Second, it was an authentic inscription and existed before the burial, and the damage was created when it was buried for more than fifteen hundred years.

(27)

Fig 2.24 “S” in 1911 photograph Fig 2.25 The current state of “S”

Fig 2.26 The second “ ” in 1911 photograph Fig 2.27 The current state, crinkle over “ ”

Fig 2.28 The crinkles found in the 1911 photograph went over “L” and “I”

Judging from the early photographs made in 1911, the inscriptions “M TITVS LVNAMIS LIBR I-L” on the separate neck protector were much easier to decipher than the other ones inscribed on the right rim of the cap. This might explain why “STABLESIA VI ” was not illustrated in the earliest color sketch. In addition, the examination suggests that two inscriptions were probably punched by different persons with different tools and styles. The sophisticated punching of “M TITVS LVNAMIS LIBR I-L” reflects what art historians believe, that it was inscribed by the certifier after it was manufactured. On the other hand, the space where “STABLESIA

VI ” was inscribed seemed to be left blank on purpose for the user or owner of the helmet. The other indirect evidence on the inscription is the fact that there was no

(28)

fluorescence reaction from foreign material under the UV examination, indicating that the appearance of the inscription was not intentionally altered by adding organic coatings.

The current surface observation of the inscription “STABLESIA VI ” and historic archives provided evidence to support that it was not a recent application. Also, the corrosion on the edge of the crack did not seem like to be formed after excavation. A reconstruction of the inscribed surface and material analysis might supply more insights to support the authenticity study. Reproducing stroke lines on a crumpled metal surface visualizes the influences from the surface morphology on the inscription. But to characterize the inscribed surface before or after excavation, the material degradation factor has to be considered in the reconstruction. The microstructure of gilded silver under long-term burial might potentially affect the inscription performances on the metal surface. To summarize, reconstruction is a method which could potentially be used for different approaches to serve different research questions. Due to time constraints, limited research was conducted on the material and manufacturing techniques of the helmet which are crucial to characterize the inscribed surface and to prepare a reconstruction. Furthermore, the highly reflective surface of the gilded silver restricted the visual examination of the inscription. A polarized optical microscope would be recommended for future research.

(29)

3 Study case 2: the “Geconfyte gember” tinplate can

This tinplate can falls into the type 4 condition in terms of the authenticity of inscription because it is unprovenanced (See table 1.1) In contrast to the Peelhelm in the first case study, the tin can is a common twentieth century industrial product. Like many other industrial consumer goods, food cans are mass produced all over the world by similar manufacturing processes. It is the historic value carried by the inscription on this can that makes it unique. In this study case, the authenticity of the inscription is defined by the period when it was applied, instead of who punched the mark. The goal of the chapter is to characterize the inscription surface and compare it to the reproduced surfaces under different conditions, in order to clarify whether the inscription is an old or fresh application. Despite the fact that the inscription appeared to be an old application, the prerequisite for an authentic inscription is that both the punched mark and the can have to be genuine. Thus, a historic investigation of the can follows the inscription examination. Figure 3.1 gives an overview of the object and the location of the inscriptions.

Fig 3.1 An overview of the “Geconfyte gember” tinplate can

3.1 Object description and current knowledge

The can is presumed to have contained candied ginger as is written on the label, and has once been opened and resealed by soldering along the lid. The yellowing and fading wrapping label on the body is color printed with fruits in the middle, and the English word “confectionery” above it; while beneath it a label printed with the old Dutch term “Geconfyte gember”, which translates as candied ginger, is stuck onto the wrapping label. The dull dark grey appearance of the can suggests it has been

(30)

oxidized and some brownish iron corrosion deposited on the surface. Several tiny holes can be found on both the top lid and bottom end with or without stains around them. Handling of the object reveals that a lump of unknown material is felt inside the container. Some unknown solid substances can be found in the storage box; they are assumed to have fallen out through the holes in the base of the can. On the lid, the inscription “ANNE MARGOT 1944 FRANK” is composed of punched dots (fig 3.2), and there are marks applied in the same way on the bottom end of the can. The provenance of this object is unclear; it is claimed to have been bought a few years ago in a local flea market by the private collector and stored in an attic for a period of time. The object is currently on loan to the Anne Frank Museum, Amsterdam.

Fig 3.2 The inscribed surface “ANNE MARGOT 1944 FRANK” on the lid

3.2 Initial investigation of the material and condition of the can

A preliminary examination of the can was carried out, including inscription identification on the bottom end, surface observation of the tinplate, the elemental analysis on the can, and tin layer thickness measurement. Due to the lack of material knowledge of modern tinplate from a conservation perspective, literature research on the tinplate and interviews with representatives of the steel industry served as an important aid prior to the inscription examination. The comprehensive result will be discussed in this section.

Inscription identification on the bottom end

The inscription on the bottom end was difficult to identify with normal raking light, because the shadow caused by the bumpy surface made it difficult to obtain a clear overview of the inscription (fig 3.3). It also could not be identified by X-radiography since the substance inside of the can interfered with X-ray penetration. Therefore, the reflectance transmission imaging (RTI) technique was employed to capture clear

(31)

images of the inscription.18 RTI is a software-based photographic method that can be used to enhance surface features of an object.19 The profiles created by this technique consist of several images taken with multi-directional light sources under the ‘monkey dome’ (fig 3.4). Then the profiles are processed by the software, RTIviewer, and the user can manipulate the computerized images to maximize the visual effect. Figure 3. 5 shows the stitched images of the individual punched marks. Although each punched mark was documented by the RTI, the inscription itself could not be recognized successfully. Only the one on the furthest right could be identified as the letter “F”.

Fig 3.3 The bumpy surface of the bottom end captrued by Hirox

Fig 3.4 The RTI profiles built by the “monkey dome” illumination system

Fig 3.5 Stitched RTI images of the punch marks on the bottom end

Visual examination and UV illumination of the tinplate can

There was no fluorescence reaction under the UV illumination suggesting that no distinguishable organic coating or paint has been applied on either the can surface or the inscription. However, under examination by normal light a unique tideline pattern can be found on the cylindrical body of the can (fig 3.6), which highly resembles the description of the “mottle” pattern of tin-iron intermetallic compound layer typically

18_{The RTI images were taken by the Micro-RTI Capture (Monkey Brain) system in RMA under the}

instruction of the photograph conservator Martin Jürgens.

(32)

found on hot-dipped tinplate.20 However, the picture provided in the research of the mottle pattern (Daniels, 1935; Hedges, 1960) was without specified magnification or scale bar (fig 3.7). Therefore, it is unclear if the mottle pattern from the reference has the same morphology as can be seen on the tin can. Both hot-dipped tinning and electrolytic tinning processes involve melting of the tin metal, as will be illustrated in chapter 3.3, so both methods could produce this type of pattern. However, the hot-dipped process has a thicker layer of tin-iron compound, so this pattern might be easier to see on the corroded surface. Although the peculiar pattern suggests a possibility that the tinplate might be manufactured by hot-dipped tinning, this cannot serve as solid evidence. The tin layer thickness analysis will be carried out to identify the manufacturing process.

Fig 3.6 The Damascene pattern observed on the tin can

Fig 3.7 The “mottle” pattern recorded in the literature21

Elemental analysis

XRF analysis was carried out on the tinplate can and soldered areas for the chemical composition.22 Due to the layer structure of the tinplate and solders on the tinplate, the quantitative results were not representative. The qualitative results showed that the object surfaces were mainly composed of elements Fe and Sn, while some trace elements Cr, Mn, Co, Ni, Cu, and Zn were detected. The composition of the solders along the side seam and the joining area between top lid to the can body were Pb and Sn. The elemental analysis confirmed that the can was made from tinplate material, and the solder used was tin-lead solder.

Tin layer thickness of the can

The measurement of tin layer thickness not only indicates the manufacturing process

20_{Hedges, 1960: 232}

21_{Picture cited from Hedges, 1960: 232.}

(33)

for the tinplate on the study object; it also helped the author to select the sample sheets for reconstruction. Theoretically, the layer thickness could be determined by the XRF technique once it is calibrated with known reference samples.

The tin layer thickness was measured by calibrated handheld XRF at TATA Steel.23 The measurement of tin coating was expressed by the conventional standard unit g/m2, then converted to μm by the formula 1 g/m2 = 0.138 μm.24 Table 3.1 lists the measurements.

Table 3.1 The tin layer thickness measured on the can surface

Measured area Thickness unit: g/m2 _{Thickness unit:}μm

Top lid 5.09 0.702

Bottom end 4.55 0.628

Body 1 6.6 0.911

Body 2 8.6 1.187

average 6.21 0.857

The common range of tin coating masses is from 2.8 to 15 g/m2 per surface for electrolytic tinplate; and from 11 to 22 g/m2 for the hot-dipped tinplate.25 From the numbers, the can seems to fall into the lower tin coating mass range of electrolytic tinning. However, it has to be kept in mind that the layer thickness was measured by calibrated XRF with contemporary industrially produced tinplate references, rather than historical tinplate. Also, it is uncertain whether the tin coating thickness measured in its current state represent the original condition. According to the interpretation of Hans van-der-Weijde (TATA Steel Europe), it is much more difficult and not economical to use the electrolytic tinning process to create a thicker tin coating such as this. In his opinion, this can might be made from hot-dipped tinplate.26

3.3 Material and manufacturing of tinplate in 20th century

In this section, a technical overview of the tinplate is given because current knowledge of tinplate for conservation study is limited. Most of the references come from the industry and focus on the practical issues encountered in business usage.

23_{The tin layer thickness of the can was measured by the calibrated handheld XRF (Oxford}

X-MET8000 Optimum) at the TATA steel, Research and Development Center.

24_{The density of tin is 7.265 g/cm3 which equals 7,265,000 g/m3; this represents 1 meter thickness}

and can cover 7,265,00 grams per square meter. Therefore, 1,000,000 μm thickness = 7,265,000 g/m2. The result is 1 μm thickness equals 7.265 g/m2, so is the conversion will be used.

25_{Barry and Thwaites, 1983: 176} 26_{Unpublished communication, 2018.}

(34)

Before answering the research question about the inscription, a fundamental understanding of the material and manufacturing should be obtained beforehand. A brief history of tinplate and development will be given first, followed by an introduction to the most important tinplate producing methods. Then the corrosion that commonly takes place on the tinplate will be discussed. This information is not only beneficial for examining the inscription and the reconstructed pieces, but also serves as a reference considering the authenticity of the tinplate can itself in the second part of the chapter.

History and industrial development of tinplate

The term “tin can” actually refers to the tinplate can which has been widely used between 1900 to 1980 before the great advent of aluminum cans.27 Modern tinplate is a thin sheet of low carbon steel coated on both sides with a very thin layer of tin. Tinplate is a product that combines the strength and formability of steel with the corrosion resistance and weldability of tin.

A historical outline of tinplate was sorted out on the basis of the literature research (Hedges, 1960; Gibbs, 1951; Barry and Thwaites, 1983; Corfield, 1985; Morgan, 1985; Brown, 1988).

Even though tin-coated iron techniques were already known to the Romans, the production of iron sheets pre-coated with tin to be fabricated into objects is considered to have originated in Bavaria in the 14th century. The process consists of immersing a hammered iron sheet into molten tin by craftsmen, after which the sheet is worked into a container. Later, the practice spread throughout Europe and in the late 17th century it was slowly developed from a craft into an industry. In the 20th century, revolutionary developments were introduced into the manufacturing process. The first innovation was the cold reduction process of steel in 1927, which gave the steel more strength. The most significant advance was the introduction of the continuous strip electroplating process, which began in Germany in 1934 and was adapted and developed in the USA in the 1940s. With electroplating, a much thinner application of tin could be achieved at lower cost, consequently, the hot-dipping method has gradually been superseded.

However, when exactly these improvements started to replace hot dipping with electroplating varies from area to area. For instance, South Wales was the leading tinplate producer in the world at the beginning of the 20th century, and a great number

(35)

of female workers were hired to polish hot-dipped tinplate until the 1950s.28 In the Netherlands, electroplating was not employed by the local steel factories until the 1970s.29 According to literature, in 1960 about three-quarters of all tinplate was produced by the electrolytic method.30 During the 1980s, over 98 % of tinplate was produced by the electrolytic process worldwide in the major production centers such as the USA, western Europe and Japan.31 Therefore, it can be stated that hot-dipped tinplate was widely used in industry worldwide before the 1980s. Even today, hot-dipping is still commercially available but rare in the manufacturing.32

Material and manufacture of tinplate

Tinplate is basically considered a steel product since the coating layer on average represents only around 0.6 % by weight of the finished product.33 Low carbon steel used in tinplate production normally contains controlled levels of impurities or trace elements. The relatively pure alloy composition not only provides good workability but also gives more corrosion resistance. The production of steel has gone through two major developments which ultimately influenced the properties of tinplate. Single reduced plate was first developed in the 1930s to produce more ductile and thicker steel plate. Later in the 1960s, double reduced plate was developed. The strength and stiffness of the plate is raised and has a very directional mechanical property. Therefore, the thickness is significantly lower than that of single reduced plate.34 The thickness and stiffness of the steel base could be used as an indicator of its production method, and possibly as indirect evidence to support a guess as to when it was made.

Two dominant tinning procedures in the tinplate industry are hot-dipped tinning and electrolytic tinning. The following description of the two tinning processes is based on the explanation of Morgan (1985).

Hot-dipped tinning vs. electrolytic tinning

For hot-dipped tinplate, the steel strip is first cut into sheets which are then individually passed through molten pure tin. A layer of tin will be coated onto the steel substrate. The electrolytic tinning method is a continuous electrodeposition process for tin directly deposited on to the steel strip in the production line. The most distinguishing feature of electrolytic tinning is that different coating weights can be

28_{Owen-Jones, 1987: 42} 29_{Information from TATA steel.} 30_{Hedges, 1960: 194}

31_{Barry and Thwaites, 1983: 182}

32_{Information form TATA steel and E.V Vulcan Co., Ltd.} 33_{Barry and Thwaites, 1983: 177}

(36)

applied on both sides of the steel surface. Normally the electrodeposited tin coating will go through a momentary flow heating treatment above the melting point of tin and then cooled rapidly, the deposited tin on the surface will convert into a lustrous bright appearance. In the modern industry, a passivation treatment in sodium dichromate electrolyte is given to the tinplate to increase the corrosion resistance. As a result, the passivated film is composed of chromium and chromium oxides as well as tin oxides.

The differences between tinplate produced by hot-dipped tinning and the modern electrolytic process can be summarized as follows.

1. In terms of the coating thickness, the coating weight is accurately controlled in the electroplating system. Electroplating perform excellently at coating thinner tin layer, and different thicknesses of tin can be applied to the two sides of the steel surface. In contrast, the hot dipped tinning process cannot achieve different coating weights on two side of the steel. Hot dipped tinning normally produces a uniform coating no lighter than 20 g/m2.

2. The hot-dipped tinning process produces a bright and lustrous coatings directly without flow-melting treatment. The average amount of tin applied to the steel is greater than in electroplating; therefore, a more continuous tin coating and more tin-iron compound FeSn2 can be formed. Research on tinplate has shown that the intermetallic compound FeSn2 plays an essential role in corrosion resistance ability.

External corrosion on tinplate can

The corrosion of tinplate has been intensively studied in the food packaging industry, but his work has focused mainly on the interior corrosion which depends on the food product. Since the inscription on the can is on the exterior of the tinplate, only external corrosion will be discussed in this section.

Tin has a more noble position than steel in the electromotive series, which means the steel will be preferentially oxidized compared to tinplate. In this context, the only protection the tin layer on the tinplate offers is a physical barrier. When the continuous layer of pure tin and intermetallic compound breaks or pinholes are formed, the tin plays the cathode and the exposed iron plays the reactive anode. Heavy localized corrosion can then occur in the breakage areas. That is why literature

(37)

refers to external attack in many ways resembling that on steel with rusting, pitting, and perforation conditions.35

The principal external factors related to the corrosion of metals are the humidity and atmospheric pollutants such as SO2 and H2S. Tin is subject to attack both from the liquid and the vapor phase. Oxide formation occurs more quickly as the temperature increases. In addition, organic and inorganic acids increase the rate of accelerated corrosion. Corrosion rates of pure tin have been measured in exposed conditions for periods of up to 20 years, and the results are given in table 3.2.36

Table 3.2 the corrosion penetration rate on pure tin

Exposed environment Corrosion rate (μm/year)

Rural 0.05

Industrial 0.125-0.175

Marine 0.175-0.275

The corrosion rate of tin is much lower than that of steel;37_{in rural environments the} corrosion rate of steel is 62.4 times that of tin.

Typically, tinplate has a full or partial coverage of tin oxides (SnOX), which may contain stannous oxide (SnO) or stannic oxide, including its hydrated forms. The presence of excessive amounts of these oxides can change the appearance of tinplate.38_{The corrosion color of tin is generally gray, and the shade of gray depends} on the amounts of the the black tin(II) oxide and the white tin (IV) oxide.39

To conclude what has been learned from the initial examination on the can and from literature, the tin layer thickness on the object indicates it might have been produced with the hot-dipped tinning process; hot-dipped tinplate production was almost eliminated in the industry after the 1980s; the dull dark gray surface layer on the can is probably mainly composed of tin oxides; tin is relatively inert therefore the tin coating thickness plays a decisive role in the corrosion behavior.

3.4 Inscription examination methods and reconstruction procedure

This section describes the non-destructive methods which were used to document the

35_{Murphy, 2006: 857}

36_{The result from Hiers and Minarcik using ASTM tests were quoted by Murphy, 2006, p. 853.} 37_{According to Oh et al (1999), the average Corrosion Penetration Rate (CPR) of steel in rural}

environments is 3.12 μm/y; in industrial is 1.65 μm/y; in Marine is 16.36 μm/y.

38_{Kunst, 2017: 9997} 39_{Selwyn, 2004: 141-148}

(38)

surface features of the inscription on the object, followed by an inscription reproduction on tinplate sheets. The aim of the reconstruction is to establish a reference for surface feature identification, and to give some explanations for the surface features observed on the object.

Optical microscopy

The aim of inscription inspection is to document the surface features, the corrosion product build-up, the current condition, and foreign materials accumulated on the surface. A digital microscope Hirox KH7700 with a MX(G)2016Z lens, RMA (fig 3.11) and the optical stereo microscope Leica DM2500M with Axiocam105 color camera and polar lens, UvA were used on the tin can (fig 3.12).

Fig 3. 11 The can was examined with a Hirox KH770

Fig 3. 12 The can was examined with a Leica DM2500M

Reconstruction experiment design

The reconstruction of the inscription on tinplate sheets intends to reproduce the marks on a variety of samples that represent different application conditions. The possible circumstances are listed in Table 3.3.

Table 3.3 Possible scenarios of inscription application and materials needed

The circumstances of the inscription application

Original surface

Post corrosion

1. The inscription was applied a long time ago on

a newly produced tinplate Uncorroded ○

2. The inscription was applied a long time ago on an oxidized tinplate

Partially

corroded ○

3. The inscription was applied a long time ago on

Inscriptions on A Roman Gilded Silver Helmet and A Twentieth Century Tinplate Can: Authenticity Examination and Surface Characterization

Conservation and Restoration Program, Master's, Metals Specialization