Analytical strategies for the identification and characterisation of thin … › bitstream › 10451 › 35143 › 1 › ulsd... · 2018-10-31 · Analytical strategies for the identification

2017

UNIVERSIDADE DE LISBOA

FACULDADE DE CIÊNCIAS

Analytical strategies for the identification and characterisation of thin layers on ancient gold artefacts

Doutoramento em Engenharia Física

Maria Isabel Leandro Pinheiro de Almeida Tissot Daguette

Tese orientada por: Doutor Pedro Manuel Ferreira Amorim

Doutora Maria Filomena Parrela Camisão Guerra Doutora Maria Luísa Dias de Carvalho de Sousa Leonardo

Documento especialmente elaborado para a obtenção do grau de doutor

2017

UNIVERSIDADE DE LISBOA

FACULDADE DE CIÊNCIAS

Analytical strategies for the identification and characterisation of thin layers on ancient gold artefacts

Doutoramento em Engenharia Física

Maria Isabel Leandro Pinheiro de Almeida Tissot Daguette

Tese orientada por: Doutor Pedro Manuel Ferreira Amorim

Doutora Maria Filomena Parrela Camisão Guerra Doutora Maria Luísa Dias de Carvalho de Sousa Leonardo

Júri: Presidente:

● Doutora Margarida Maria Telo da Gama Vogais:

● Doutor Régis Bertholon ● Doutor Salvatore Siano ● Doutora Maria Filomena Parrela Camisão Guerra ● Doutor João Filipe Calapez de Albuquerque Veloso ● Doutora Marta Catarino Lourenço

Documento especialmente elaborado para a obtenção do grau de doutor

Instituições financiadoras: Fundação para a Ciência e a Tecnologia com bolsa de doutoramento em

empresa (SFRH/BDE/51439/2011) e Archeofactu

All that is gold does not glitter, Not all those who wander are lost;

The old that is strong does not wither, Deep roots are not reached by the frost.

J.R.R. Tolkien, The Fellowship of the Ring

i

Acknowledgments

To my supervisors, Professor Maria Filomena Guerra, Professor Luisa Carvalho and Professor Pedro Amorim, I would like to express my sincere gratitude for their trust and tireless support during this stage of my life. Bem-hajam! A special word to Professor Maria Filomena Guerra, for whom I have a deep admiration, and has been the pillar of my scientific education. In addition to her great scientific and personal generosity, she always knew how to compensate exigence with patience and care.

Several people contributed to my scientific development. I would like to emphasise three persons that throughout this time have been a support and for whom I have special affection and admiration: Doctor Alexandra Barreiros (LNEG), Doctor Olinda Monteiro and Professor Jorge Correia (IEG/Departamento de Química e Bioquímica, FCUL), my most sincere gratitude.

To LIBPhys (UNL) and to the Interfacial Electrochemistry Group (CQB/UL) group leaders and members for receiving me in their laboratories and for supporting me during my thesis.

To Professor José Rebordão (FCUL) who supported this thesis by being always available for the resolution of administrative issues.

To the direction and collaborators of the National Museum of Archeology (Lisbon), the Calouste Gulbenkian Museum (Lisbon), the National Museums of Scotland and the Garstang Museum (Liverpool) for supporting the projects that allowed to study the gold objects presented in this work.

To the Foundation for Science and Technology for awarding me with a Ph.D. scholarship in a business context (SFRH/BDE/ 51439/2011).

To Pedro Pedroso, “boss” and great friend, enthusiastic of all projects and that did everything he could for this thesis to move forward supporting it both as Archeofactu responsible as well as a friend.

To my colleagues and friends, Marta Manso and Virgínia Ferreira. Also, my thanks to all those with whom I had the opportunity to work during this time, Sofia Pessanha (LIBPhys), Lore Troalen (National Museums of Scotland), Michel Dubus (C2RMF), Victoria Corregidor (CTN/IST), Luis Cerqueira Alves (CTN/IST), Marc Aucouturier, Rui Xavier (FCG) and the AGLAE team.

To my friends, whom I deeply deprived myself during this time, Mana Bé, Manel (let the party begin!), João Petisca, Rita (Rute), Ana. My deep appreciation for supporting me during this project, even without knowing what I have been doing, and for respecting my moments of isolation.

To Antoine Mattei, for the hours of sun he lost in Lisbon.

To my parents, the pillars of my life, and to my sisters, Joana and Marta, unconditional supporters of my adventures and comforters of my misadventures.

To my children, Catarina and Tomás, and to Matthias for making me believe that life will always be better!

ii

iii

Resumo

Este trabalho define estratégias analíticas para a caracterização de camadas finas de corrosão de

objetos em ligas de ouro expostos ou armazenados em museus. As superfícies corroídas destas

ligas podem apresentar uma variação de cor entre o amarelo e o vermelho, passando por uma

variedade de tonalidades (violeta, azul, castanho, …). Os produtos de corrosão formados alteram,

de maneiras diversas, o aspeto dos objetos, sendo por isso necessário realizar uma intervenção

de conservação e restauro, a qual se deve basear no conhecimento detalhado da camada de

alteração resultante do processo de corrosão atmosférica. O desconhecimento deste processo

levou à necessidade de estabelecer um protocolo analítico que permitisse caracterizar a cor, a

morfologia, a composição elementar e a composição estrutural das camadas finas de corrosão

formadas em substratos de ouro. Para tal, consideraram-se os requisitos para a análise de objetos

de ouro expostos e armazenados em museus, utilizando técnicas não-invasivas e, de preferência,

com equipamentos portáteis. Baseado nas vantagens de cada técnica analítica selecionada e nas

características da camada de corrosão, apresenta-se e discute-se um protocolo analítico.

O estudo desenvolveu-se em duas partes. Na primeira, as técnicas analíticas selecionadas foram

testadas e o protocolo analítico fundamentado na investigação de três casos de estudo: a Coleção

de Ourivesaria Arcaica do Museu Nacional de Arqueologia (Lisboa); as joias da Coleção de René

Lalique do Museu Calouste Gulbenkian (Lisboa); e folhas Egípcias em ouro dos National Museums

of Scotland e do Garstang Museum (Liverpool). Numa segunda parte, e para superar as restrições

relacionadas com o transporte de objetos de ouro dos museus, as mesmas técnicas foram

aplicadas ao estudo de amostras de ligas binárias e ternárias em ouro, fabricadas especificamente

para este trabalho e submetidas a corrosão artificial.

As vantagens e limitações de cada técnica são discutidas e, com base nos resultados obtidos,

evidenciou-se a impossibilidade de caracterizar a camada de corrosão das ligas de ouro in-situ

com os equipamentos portáteis disponíveis. Foi possível caracterizar a morfologia da camada de

corrosão por microscopia eletrónica de varrimento (SEM - Scanning Electron Microscopy) e

analisar a cor das superfícies corroídas por espectrofotometria de ultra-violeta-visível (UV-Vis) e

por elipsometria. Esta última técnica também permitiu determinar a espessura das camadas de

corrosão. A composição elementar foi determinada por espectrometria de Raios-X dispersiva em

energia (EDS - Energy Dispersive Spectrometry) e por espectrometria de fluorescência de Raios-X

(XRF - X-ray fluorescence) e a composição estrutural obteve-se por difração de Raios-X (XRD - X-

ray diffraction).

iv

A caracterização da corrosão dos objetos em ouro desenvolveu-se segundo três eixos: i) avaliação

das condições ambientais, ii) identificação da composição das ligas de ouro e das técnicas de

fabrico e iii) identificação dos produtos de corrosão e da espessura da camada de corrosão. A

possibilidade de deslocar um conjunto de objectos para análise em laboratório permitiu inferir

sobre estes três eixos de investigação e identificar as vantagens e limitações das técnicas

selecionadas para a caracterização da camada de corrosão das ligas de ouro. O estudo da coleção

do Museu Nacional de Arqueologia permitiu relacionar as características ambientais da sala de

exposição com os produtos de corrosão formados, na maioria compostos de enxofre. A

degradação do tecido utilizado nas vitrinas, à base de lã, em conjunto com valores elevados de

humidade e temperatura foram identificados como sendo os responsáveis pela corrosão dos

objetos da coleção. Apesar da composição da liga de ouro ter influência no desenvolvimento da

corrosão, os resultados obtidos para os diferentes casos de estudo não evidenciaram claramente

esse fenómeno. Os processos de corrosão também dependem dos processos termomecânicos,

aos quais o objeto foi submetido durante o fabrico. A observação por SEM de uma joia de René

Lalique demonstrou que defeitos das técnicas de fabrico podem resultar na formação de

microestruturas que influenciam o desenvolvimento da corrosão, criando efeitos de reflexão de

luz diferentes das restantes áreas dos objetos. O estudo das folhas Egípcias permitiu evidenciar

que a corrosão se desenvolve numa estrutura em camadas e que cada camada é constituída por

produtos de corrosão com morfologias e composições distintas.

Para ultrapassar as restrições relacionadas com o transporte de objetos de ouro dos museus,

propõe-se uma metodologia que se baseia no fabrico de amostras de ligas de ouro com as mesmas

composições dos objetos e submetidas às mesmas condições de corrosão atmosférica, para serem

posteriormente analisadas em laboratório. Neste trabalho fabricou-se um conjunto de amostras

de ligas ternárias de ouro, às quais se adicionaram para estudo, ligas binárias de ouro, liga de prata

(prata de lei 925 ‰) e ainda amostras de prata pura e cobre puro, para efeitos de comparação. As

amostras foram corroídas por imersão numa solução contendo enxofre e posteriormente

analisadas com as técnicas analíticas selecionadas para este estudo.

A caracterização das amostras permitiu demonstrar a influência da prata e do cobre no

desenvolvimento da corrosão das ligas de ouro, determinar a espessura da camada de corrosão e

discutir a relação entre a cor das superfícies corroídas e a morfologia dos produtos de corrosão.

Os produtos de corrosão formados nas superfícies das amostras de prata e cobre são compostos

principalmente por sulfureto de prata e sulfuretos de cobre, respetivamente, e as diferentes cores

das superfícies corroídas dependem da espessura da camada de corrosão. No caso da liga de prata

v

e das ligas de ouro, as várias cores das superfícies corroídas estão também relacionadas com a

presença de diferentes produtos de corrosão com morfologias distintas, numa formação de

corrosão por camadas, dando origem a uma estrutura de multicamadas. Para a liga de prata,

revelou-se, para os primeiros instantes de corrosão, uma prevalência da formação de compostos

à base de cobre seguida por compostos à base de prata. Para as ligas de ouro, mostrou-se para a

formação de um filme composto por duas camadas, uma camada interna com 60 nm de espessura

e uma externa com 20 nm de espessura. A fina espessura das camadas de corrosão das ligas de

ouro e a sua composição limitam o uso das técnicas selecionadas. A análise por UV-Vis das

amostras de prata, cobre e liga de prata permitiu diferenciar produtos de corrosão, compostos

por partículas com diferentes dimensões. No entanto, no caso de ligas de ouro, esta técnica só

permitiu demonstrar que existe uma diminuição da refletância nas superfícies corroídas. A análise

por EDS possibilitou a determinação da composição elementar quantitativa da camada corroída

das amostras de prata e cobre, bem como de cada camada constituinte da estrutura de

multicamadas desenvolvida sobre a liga de prata. Porém, no caso de ligas de ouro, a deteção do

enxofre é dificultada pela cauda do pico intenso respeitante às riscas M do ouro muito próximas

das riscas K do enxofre. A análise por XRD permitiu identificar os compostos cristalinos presentes

nas amostras de prata pura e de cobre puro. Contudo, a resolução espacial do equipamento

utilizado é inadequada para identificar os compostos que constituem cada camada da estrutura

multicamada que se desenvolve na liga de prata e nas ligas de ouro.

Palavras-chave: Ligas de ouro; corrosão atmosférica; filmes finos; estratégia analítica;

património cultural

vi

vii

Abstract

The colour, morphology, thickness, elemental and structural composition of the thin atmospheric

corrosion layer that develops on ancient gold objects exhibited or stored in museums were

investigated using analytical techniques chosen after establishing the main requirements for those

objects. An analytical protocol based on the advantages of each selected analytical technique and

on the characteristics of the corrosion layer is presented and discussed.

The study was carried out in two steps. Firstly, non-invasive portable and stationary techniques

were tested directly on prehistoric goldwork in the collection of the National Museum of

Archaeology in Lisbon, on jewellery by René Lalique in the collection of the Gulbenkian Foundation

Museum, and on Egyptian gold foils in the collections of the Garstang Museum in Liverpool and

the National Museums Scotland in Edinburgh. Secondly, and to overcome the restrictions related

to moving gold objects from museums, the same techniques were applied to the study of samples

made from binary and ternary gold alloys fabricated in this work to be subjected to natural and

artificial corrosion in similar atmospheric conditions.

The advantages and limitations of each technique are discussed. Based on the results obtained, it

could be shown that portable equipment for in-situ analysis is not suitable for this type of

investigation, but Scanning Electron Microscopy (SEM) could assess the morphology of the

corroded layer, and the colour analysed by Ultraviolet-visible spectrophotometry (UV-Vis). The

corrosion layer thickness could be estimated by ellipsometry. The elemental composition could

be determined by micro-X-ray fluorescence spectrometry (µXRF) and by Energy Dispersive

Spectrometry (EDS), and the structural composition obtained by X-ray Diffraction (XRD).

The Treasure room of the National Museum of Archaeology exhibits gold objects with an

accentuated corroded surface due to atmospheric corrosion. The main pollutants for indoor

corrosion could be inferred and their sources identified. The environmental conditions were

related to the corrosion products formed on the surface of the objects. Sulphur was identified as

the principal pollutant of gold alloys in indoor environments; the main corrosion products are,

thus, sulphur-based compounds. Data on René Lalique’s jewellery showed the influence of the

fabrication techniques on the corrosion development and the study of the Egyptian foils allowed

relating the colours of the corroded surfaces to the corrosion products formation. It could be

shown that the atmospheric corrosion process leads to the development of a layer-by-layer

structure consisting of corrosion products with different morphologies and distinct compositions.

viii

The characterisation of binary and ternary gold alloys samples corroded in sulphide-containing

solutions was complemented with the study of corroded pure silver, pure copper and sterling

silver samples. The role of the constitutive elements on the corrosion process of gold alloys was

defined, the corrosion layer thickness was determined, and the relation between the colour of the

corroded layer and the morphology of the corrosion products was discussed. The corrosion

products formed on the silver and copper samples surfaces are mainly composed of silver sulphide

and copper sulphides, and their different corroded colours depend on the corrosion layer

thickness. The different colours of corroded sterling silver and gold alloy surfaces are related to

the formation of distinct corrosion products with different morphologies, in a layer-by-layer

structure. For sterling silver, it was revealed that at early corrosion stages, there is a prevalence

of the formation of copper-based compounds, followed by the formation of silver-based

compounds. For gold alloys, it was shown the formation of a two-layer film composed of a 60 nm

thick inner-layer and a 20 nm thick outer-layer.

The corrosion layer thickness of gold alloys and its composition limits the use of the selected

techniques. UV-Vis analysis of the silver, copper and sterling silver samples allowed to

differentiate corrosion products composed of particles with distinct sizes, but in the case of gold

alloys it was only possible to show a reflectance decrease for the corroded areas. EDS analysis

provided the quantitative elemental composition of the corroded layer of silver and copper

samples as well as of each layer of the layer-by-layer structure that developed on sterling silver.

However, in the case of gold alloys the identification of sulphur is difficult due the right tail intense

peak corresponding to the Au-M lines close to the S-K line. XRD identified the compounds present

on the Ag and Cu samples. The spatial resolution of the equipment is, however, inappropriate to

identify the compounds that constitute each layer of the layer-by-layer structure that develops on

corroded sterling silver and gold alloys.

Keywords: Gold alloys; atmospheric corrosion; multi-layered thin films; sulphur compounds;

analytical protocol

ix

Table of contents

Acknowledgments ---------------------------------------------------------------------------------- i Resumo ------------------------------------------------------------------------------------------------ iii Abstract ----------------------------------------------------------------------------------------------- vii Contents ----------------------------------------------------------------------------------------------- ix List of figures ----------------------------------------------------------------------------------------- xi List of tables ------------------------------------------------------------------------------------------ xv Chapter 1 – Introduction 1.1 Gold as a cultural heritage material -------------------------------------------------- 3 1.1.1 Gold alloys ------------------------------------------------------------------------------------- 5 1.1.2 The colour of the alloys --------------------------------------------------------------------- 7 1.2 Atmospheric corrosion of gold alloys ------------------------------------------------ 8 1.3 Conservation of gold alloys objects -------------------------------------------------- 10 1.4 Research objectives and structure of the thesis ---------------------------------- 11 1.5 References ---------------------------------------------------------------------------------

13

Chapter 2 - Analytical methods and instrumentation 2.1 Introduction -------------------------------------------------------------------------------- 19 2.1.1 Analysis of gold objects -------------------------------------------------------------------- 19 2.1.2 Analysis of corrosion layers on gold objects ------------------------------------------- 20 2.2 The analytical approach ----------------------------------------------------------------- 22 2.2.1 Ultraviolet-visible spectrophotometry -------------------------------------------------- 27 2.2.2 Ellipsometry ----------------------------------------------------------------------------------- 29 2.2.3 X-ray fluorescence spectrometry --------------------------------------------------------- 31 2.2.4 Scanning electron microscopy with energy dispersive X-ray spectrometry ---- 34 2.2.5 X-ray diffraction ------------------------------------------------------------------------------ 35 2.3 References ---------------------------------------------------------------------------------

38

Chapter 3 - Atmospheric corrosion of gold alloy objects: case studies 3.1 Introduction-------------------------------------------------------------------------------- 43 3.2 Exhibition and storage environment assessment: the National Museum

of Archaeology, Lisbon 45

3.2.1 The “Treasures of Portuguese Archaeology” collection conservation assessment -------------------------------------------------------------------------------------------

45

3.2.2 Climate and materials characterisation ------------------------------------------------ 46 3.2.3 Characterisation of the corroded layer ------------------------------------------------- 51 3.2.4 Final remarks---------------------------------------------------------------------------------- 54 3.3. Composite objects: the jewellery of René Lalique 56 3.3.1 The jewellery of René Lalique collection ------------------------------------------------ 56 3.3.2 The gold alloy composition and fabrication techniques ---------------------------- 57 3.3.3 Final remarks---------------------------------------------------------------------------------- 62

x

3.4 Ancient Egypt gold objects 63 3.4.1 The corrosion of gold Egyptian objects-------------------------------------------------- 63 3.4.2 Characterisation of the corroded surface ---------------------------------------------- 64 3.4.2.1 Observation ------------------------------------------------------------------------- 64 3.4.2.2 Analysis ------------------------------------------------------------------------------- 65 3.4.3 Final remarks---------------------------------------------------------------------------------- 69 3.5 Conclusions--------------------------------------------------------------------------------- 70 3.6 References----------------------------------------------------------------------------------

72

Chapter 4 - Characterisation of corroded binary and ternary alloys in sulphide environments

4.1. Introduction ------------------------------------------------------------------------------- 77 4.2 Sample preparation 79 4.2.1 Silver, copper and sterling silver samples----------------------------------------------- 79 4.2.2 Gold alloy samples--------------------------------------------------------------------------- 79 4.3. Corrosion test method and conditions --------------------------------------------- 80 4.4. Corroded surface characterisation 82 4.4.1 Silver -------------------------------------------------------------------------------------------- 82 4.4.2 Copper ------------------------------------------------------------------------------------------ 88 4.4.3 Sterling silver ---------------------------------------------------------------------------------- 94 4.4.4 Gold-silver-copper alloys ------------------------------------------------------------------- 105 4.5. Conclusions-------------------------------------------------------------------------------- 114 4.6. References---------------------------------------------------------------------------------

116

Chapter 5 - Towards to a definition of an analytical strategy 5.1 Introduction-------------------------------------------------------------------------------- 121 5.2 Exhibition/storage context assessment---------------------------------------------- 122 5.3 Gold alloy composition and fabrication techniques assessment-------------- 123 5.4 Corrosion products identification----------------------------------------------------- 125 5.5 Conclusions--------------------------------------------------------------------------------- 128 5.6. Outlook ------------------------------------------------------------------------------------- 130 5.7 References----------------------------------------------------------------------------------

132

Appendixes Appendix 1 - Analysis of a gold alloy corrosion layer by Raman spectroscopy - 135 Appendix 2 - Analysis of a gold alloy corrosion layer by PIXE and RBS------------ 141 Publications resulting from this work---------------------------------------------------------- 147

xi

List of figures

Chapter 1 Figure 1.1 Schema of the different stages of the production (from mineral to fabrication

techniques), use, discard and musealisation of gold alloy objects. 4

Figure 1.2 Binary phase equilibrium diagrams of the Au-Ag (A) and Au-Cu (B) systems [10].

5

Figure 1.3 Binary phase equilibrium diagram of the Ag-Cu system [10]. 6

Figure 1.4 Au-Ag-Cu ternary phase equilibrium diagram of some isotherms on the liquidus surface (A) and some isothermal solid state boundaries of the immiscibility field (B). Projections on the room temperature plane [11].

6

Figure 1.5 Relationship between the composition of Au-Ag-Cu alloys and their colour. Adapted from [11].

7

Figure 1.6 Archer armband (c.1700 BC) from the National Museum of Archaeology in Lisbon (Portugal). The corrosion layer shows a reddish colour with a multi-hued effect.

8

Figure 1.7 Schema of the thesis structure. 12 Chapter 2

Figure 2.1 Schematic representation of a corroded gold alloy. 21 Figure 2.2 Simulation of trajectories and ranges 3 MeV protons (A) and helium (B) ions in

a bulk Cu target. The simulation was obtained by SRIM 2003 software [29]. 25

Figure 2.3 In-situ ellipsometric measurements during the corrosion of a gold alloy sample (A) and the electrochemical cell used for the in-situ measurements (B) [47].

31

Figure 2.4 Portable XRF during the analysis of a gold necklace (NMA-Au47) from the collections of the National Museum of Archaeology.

33

Figure 2.5 SEM-EDS equipment (A) and detail of the SEM chamber prior to the exam and analysis of brooch Ophelia (CGM-1138) from the René Lalique collection of the Calouste Gulbenkian Museum (B).

35

Figure 2.6 Analysis of a torc fragment from Serpa (NMA-Au293) of the National Museum of Archaeology with the D8 Discover diffractometer from the Laboratório José de Figueiredo, Lisbon.

37

Chapter 3

Figure 3.1 The Treasure room of the NMA. 45 Figure 3.2 Iron Age torc from Codeçais, Bragança (NMA - Au 1139) and the detail of the

terminal illustrating the corroded surface in 2009 (A) and 2012 (B and C). 46

Figure 3.3 RH and T measurements for one month taken inside (A) and outside (B) the Treasure room.

47

Figure 3.4 RH and T measurements for one month taken inside the Treasure room in a period during which the room was temporarily closed.

48

Figure 3.5 SEM image of the blue fabric showing the wool fibres with scales characteristics of the wool fibres morphologies mixed with polyester fibres, exhibiting a flatter surface [23].

50

Figure 3.6 Bronze Age torc (MNA- Au 293) from Serpa (A); fragment of a Bronze Age torc (MNA-Au 283) from Alentejo (B) and Iron Age earring (MNA-Au 574) from Cabeço de Vaiamonte, Monforte (C).

51

Figure 3.7 SEM image of the corroded surface of the torc from Alentejo (NMA-Au283) composed of small rounded particles. Areas with a more developed corrosion show a structure that can be related to distinct corrosion products.

53

Figure 3.8 EDS spectra of the fragment torc from Alentejo (MNA-Au 283) obtained in areas with and without visible corrosion. The presence of higher Ag and S

54

xii

contents in the corroded areas indicates the possible formation of S-based and Ag-based corrosion products.

Figure 3.9 Details of red corroded surfaces of the pendants Peacocks on prunus (CGM-1203) (A) and Nymph in a tree (CGM-1165) (B).

56

Figure 3.10 Ternary diagrams (Au,Ag,Cu) in wt% of the pendants and brooches analysed by XRF, with the colour (A) and the melting temperatures (B) adapted from McDonald and Sistare [29].

57

Figure 3.11 Dendritic structure at the surface of the bracelet Pair ensnared (CGM-1180), a result of the metal solidification process.

58

Figure 3.12 SEM image of brooch Ophelia (CGM-1138) showing the microstructure characteristic of porosity defects.

58

Figure 3.13 SEM images of corroded areas of brooch Ophelia (CGM-1138). Different morphologies were observed: small rounded particles corresponding to the nucleation of corrosion products (A) and a uniform corrosion film (B).

59

Figure 3.14 SEM image of the file tool marks on brooch Ophelia (CGM-1138). 59 Figure 3.15 Cracks in the gold alloys of pendants The Abduction (CGM-1173) (A) and

Peacocks on prunus (CGM-1203) (B). 61

Figure 3.16 Corrosion of the silver foil on the gold plate under the enamels of the pendant Peacocks on prunus (CGM-1203), due to enamel cracks.

61

Figure 3.17 Fragment foil (GMA-432-25A) from Abydos excavations showing a heterogeneous corroded surface with colours varying from dark yellow, red and blue (A); and bead from Harageh (NMS-A1914.1096) with homogeneous corrosion with a red colour surface (B).

64

Figure 3.18 Au and Ag contents obtained by µXRF for the different corrosion colours of the Abydos foil (GMA-432-25A).

65

Figure 3.19 Au and S contents obtained by μXRF for the corroded and non-corroded areas of the Abydos foil (GMA-432-35B) and the Harageh bead (NMSA1914.1096) with homogenous corrosion.

66

Figure 3.20 SEM images for the heterogeneous colour corrosion surfaces on Abydos foil (GMA-432-25A). Morphology of the corrosion products featuring thin tubes (A); Surface morphology of the red corroded area (B) and of the blue area (C). The blue area is composed of three distinct layers: 1) one nearer the substrate, morphologically similar to the red corroded areas, 2) a second with corrosion products featuring agglomerates and 3) a third with corrosion products featuring thin tubes.

67

Figure 3.21 SEM images for the homogeneous colour corrosion surface of the Abydos foil (GMA- 432-25B) (A); a detail of the polycrystalline tubular formations of the corrosion products (B) and homogeneous colour corrosion on the bead from Harageh (NMS-A1914.1096) composed by corrosion products mixed with silicates from burial (C).

68

Chapter 4

Figure 4.1 Surface colour of the corroded Ag samples immersed during 1, 3, 5, 7, 15, 30, 60, 120, 240, 480 and 1020 min in a 0.1M Na2S aqueous solution.

82

Figure 4.2 UV-Vis absorption spectra obtained for the Ag samples corroded during 1, 3, 5 and 7 min (A); 15, 30, 60 and 120 min (B); 240, 480 and 1020 min (C) in a 0.1M Na2S aqueous solution.

84

Figure 4.3 SEM images of Ag (A) and corroded Ag by immersion in a 0.1 M Na2S aqueous solution during 3 min (B), 5 min (C), 7 min (D), 15 min (E), 30 min (F), 60 min (G), 240 min (H) and 1020 min (I) (scale bar is 2 µm for A to H and 5 µm for I).

85

Figure 4.4 X-ray diffractograms obtained for the 1, 3, 5, 15, 60, 240 and 1020 min corroded pure Ag. The diffraction peaks correspond to Ag2S ( PCD file nº00-014-0072).

86

xiii

Figure 4.5 X-ray diffractogram obtained for the 240 min corroded Ag sample with the identification of the main diffraction peaks of Ag and Ag2S. Ag (PCD file nº 00-004-0783); Ag2S (PCD file nº 00-014-0072).

87

Figure 4.6 Surface colour of the corroded Cu samples immersed during 1, 3, 5, 7, 15, 30, 60, 120, 240, 480 and 1020 min in a 0.1M Na2S aqueous solution.

89

Figure 4.7 UV-Vis absorption spectra for the Cu samples corroded during 1, 3, 5 and 7 minutes (A); 15, 30 and 60 min (B); 120, 240 and 1020 min (C) in a 0.1M Na2S aqueous solution.

90

Figure 4.8 SEM images of Cu (A) and corroded Cu by immersion in a 0.1 M Na2S aqueous solution during 3 min (B), 5 min (C), 7 min (D), 15 min (E), 30 min (F), 60 min (G), 240 min (H) and 1020 min (I). (scale bar is 2 µm for A to H and 5 µm for I).

91

Figure 4.9 X-ray diffractograms obtained for the samples immersed 60, 120, 240 and 1020 min with the identification of main diffraction peaks of Cu2S and Cu2O. Cu2S (PCD file nº00-033-0490); Cu2O (PCD file nº00-077-0199).

93

Figure 4.10 Colour evolution of the corroded sterling silver samples for different immersion times (1, 3, 5, 7, 15, 30, 60, 120, 240, 480 and 1020 min) in a 0.1M Na2S aqueous solution.

95

Figure 4.11 UV-Vis absorption spectra for the sterling silver samples corroded during 1, 3, 5 and 7 min (A); 15, 30 and 60 min (B); 120, 240, 480 and 1020 min (C) in a 0.1M Na2S aqueous solution.

97

Figure 4.12 Experimental values of Ψ vs. Δ collected by ellipsometry for sterling silver samples corroded until 60 min immersion time. The angle of incidence was fixed at 70 ⁰.

98

Figure 4.13 SEM images of the samples prepared by immersion in a 0.1 M Na2S aqueous solution at different times: 0 min, bare silver (A), 1 min (B), 3 min (C), 5 min (D), 7 min (E) and 15 min (F).

100

Figure 4.14 SEM images of the samples prepared by immersion in a 0.1 M Na2S aqueous solution at: 60 min (A), 120 min (B), 240 min (C), 480 min (D) and 1020 min (E).

102

Figure 4.15 X-ray diffractograms obtained for the 2, 4, 5, 15 and 60 min corroded Ag-Cu alloy. ● Ag (PCD file nº 00 004 0783); Cu (PCD file nº 00 002 1225) diffraction peaks.

103

Figure 4.16 X-ray diffractogram obtained for the 60 min corroded Ag-Cu alloy with the identification of the Ag, Cu and Ag3CuS2 patterns. ● Ag (PCD file nº 00-004-0783); Cu (PCD file nº 00 002 1225), □ Ag3CuS2 (PCD file nº 00 012 0207) diffraction peaks.

104

Figure 4.17 UV-Vis diffuse reflectance spectra of sample T6 before (A) and after corrosion test (B) and of pure metals used to obtained the alloy Ag, Cu and Au.

106

Figure 4.18 Experimental values of Δ vs. Ψ () acquired during the corrosion of T6 sample in a 0.1M Na2S solution and the obtained theoretical curves for the two-layer corrosion film: Inner layer (—) and outer layer (—).

107

Figure 4.19 SEM images of the T6 sample before corrosion. 108 Figure 4.20 SEM images of the T6 sample surface after immersion during 1 month (A and

B) and 12 months in a 50 mM Na2S solution (C and D) and during 35 hours in a 0.1 M Na2S solution (E and F).

109

Figure 4.21 SEM images of B4 surface sample after immersion during 12 months in a 50 mM Na2S solution.

110

Figure 4.22 X-ray diffractograms obtained for T6 sample after corrosion in a 50 mM Na2S solution during 12 months with the identification of the main peaks of Au (PCD file nº00-004-0784), Ag (PCD file º 00-004-0783), Cu (PCD file nº 00 002 1225) and AuAgS (PCD file nº 00 038 0396).

112

xiv

Chapter 5 Figure 5.1 Sterling silver box used for testing the surface replica method carried out with

a 30 µm cellulose acetate sheet (A). SEM images of the box corroded layer (B) and of the negative replica of the surface showing the imprinted corrosion products (C). The cellulose acetate sheet was coated with gold during 20 s at 100 mA in a EMITECH K550X sputter coater.

125

Appendix 1

Figure app1.1 Raman spectra obtained for the non-corroded (―) and corroded (―) Au alloy sample.

137

Appendix 2

Figure app2.1 Variable angle sample holder with a gold sample in front of the extracted proton beam for analysis of the same sample at different angles.

142

Figure app2.2 PIXE spectra obtained for the corroded Au alloy using incident angles of 90° and 100°.

143

Figure app2.3 PIXE spectra obtained for the non-corroded and corroded Au alloy (3 MeV proton beam, 90° incident angle), showing that is impossible to clearly identify the presence of S, (Kα1 2.308 keV ; Kα2 2.307 keV; Kβ 2.464 keV ), due to the close energies of the Au M-Lines (Mα1 2.142 keV ; Mα2 2.133 keV; Mβ 2.220 keV; Mƴ 2.404 keV).

143

Figure app2.4 RBS experimental and simulated spectra for the corroded gold alloy (data obtained with a 3.1 MeV proton incident beam).

144

xv

List of tables

Chapter 2 Table 2.1 Characteristics of the techniques for thin film depth profiling analysis [28]. 24 Table 2.2 Energy loss (dE/dx), range (R) and lateral spread (σx) of ions of the beam at the end

of the trajectory for 3 MeV hydrogen and helium ions for Cu, Ag and Au [29]. 25

Table 2.3 Characteristics of the analytical techniques for corrosion layer and substrate characterisation selected for this work.

27

Chapter 3 Table 3.1 Gaseous pollutants concentrations in μg g-1 measured at 55 % RH and 18 °C T

inside a showcase from the Treasure room by using diffusion tubes. 49

Table 3.2 Composition of the base-alloy obtained by XRF for the two Bronze Age fragment torcs and the Iron Age earring.

52

Table 3.3 Composition of the corrosion layer obtained by EDS compared to the Au-base alloy of the fragment torc from Alentejo (MNA-Au283).

54

Table 3.4 Composition normalised to 100 wt% of the corrosion areas and of the Au-based alloy obtained by EDS and XRF.

60

Table 3.5 Egyptian gold objects selected for surface corrosion characterisation with indication of the provenance, date and local of conservation.

64

Table 3.6 Composition of the corrosion layer obtained by EDS and composition of the Au-base alloy obtained by XRF for the Abydos foil (GMA-432-25A) presenting a heterogeneous colour surface.

67

Table 3.7 Composition of the corrosion layer obtained by EDS and composition of the Au-base alloy obtained by μXRF for the Abydos foil (GMA-432-25B) and for the Harageh bead (NMS-1914.1096) presenting a homogeneous colour surface.

69

Chapter 4 Table 4.1 Elemental composition obtained by µXRF, normalised to 100 % of the Ag, Cu and

sterling silver samples used in this work. 79

Table 4.2 Elemental composition obtained by μXRF, normalised to 100 % of the binary (B) and ternary (T) Au alloys fabricated for this work.

80

Table 4.3 Elemental composition by EDS (wt%), normalised to 100 %, of Ag corroded samples.

86

Table 4.4 Complex refractive index (ñ) of the corrosion layer formed on Ag samples and their thicknesses (d).

88

Table 4.5 Elemental composition by EDS (wt%), normalised to 100 %, of Cu corroded samples.

92

Table 4.6 Complex refractive index (ñ) of the corrosion layer formed on Cu samples and their thicknesses (d).

93

Table 4.7 Complex refractive index (ñ) of the corroded sterling silver alloy samples after different immersion times and calculated layer thickness (d).

99

Table 4.8 Elemental composition by EDS (wt%), normalised to 100 %, of the set of artificially corroded sterling silver samples.

101

Table 4.9 Crystalline elements and compounds identified by XRD on the surface of the corroded samples.

103

Table 4.10 Composition of the corroded layer obtained by EDS and composition of the Au-base alloy obtained by µXRF for the Au alloy samples after immersion during 1 and 12 months in 50 mM Na2S solution and for the T6 sample also in immersion during 35h in a 0.1 M Na2S solution.

111

Chapter 5 Table 5.1 Overall evaluation of the selected analytical techniques for the characterisation of

corroded gold alloy surfaces. 129

xvi

Chapter 1

Introduction

Chapter 1 - Introduction

3

1. Introduction

1.1 Gold as a cultural heritage material

Ancient civilisations used gold to manufacture different types of objects. In Europe,

archaeological remains found during the 1970s in Varna, present Bulgaria, indicate that gold

objects have been present in societies since the 5th millennium BC [1,2].

Gold objects are material expressions of past civilisations: gold can be found in jewellery,

ornamental and sacred objects and coins, and can be used as coating of different supports. Due

to its relative scarcity, chemical stability, and physical properties like specular reflection,

responsible for its characteristic lustrous surface [3], gold is connected by different cultures to

power, religion, wealth, etc. Gold artefacts are thus testimonies on a double legacy: a material

and a culture identity.

In Nature gold occurs predominantly in the native state (for example, as “visible” gold in

auriferous quartz and in alluvial deposits) and in primary deposits as a constituent of the ores

[4]. “Free” gold (in powder and nuggets) are gold alloys containing a variable silver content that

might attain high values and a copper content generally under 2 wt%.

In early periods, when metallurgy was not yet well established, different gold alloys could be

obtained by exploitation of distinct gold deposits where silver contents in gold were different.

By addition of silver was then possible to produce a large range of colour nuances. Over time,

refining techniques were developed, and the separation of gold from the other elements

present in native gold was achieved first by cupellation, the separation of gold from other base

metals, and then by parting that separates gold from silver [5,6]. Pure gold could then be alloyed

with silver and copper to obtain alloys with specific properties, like colour and hardness [4,5].

This very ductile “material” was shaped into objects by plastic deformation (hammering,

lamination) or by casting. In the case of composite objects, made up of several parts, these parts

could be mechanically joined (with rivets) or soldered [4]. An object could then be decorated

with motives obtained by chasing, engraving, and repoussé or by adding elements made from

either several different materials (precious stones, enamel, niello, etc.) or from gold alloys

(filigree and granulation).

When finished, the object was put to use and submitted to breakage. At this stage, it could be

repaired for reuse or just be discarded. When reused, its function could be modified. When

discarded, it could be submitted to different environments like burial and underwater contexts.


4

In the case of gold objects, these contexts determine the corrosion processes. When found, the

object is musealised and submitted to a new environment context. In a museological context,

depending on its state of conservation, the object can undergo a conservation process.

Information on the different steps of the construction, use, reuse, and degradation processes of

a gold object is “registered” in the material (figure 1.1). This information can be “read” through

the application of analytical techniques based on physical and chemical phenomena. Because

the large number of techniques available provide different information on the object, the

analytical study of gold objects can be separated into three main areas: 1) raw material

provenance and circulation of materials and artefacts; 2) metallurgy (material transformation)

and objects manufacturing techniques; and 3) conservation [4].

Conservation studies are related to the material preservation and require the full understanding

of the corrosion processes. This work is devoted to the atmospheric corrosion of objects made

from gold alloys in museological context. To approach the atmospheric corrosion processes, it

is mandatory to have a detailed knowledge of the material composition, of the fabrication and

decoration techniques used in the object production, and of the environment conditions in

which the object has been conserved.

Figure 1.1 – Schema of the different stages of the production (from mineral to fabrication techniques), use, discard and musealisation of gold alloy objects.


5

1.1.1 Gold alloys

As above-mentioned, gold objects are in general made from gold (Au) alloys containing silver

(Ag) and copper (Cu). The Au-Ag-Cu alloys properties are defined by the interrelation of these

three elements that can be described by the conjunction of the Au-Ag, Au-Cu and Ag-Cu systems.

In the Au-Ag system, completely miscible alloys are formed even at low temperatures, as it can

be depicted by the Au-Ag phase equilibrium diagram (figure 1.2A). As a consequence of the very

similar lattice constants of the face-centered cubic (FCC) unit cells of Au (a= 4.078 Å) and Ag (a=

4.085 Å) [7,8], a single solid phase (α-phase) solution is formed at all concentrations. The Au-Cu

system also shows a complete series of α-phase solutions for all concentrations, for

temperatures above 410 °C. The liquidus and the solidus occur at a composition of 80.1 wt% Au

at 911 °C, indicating that Au-Cu alloys have a narrow solidification range. For certain composition

ranges and at temperatures below 410 °C, ordered structures like AuCu, AuCu3 and Au3Cu are

formed (figure 1.2B) [9]. These structures are a result of a disorder-to-order transformation, and

their formation implies a severe distortion of the crystal lattice in one crystal lattice direction. A

change from a FCC structure for the disordered phase to a face-centered tetragonal (FCT)

structure occurs. This change increases the hardness of the alloy, but decreases its ductility and

workability [9].

Figure 1.2 – Binary phase equilibrium diagrams of the Au-Ag (A) and Au-Cu systems (B) [10].

The Ag-Cu system is of eutectic type (figure 1.3). The alloys have limited solubility in the solid

state, as a consequence of the different lattice constants of the FCC units (Ag: a= 4.085; Cu: a=

3.614 Å) [8]. The solubility of Cu in Ag increases with temperature until it reaches the limit of 8.8


6

wt% Cu at 779.1 °C. Ag can be dissolved in Cu up to a maximum of 8 wt%. The eutectic point

occurs at a composition of 28.1 wt% Ag at 779.1 °C.

Figure 1.3 – Binary phase equilibrium diagram of the Ag-Cu system [10].

The resulting Au-Ag-Cu ternary phase diagram, presented in figure 1.4A shows a liquidus valley

that extends from the Ag-Cu eutectic point at 779.1 °C to the Au-Cu liquidus minimum at 911 °C

[11]. The Au-Ag-Cu system is also influenced by the solid-state immiscibility field of the Ag-Cu

system as it can be depicted by figure 1.4B which represents Au-Ag-Cu phase diagram of some

isothermal solid state boundaries of the immiscibility field [11]. The composition and the

solidification temperatures influence the alloy microstructures.

Figure 1.4 – Au-Ag-Cu ternary phase equilibrium diagram of some isotherms on the liquidus surface (A) and some isothermal solid state boundaries of the immiscibility field (B). Projections on the room temperature plane [11].


7

1.1.2 The colour of the gold alloys

The colour of gold alloys is an important characteristic for the identification of a corroded gold

object (as discussed in section 1.2). When adding Ag and Cu to Au, the alloy acquires different

colours. As depicted in figure 1.5, by adding Ag, the alloy acquires a green-yellow to white colour.

The addition of Cu changes to red colour.

Figure 1.5 – Relationship between the composition of Au-Ag-Cu alloys and their colour. Adapted from [11].

The colour of gold alloys is the result of their reflectivity as a function of the incident light

frequency, which results directly from its electronic band structure [12]. Absorption processes

in metal result from transitions of electrons from the conduction band to energetically higher

bands, or transitions from lower bands to energy states of the conduction band, which are

situated above the Fermi level [9]. For Cu, interband transitions are possible from the 3d bands

below the Fermi energy to unoccupied levels in the 4s above the Fermi energy. The transition

energy is 2.0 eV, which corresponds to a wavelength of circa 620 nm, fact that explains the Cu

reddish colour. In Au, the interband absorption threshold occurs at a slightly higher energy (2.2

eV) than Cu, corresponding to a selective absorption at 550 nm, explaining the yellow colour.

The interband absorption edge of Ag is ≈ 4 eV. This frequency is in the ultraviolet spectral range,

and the reflectivity remains high throughout the visible spectrum [13].


8

1.2 Atmospheric corrosion of gold alloys

In the 1980s, the corrosion of gold alloys emerged as a research domain in modern jewellery

[14,15] and dentistry, due to the wide application of gold alloys as restoration material [16-18].

More recently, the use of Au and Ag nanoparticles in catalysis, biological sensors and

nanoelectronics domains increased the research on gold and silver corrosion, particularly to

characterise corroded nanosurfaces [19,20] and produce nanoporous metals (by using corrosion

processes, like chemical and electrochemical dealloying [21]). However, only occasional papers

are devoted to case studies of gold alloys corrosion in the field of cultural heritage [22-27].

The corrosion of gold objects is visually characterised by a surface colour change. The surface

acquires a reddish hue referred by several authors as tarnish (figure 1.6). The terminology is

however, inaccurate. The term “tarnish” is used to indicate the alteration of the surface colour,

while the term “corrosion” is used when the formation of corrosion products occurs [28,29]. Yet,

“tarnish” implies a chemical transformation of the surface by a corrosion process.

Figure 1.6 – Archer armband (c.1700 BC) from the National Museum of Archaeology in Lisbon (Portugal). The corrosion layer shows a reddish colour with a multi-hued effect.

Pure gold is chemically stable, but the main alloying elements in gold alloys, Ag and Cu, may

corrode [7,30]. In addition to the influence of relative humidity and temperature, the gaseous

pollutants present in the atmosphere, such as H2S, NOx, SO2 and O3, contribute to the Ag and

Cu atmospheric corrosion development [31-36]. Some authors suggest that gold alloys corrosion

growing is related to the increase of sulphur concentration in the atmosphere due to air

pollution [22]. The corrosion due to chloride was also suggested and approached by several

authors, but the influence of this element on the atmospheric corrosion of gold objects is

insignificant when compared to the influence of sulphur [17,37-39].


9

Corrosion resistance depends on the gold alloy composition and its microstructure [29]. With

the increase of the Ag and Cu contents, the gold alloy corrosion resistance decreases. Contrarily

to gold binary alloys (Au-Ag and Au-Cu) with a homogenous single-phase, the ternary alloys (Au-

Ag-Cu) can have multiple phases that differ with the alloy composition. The presence of a two-

phase microstructure promotes tarnish formation due to microgalvanic coupling [17,28].

Furthermore, the grain size may also affect the corrosion resistance. Smaller grain sizes and

greater homogeneity leads to higher corrosion resistance [28,40]. Nevertheless, Randin et al.

[41] propose that when compared to the Au-Ag and Au-Cu binary alloys, Au-Ag-Cu ternary alloys

are more corrosion resistant. Several authors suggest that above about 50 at% of Au, corrosion

is limited [42]. This value corresponds to a concentration of 75.6 wt% Au in Au-Cu alloys and

64.6 wt% Au in Au-Ag alloys.

The corrosion resistance is also influenced by the technology of production, because the metal

is subjected to thermo-mechanical processes that change its microstructure, such as mechanical

deformations and recrystallization annealing. When gold alloys undergo plastic deformation

during forming processes, their mechanical properties change: for example, due to grains

elongation hardness increases while ductility decreases, but the growing of new grains by

heating the metal increases ductility again and reduces strength. During the cited recovery

process, the residual stresses are reduced, however, as they have not been removed by stress-

relieving annealing, their presence can induce the development of stress corrosion cracking,

which decreases the corrosion resistance [28,40,43,44]. The susceptibility to stress corrosion of

a gold alloy depends on its composition [45]. However, some authors reported that annealed

microstructures with a low level of residual stresses can show high corrosion resistance

independently of the gold alloy composition [40].

Therefore, as above mentioned, the corrosion products formed on the surface of a gold object

are a result of the alloy composition, of the fabrication techniques, and of the type of

atmosphere to which they are exposed.

Although it is recognised the influence of both the gold alloy composition and the fabrication

techniques employed on the corrosion development, the effect of these two factors on the

corrosion mechanism and the identification of the corrosion products are still not clearly

understood or fully described. Also, two important parameters for the corrosion layer

characterisation, its morphology and thickness, have seldom been investigated. The scarce

references on the thickness of the corroded layer (which is said to be thinner for red than for


10

dark red areas) that also contributes to the surface colour, is reported only for high-purity gold

alloys (Au > 99 wt%) and it is estimated between 5 and 350 nm [22,46].

The red to brownish-black coloured corroded surfaces of gold objects have been mainly ascribed

to the presence of acanthite Ag2S [26,47], but other corrosion products such as covellite (CuS),

chalcocite (Cu2S), petrovskaite (AuAgS) and uytenbogaardtite (Ag3AuS2) have also been

identified [48-50]. It is also referred that the corrosion products appear to be similar even when

the base alloy composition is distinct and that in sulphide environments there is a selective

reaction of Cu and Ag [51].

1.3 Conservation of gold alloys objects

The lack of knowledge on the corrosion of gold alloys has a direct impact on the conservation of

gold alloy objects. In the last 15 years, published work on gold alloys corrosion, particularly in

the case of museum objects, has increased [22-27].

Gold alloys corrosion has been reported since the early twentieth century [52], but excluding

the work published by David Scott on corroded objects made from tumbaga (Au-Cu alloys with

high Cu contents) from Central and South America [53], research on corrosion mechanisms is

rare.

Conservation methodologies are, however, inexistent. Some authors refer that pure gold and

high gold content alloys do not require any conservation treatment, because they are resistant

to corrosion. Others affirm that the corrosion products formed on the gold alloys surface are a

result of the corrosion of the alloying elements, Cu and Ag, and suggest that the same cleaning

processes used for those metals should be directly applied to gold alloys [54].

Hence, the removal of corrosion products of gold alloys is based on mechanical, chemical, and

electrochemical methods usually applied for silver alloys cleaning. Other methods like laser

cleaning [55-57], cold plasma [58] and UV/Ozone treatment [59] are under investigation to be

implemented in the case of silver objects. Mechanical methods use abrasives (SiO2, Al2O3, TiO2,

CaCO3, etc.) mixed with organic substances (soaps, fatty acids, etc.) [60,61]. Chemical methods

are widely based on the use of acidified thiourea (CS(NH2)2), which acts as a chelating agent

dissolving Ag2S. Other weak acids, such as formic acid, and chelating agents, like

ethylenediaminetetraacetic acid (EDTA), are also used. Electrochemical methods consider

mainly the reduction of Ag2S and AgCl [62].


11

These methods were recently compared for corroded silver alloys cleaning [60,63]. The

published data indicate that the result of cleaning processes depends on the alloy composition.

None of the tested methods presents optimal results when the visual appearance, base metal

loss, cleaning residues left on the surface and re-corrosion are assessed [60]. This demonstrates

as well that silver alloys conservation treatments are still not fully determined.

For corroded gold objects, this assessment was never carried out, and it is compulsory to

characterise and to be able to identify the corrosion layer. That way, it will be possible to assess

and to establish the adequate conservation actions to be implemented either by direct

(definition of cleaning methods) or by indirect methods (preventive conservation).

1.4 Research objectives and structure of the thesis

The present work considers the atmospheric corrosion of gold alloy objects exhibited or stored

in museums. The main objective is to define and to establish analytical strategies to characterise

thin corrosion layers on gold objects. The analytical strategies developed should then be applied

to the definition of conservation strategies of gold objects collections.

Figure 1.7 outlines the structure of the thesis indicating the main objectives by chapter. The

characterisation of thin corrosion layers on gold alloys requires information on its colour,

morphology, thickness, and on its elemental and structural compositions. A first analytical

approach to the corrosion layer characterisation was based on these five parameters and on the

particular requirements related to gold objects, namely the use of non-invasive techniques and

the transport limitations to external laboratories.

The discussion of the available analytical techniques and their selection is presented in Chapter

2. The third chapter is devoted to three case studies. Corroded gold objects from national and

foreign museums were studied to assess the applicability of the analytical techniques selected

in Chapter 2. The characterisation of the corroded layers considered the elemental composition,

the fabrication techniques, and the exhibition environments.

Chapter 4 presents the detailed characterisation of samples made from binary and ternary alloys

(Ag-Cu, Au-Cu, Au-Ag, and Au-Ag-Cu) corroded in a sulphide environment, the pollutant that

revealed to be responsible for the atmospheric corrosion in museums. The samples

compositions were selected based on the results obtained for the three case studies. This

approach allowed to evaluate the limitations of each technique for the corrosion layer

characterisation and to propose an analytical strategy which is described in Chapter 5.


12

Figure 1.7 – Schema of the thesis structure.


13

1.5 References

[1] T. Higham, J. Chapman, V. Slavchev, B. Gaydarska, N. Honch, Y. Yordanov, B. Dimitrova, New perspectives on the Varna cemetery (Bulgaria) – AMS dates and social implications, Antiquity 81 (2007) 640‒654. [2] T. Dzhanfezova, Are the “new” AMS Varna dates older?, Bulgarian e-Journal of Archaeology 3 (2013) 31‒66. [3] G. Okazawa, K. Koida, H. Komatsu, Categorical properties of the colour term “Gold”, J. Vision 11 (2011) 1‒19. [4] M.F. Guerra, Ouro e tesouros patrimoniais: Compreender e conservar, Gazeta de Física 30 (2007) 7‒11. [5] M.F. Guerra, An overview on the ancient goldsmith’s skill and the circulation of gold in the past: the role of x-ray based techniques, X-ray Spectrom. 37 (2008) 317‒327. [6] P. Craddock, The process of gold refining in King Croesus’ Gold, Excavations at Sardis and the History of Gold refining (A. Ramage, P. Craddock eds.), 2000, British Museum Press, London, pp. 10‒13. [7] R.E. Krupp, T. Weiser, On the stability of gold-silver alloys in the weathering environment, Miner. Deposita 27 (1992) 268‒275. [8] G.V. Raynor, The alloying behaviour of gold, Gold Bull. 9 (1976) 12‒19. [9] M. Grimwade, The metallurgy of gold, Interdisciplinary Science Reviews, 17 (1992) 371‒381. [10] ASM Handbook vol 3: Alloy phase diagrams, (H. Okamoto, M.E. Schlesinger, E.M. Mueller, eds.) ASM International, New York, 1992. [11] A.S. McDonald, G.H. Sistare, The metallurgy of some carat gold jewellery alloys, Gold Bull. 11 (1978) 128‒131. [12] J. Henning, Phase transformations in 18-Carat gold alloys studied by mechanical spectroscopy, PhD Thesis, École Polytechnique Fédérale de Lausanne, Suisse, 2010. [13] M. Fox, Optical properties of solids, Oxford University press, 2001. [14] C. Courty, H.J. Mathieu, D. Landolt, Tarnishing of Au-Ag-Cu alloy studied by Auger electron spectroscopy and coulometry, Mater. Corros. 42 (1991) 288‒295. [15] O. Carvalho, D. Soares, A. Fonseca, F.-S. Silva, Tarnish and corrosion evaluation of a blue-gold based alloy, Mater. Corros. 60 (2009) 355‒359. [16] K.J. Fioravanti, R.M. German, Corrosion and tarnishing characteristics of low content dental casting alloys, Gold Bull. 21 (1988) 99‒110. [17] P.P. Corso, R.M. German, H.D. Simmons, Corrosion evaluation of gold-based dental alloys, J. Dent. Res. 64 (1985) 848‒853. [18] M.S. Chana, A.T. Kuhn, A critique of the Tuccillo-Nielsen wheel method for tarnish testing of dental alloys, J. Dent. 12 (1984) 314‒318. [19] V.J. Keast, T.A. Myles, N. Shahcheraghi, M.B. Cortie, Corrosion process of triangular silver nanoparticles compared to bulk silver, J. Nanopart. Res. 18 (2016) 45‒66. [20] J.L. Elechiguerra, L. Larios-Lopez, C. Liu, D. Garcia-Gutierrez, A. Camacho-Bragado, M.J. Yacaman, Corrosion at the nanoscale: the case of silver nanowires and nanoparticles, Chem. Mater. 17 (2005) 6042‒6052. [21] J. Weissmüller, R.C. Newman, H.J. Jn, A.M. Hodge, J.W. Kysar, Nanoporous metals by alloy corrosion: formation and mechanical properties, MRS Bull. 34 (2009) 577‒586. [22] T. Lu, J. Zhang, Y. Lan, Y. Ma, H. Chen, J. Ke, Z. Wu, M. Tang, Characterization of tarnish spots in Chinese high-purity gold jewelry, Gems & Gemology 51 (2015) 410‒417. [23] V. Corregidor, L.C. Alves, J. Cruz, Analysis of surface stains on modern gold coin, Nucl. Instrum. Meth. B 306 (2013) 232‒235. [24] C. Liang, C. Yang, N. Huang, Investigating the tarnish and corrosion mechanisms of Chinese gold coins, Surf. Interface Anal. 43 (2011) 763‒769.


14

[25] E. Campo, E. Vera, J. Göllner, C.A. Ortiz, R. Lleras, Degradación de piezas arqueológicas “Colección Calima” Museo del Oro, Suplemento de la Revista Latinoamericana de Metalurgia y Materiales S1(2) (2009) 657‒661. [26] D.M. Bastidas, E. Cano, A.G. González, S. Fajardo, R. Lleras-Pérez, R., E. Campo-Montero, F.J. Belzunce-Varela, J.M. Bastidas, An XPS study of tarnishing of a gold mask from a pre-Columbian culture, Corros. Sci. 50 (2008) 1785‒1788. [27] G. Gusmano, R. Montanari, S. Kaciulis, G. Montesperelli, R. Denk, “Gold corrosion”: red stains on a gold Austrian ducat, Appl. Phys. A 79 (2004) 205–211. [28] W.S. Rapson, The metallurgy of the coloured carat gold alloys, Gold Bull. 29 (1996) 61–69. [29] L.W. Laub, J.W. Stanford, Tarnish and corrosion behaviour of dental gold alloys, Gold Bull. 14 (1981) 13–18. [30] T.E. Jones, S. Piccinin, C. Stampfl, Relativity and the nobility of gold, Mater. Chem. Phys. 141 (2013) 14‒17. [31] R. Wiesinger, R. Grayburn, M. Dowsett, P.-J. Sabbe, P. Thompson, A. Adriaens, M. Schreiner, In situ time-lapse synchrotron radiation X-ray diffraction of silver corrosion, J. Anal At. Spectrom. 30 (2015) 694–701. [32] R. Wiesinger, I. Martina, Ch. Kebler, M. Schreiner, Influence of relative humidity and ozone on atmospheric silver corrosion, Corros. Sci. 77 (2013) 69–76. [33] B.V. Salas, M.S. Wiener, R.Z. Koytchev, G.L. Badilla, R.R. Irigoyen, M.C. Beltrán, N.R. Nedev, M.C. Alvarez, N.R. Gonzalez, J.M.B. Rull, Copper corrosion by atmospheric pollutants in the electronics industry, ISRN Corros. (2013) Article ID 846405. [34] R. Wiesinger, M. Schreiner, Ch. Kebler, Investigations of the interactions of CO2, O3 and UV light with silver surfaces by in situ IRRAS/QCM and ex situ TOF-SIMS, Appl. Surf. Sci. 256 (2010) 2735–2741. [35] H. Kim, J.H. Payer, The tarnish process of silver in H2S environments, Corr. Sci. Tech. 6 (2006) 206–212. [36] H. Kim, Corrosion process of silver in environments containing 0.1 ppm H2S and 1.2 ppm NO2, Mater. Corros. 54 (2003) 243–250. [37] D. Upadhyay, M.A. Panchal, R.S. Dubey, V.K. Srivastava, Corrosion of alloys used in dentistry: A review, Mat. Sci. Eng. A – Struct. 432 (2006) 1‒11. [38] C.C. Merriman, D.F. Bahr, M.G. Norton, Environmentally induced failure of gold jewelry alloys, Gold Bull. 38 (2005) 113–119. [39] N.K. Sarkar, R.A. Fuys, J.W. Stanford, The chloride corrosion of low-gold casting alloys, J. Dent. Res. 58 (1979) 568‒575. [40] C. Cason, L. Pezzato, M. Breda, F. Furlan, M. Dabalà, Effect of microstructure and residual stresses, generated from different annealing and deformations processes, on the corrosion and mechanical properties of gold welding alloy wires, Gold Bull. 48 (2015) 135‒145. [41] J.P. Randin, P. Ramoni, J.P. Renaud, Tarnishing of AuAgCu alloys. Effect of the composition, Werkst. Korros. 43 (1992) 115–123. [42] A.J. Forty, Micromorphological studies of the corrosion of gold alloys, Gold Bull. 14 (1981) 25–35. [43] M.G. Alvarez, S.A. Fernández, J.R. Galvele, Effect of temperature on transgranular and intergranular stress corrosion crack velocity of Ag-Au alloys, Corros. Sci. 44 (2002) 2831‒2840. [44] W.S. Rapson, The metallurgy of the coloured carat gold alloy, Gold Bull. 23 (1990) 125‒133. [45] J.M.M. Dugmore, C.D. DesForges, Stress corrosion in gold alloys, Gold Bull. 4 (1979) 140‒144. [46] G. Gusmano, R. Montanari, S. Kaciulis, A. Mezzi, G. Montesperelli, L. Rupprecht, Surface defects on collection coins of precious metals, Surf. Interface Anal. 36 (2004) 921–924. [47] Y. Changjiang, L. Chenghao, W. Peng, Investigation of the tarnish on the surface of a panda gold coin, Rare Metals 26 (2007) 68‒73. [48] M. Griesser, R. Traum, K.E. Mayerhofer, K. Piplits, R. Denk, H. Winter, Brown spot corrosion on historic gold coins and medals, Surface Eng. 21 (2005) 385–392. [49] J.H. Frantz, D. Schorsch, Egyptian red gold, Archeomaterials 4 (1990) 133–152.


15

[50] C. Courty, H.J. Matthieu, D. Landolt, Tarnishing of Au-Ag-Cu alloy studied by Auger electron spectroscopy and coulometry, Werkst. Korros. 42 (1991) 288–295. [51] E. Angelini, E. Cordano, S. Kaciulis, G. Mattogno, L. Pandolfi, M.R. Pinasco, F. Rosalbino, XPS and electrochemical characterization of tarnish films on the dental alloys, Surf. Interface Anal. 30 (2000) 50‒55. [52] A. Lucas, Ancient Egyptian Materials and Industries, Edward Arnold Publishers Ltd., 1926. [53] D. Scott, The deterioration of gold alloys and some aspects of their conservation, Studies in Conservation 28 (1982) 194‒203. [54] E. Angelini, S. Grassini, S. Tusa, Underwater corrosion of metallic heritage artifacts in Corrosion and conservation of cultural heritage metallic artefacts (P.Dillmann, D. Watkinson, E. Angelini, A. Adriaens, eds.) Woodhead Publishing Limited, 2013, pp. 236‒259. [55] J.M. Lee, J. Yu, Y. Koh, Experimental study on the effect of wavelength in the laser cleaning of silver threads, Journal of Cultural Heritage 4 (2003) 157‒161. [56] R. Pini, S. Siano, R. Salimbeni. M. Pasquinucci, M. Miccio, Tests of laser cleaning on archaeological metals artefacts, J. Cult. Herit. 1 (2000) 129‒137. [57] C. Degrigny, E. Tanguy, R. Le Gall, V. Zafiropulos, G. Marakis, Laser cleaning of tarnished silver and copper threads in museum textiles, J. Cult. Herit.4 (2003) 152‒156. [58] E.G. Ioanid, A. Ioanid, D.E. Rusu, F. Doroftei, Surface investigation of some medieval silver coins cleaned in high-frequency cold plasma, Journal of Cultural Heritage 12 (2011) 220‒226. [59] A.-M. Hacke, C. Carr, A. Brown, D. Howell, Investigation into the nature of metal threads in a Renaissance tapestry and the cleaning of tarnished silver by UV/Ozone (UVO) treatment, J. Mater. Sci. 38 (2003) 3307‒3314. [60] T. Palomar, B.R. Barat, E. García, E. Cano, A comparative study of cleaning methods for tarnished silver, J. Cult. Herit. 17 (2016) 20‒26. [61] G. Wharton, S.L. Maish, W.S. Ginell, A comparative study of silver cleaning abrasives, JAIC 29 (1990) 13‒31. [62] C. Degrigny, Use of Electrochemical Techniques for the Conservation of Metal Artefacts: A Review, J. Solid State Electr. 14 (2010) 353–361. [63] P. Storme, O. Schalm, R. Wiesinger, The sulfidation process of sterling silver in different corrosive environments: impact of the process on the surface films formed and consequences for the conservation-restoration community, Herit. Sci. 3 (2015) 25‒40.


16

Chapter 2

Analytical methods and instrumentation

Chapter 2 – Analytical methods and instrumentation

19

2. Analytical methods and instrumentation

2.1 Introduction

2.1.1 Analysis of gold objects

The study of gold objects covers their production (from ore exploitation to fabrication techniques)

and the identification of their function, use-wear and degradation [1]. Information on these

different aspects of the object is contained in its morphology and in the materials used, and it can

be gathered by using techniques based on physicochemical phenomena to be added to

information based on typology, iconography, etc. in order to fully understand the object in a social

and cultural context.

Many science-based techniques are nowadays available to contribute to the study of gold objects.

These techniques provide, in a general way, information on morphology at different scales and

depths and on the chemical composition. The first, included in the techniques of exam, can be

achieved by constructing images with different incident and emitted radiations. The main exam

techniques used for gold objects are optical microscopy, X-ray radiography, and scanning electron

microscopy (SEM) [2].

Analytical techniques provide information on the elemental, structural and isotopic composition

of the materials used. Several analytical methods are available for this, but their selection depends

on the type of information to be searched. For example, in the case of precious metals, isotopic

techniques provide data that can be linked to the material provenance, but usually not to the

object manufacturing. The material or the object provenance and circulation can also be searched

by measuring its elemental composition, but good detection limits are required because low

amounts of characteristic elements should be determined (at µg.g-1 level, in general) [3].

Structural techniques provide useful information on certain manufacture techniques and allow

the characterisation of surface layers that correspond to intentional patinas or natural corrosion.

One of the main requirements when analysing gold objects is the use of non-invasive techniques.

For this reason, elemental techniques are often applied to their study. These techniques are wide-

ranging and can be applied to determine major, minor and trace elements. The most widely used

techniques are X-ray fluorescence (XRF), scanning electron microscopy with an energy dispersive

X-ray system (SEM-EDX) and particle induced X-ray emission (PIXE). X-ray diffraction (XRD) a

structural technique is often added to the list, because non-destructive systems became routinely

portable [4-6]. Non-destructiveness restricts the use, for example, of inductively coupled plasma


20

(ICP) based spectrometry techniques. ICP-MS, ICP-AES and ICP-OES (mass, atomic and optical

emission) are generally applied to particular studies of gold provenance [2,7,8].

Another important requirement when analysing gold objects is related to transport restrictions

due to security reasons and sometimes to the objects fragility. Portable equipment can be used

to overcome this situation. Today, portable equipment is available for some analytical techniques

like XRF, SEM and XRD, among others. However, not all of them present the same level of

portability and the detection limits and spatial resolution that can be reached are far from those

obtained with stationary equipment [9-12].

The definition of an analytical protocol to be applied to the study of one, or a set of gold objects,

should consider the objects characteristics, the issues to be researched, and security and fragility

requirements. Several criteria should be used to select the type of information that should be

search (elemental, structural, isotopic information according to the question) and the more

appropriate analytical techniques for each case. For example, when the elemental composition is

searched, the choice of the analytical techniques is usually based on parameters such as the spatial

resolution (important when studying small details such as solders and decoration elements),

detection limits (for provenance issues), penetration depth (when information on both the

substrate and the surface is necessary).

2.1.2 Analysis of corrosion layers on gold objects

The selection of a conservation methodology should be based on the detailed characterisation of

the objects surface; therefore, an analytical protocol should be defined. The aim of this work is to

characterise the corrosion layer developed on gold objects exhibited or stored in museums,

subjected to atmospheric corrosion. This protocol should be applied to objects that can be, or not,

moved to laboratories to be analysed.

A typical surface of a corroded gold alloy, figure 2.1, consists on a layer in which the corrosion

products formed are located, and a substrate that corresponds to the original gold alloy. The

analytical techniques chosen to integrate the protocol should consider the characterisation of

both the corrosion layer and the substrate. However, it should be underlined that the corrosion

behaviour of gold alloys has not yet been fully understood, which means that the corrosion

mechanisms and the nature of the formed corrosion products are not entirely described.


21

Figure 2.1 – Schematic representation of a corroded gold alloy.

In fact, the scarce published studies on corroded gold alloys in the field of cultural heritage focus

on jewellery, coins and medals, but do not consider the surface structure (corrosion layer and

substrate) [13-15]. In the case of jewellery, it should be cited the work by Bastidas et al. [16], on

the corrosion of a Pre-Columbian gold mask, and the work by Frantz and Schorsch [17], on

Egyptian objects. The former was carried out by X-ray photoelectron spectroscopy (XPS) on a

sample and the latter by optical microscopy (OM), SEM-EDS, infra-red (IR) spectrometry and XRD,

both directly on the object and on samples.

In the case of gold alloys, the few publications providing an identification of the corrosion products

and the characterisation of the corroded surfaces were carried out mainly by OM, SEM-EDS, ICP-

MS, electron microprobe analysis (EPMA), XPS, Auger electron microscopy (AES), secondary ion

mass spectrometry (SIMS) and XRD. With the exception of OM, SEM-EDS and XRD, the other

techniques are not suitable in the present case, because they require a surface ablation (ICP-MS,

XPS, AES, SIMS) and are, therefore, too invasive for the objects.

In this work, the objective is to select analytical techniques that allow to fully characterise the

corrosion layer, and to provide the composition of the substrate without sampling. For this

purpose, and in addition to the elemental composition of the substrate, the colour, morphology,

thickness, and the elemental and structural compositions of the surface layer are assessed.

Considering this, the limits of the analytical techniques currently applied to the study of gold

objects were the first tested. To these techniques were added others that provide information on

the colour of the corroded surface and on the thickness of the corroded layer, parameters that

were not approached by the above-mentioned publications. It should be underlined that this

selection was constrained by the lack of knowledge on the corrosion behaviour of gold alloys.


22

2.2 The analytical approach

The characterisation of the corrosion layer requires information on its morphology, colour,

thickness and on its elemental and structural compositions.

Because of the low resolution of optical microscopy, a few hundred of µm for white light, the

morphology of the corrosion layer was in this work studied by SEM. The use of an accelerated

electron beam, with shorter wavelength than white light, allows lowering down the resolution to

a few ten of nm. In addition, since the samples are metallic, they are imaged without stain-coating.

Field Emission Gun Scanning Electron Microscopy (FEG-SEM) provides better spatial resolution,

higher brightness (108 A cm-2 sr-1 at 20 kV) and lower energy spread (0.3‒1.0 eV) when compared

to a regular SEM with a tungsten filament technology, with a brightness of 105 A cm-2 sr-1 at 20 kV

and with 1‒3.0 eV energy spread [18,19]. The main problems remaining are the size of the sample

chamber that limits the size of the object to be imaged, and the need of vacuum pumping in non-

environmental chambers that may cause damage to fragile objects.

The colour is an important hint to describe the corroded surfaces. Its analysis combined with

morphological data provides detailed information on the corroded layer. In this work, the colour

was investigated by UV-Vis spectrophotometry, widely used for colour analysis, including the

colour of gold alloy [20,21]. This technique is fast, non-invasive and both stationary and portable

equipment versions are available. UV-Vis spectrophotometry is based on the measurement of the

reflected or transmitted UV and visible light by a sample. Depending on the sample characteristics,

several information like reflectance (specular or diffuse), absorbance, transmittance, and

fluorescence can be used to obtain information on its optical properties and chemical

composition. For example, specular reflectance measurement is suitable for mirrored surfaces

while diffuse reflectance is more suitable for non-specular surfaces. Unlike specular reflection,

where the incidence angle is equal to the reflection angle, in the case of diffuse reflection, the

incoming radiation is spread over the half-space above the surface. The measurement of the

diffuse reflection allows obtaining the scattered radiation [22]. In this work, considering the

expected surface roughness of the samples, the diffuse reflectance was used to analyse the colour

of the corroded surfaces.

The colours of corrosion layers on metallic surfaces are dependent on their thickness [23]. For this

reason, the estimation of the corrosion layer thickness was considered in this work. Several

techniques are used to estimate thin films thickness, like among others XPS, AES, SIMS, Rutherford

backscattering spectrometry (RBS) and ellipsometry [24]. XPS, AES and SIMS were not considered

as they are invasive techniques. RBS was tested in this work, as discussed later, once its


23

characteristics are appropriate to determine the composition depth profile of the corrosion layer.

However, only ellipsometry, a non-invasive technique based on the measurement of the change

in polarisation state of a polarised collimated light beam upon reflection by a surface can, in

addition to thickness, determine the optical constants (n, k) of the corroded layers. This allows an

integrated analysis of the colour and the thickness of the corroded surface [25].

For the determination of the elemental and structural composition of the corrosion layer and the

substrate, several methods may be considered. XRD was immediately chosen to identify the

crystalline corrosion products. It is non-invasive, and both portable and stationary equipment are

available. Alike XRD, Raman spectrometry is also non-invasive and can be found as portable

system. This technique has other advantages, such as high surface sensitivity and high lateral

resolution that allow obtaining information on less crystallised phases, difficult to be identified by

XRD [26]. As Raman spectrometry has been used for the identification of the structural

composition of corrosion products of metals, namely silver [26], it was also considered its

application to the characterisation of the corrosion products growing on gold alloys. A first

approach was thus carried out in this work, but the preliminary results, which can be found in

Appendix I, are hardly exploitable. The absence of fingerprints characteristic of the corrosion

products of gold alloys presently disables their identification.

The elemental composition of the first layer and of the substrate may be determined by many

different analytical methods available in portable and stationary configurations. As the nature of

corrosion products and corrosion layer thickness are unknown, it was decided to test several non-

invasive techniques that are regularly applied to the study of gold objects.

The most widely used methods are X-ray based, like ED-XRF, EDS and PIXE. Other ion beam

analysis (IBA) techniques, such as RBS, are also widely used, particularly when studies involve the

characterisation of thin films [27]. In their recent review on thin film analysis, Jeynes and Colaux

[28] pointed out the advantages and disadvantages of several techniques when depth profiling is

required. The main general characteristics of the discussed techniques are given in table 2.1. They

propose a discussion on one important question: “what to do when the routine method fails for

the exceptional problems?” and they suggest the use of IBA techniques.


24

Table 2.1 – Characteristics of the techniques for thin film depth profiling analysis [28].

SIMS*a XTEM*b SAM*c GD-OES*d

XPS*e LA-ICP-MS*f IBA

Primary beams

keV ions ≈ 100 keV electrons

≈ 100 keV electrons

Plasma X-rays Pulsed laser ≈ 3 MeV light ions ≈ 30 MeV heavy ions

Detected signal

Sputtered ions

Primary electrons in phase contrast

Auger electrons

Visible photons

Photo-electrons

Evaporated ions

X-ray; nuclear reaction products: scattered primaries, target recoils and γ-rays

Destructive Yes Yes Yes Yes Yes Yes No Depth resolution

2 nm 0.1 nm 2 nm 20 nm 2 nm 50 nm 2 nm

Information depth

500 nm 100 nm 500 nm 50 μm 500 nm ≈ 5 μm 15 μm

Lateral resolution

50 nm 0.1 nm 2 nm 1 mm 3 μm 10 μm 500 nm

Elemental imaging

Yes EELS*g, EDX*h

Yes No No No No

Molecular information

Yes No? Yes No? Yes No? No

Ambient analysis

No No No No No Yes Yes

Sample preparation

No Yes UHV*i No UHV No No

Quantitative ? No Yes Yes Yes Yes Yes Standards needed

Yes - Yes Yes Yes Yes No

Elemental Sensitivity

10-8 10-1 10-3 10-6 10-3 10-9 10-6

Accuracy ― ― 10 % 10 % 5 % 5 % 1 % Traceability ― ― ― ― Yes Yes Primary

*a Secondary ion mass spectrometry (SIMS); b Cross-sectional transmission electron microscopy (XTEM); c Scanning Auger microscopy (SAM); d Glow discharge optical emission spectroscopy (GD-OES); e X-ray photoelectron spectrometry (XPS); f Laser-ablation coupled plasma mass spectrometry (LA-ICP-MS); g Electron energy loss spectrometry (EELS); h energy-dispersive X-ray analysis; I ultra-high vacuum (UHV).

As referred, the two IBA techniques currently used in the analysis of gold objects are PIXE and RBS

with an external beam facility not to limit the object size and shape. The two IBA techniques that

can in certain facilities be combined (data obtained with a single measurement) probe the near-

surface (from about 3 to 20 µm, with a depth resolution better than 100 nm) of archaeological

materials [29,30].


25

Table 2.2 shows the energy loss, range and lateral spread of ions for Cu, Ag and Au when 3 MeV

alpha and proton beams are used as incident beams. The range of penetration is represented for

Cu in figure 2.2 [29]. The protons have a longer and a wider range distribution compared to the

helium ions, being evident the complementary information obtained by association of PIXE and

RBS [29].

Table 2.2 – Energy loss (dE/dx), range (R) and lateral spread (σx) of ions of the beam at the end of the trajectory for 3 MeV hydrogen and helium ions for Cu, Ag and Au [29].

Material ρ (g cm-3) Protons 3 MeV Helium 3 MeV dE/dx

(keV μm-1) R

(μm) σx

(μm) dE/dx

(keV μm-1) R

(μm) σx

(μm) Cu 8.9 55 35 2.7 490 5.5 0.5 Ag 10.5 53 36 3.7 456 5.4 0.5 Au 19.3 72 27 3.9 560 4.7 0.6

Figure 2.2 – Simulation of trajectories and ranges 3 MeV protons (A) and helium (B) ions in a bulk Cu target. The simulation was obtained by SRIM 2003 software [29].

PIXE has a poor depth profiling capability that can be increased by varying the energy of the

incident beam (absorbers can be used instead), limited by the maximum energy that can be

attained. RBS uses the relation between the energy of the elastically scattered ions and the mass

of the target to determine layer thicknesses and depth profiles, particularly for intermediate and

heavy elements in a light matrix. The choice of the beam particles and energy depends on the

material nature and on the thickness to be determined. The third-generation set-up of the AGLAE

accelerator (C2RMF laboratory) allows the simultaneous use of PIXE and RBS (2 MeV proton-beam


26

and 6 MeV alpha-beam), providing fast measurements on the same spot with good mass and

depth resolution for RBS, even if PIXE analysis remains limited to low- and intermediate-Z

elements [31].

In this work, it was possible to test the possibilities of the combination PIXE-RBS at the AGLAE

facility for the analysis of gold alloy samples artificially corroded. Expecting a corrosion layer

composed of low-and intermediate-Z elements, like oxygen and sulphur, grazing incidence (at

several angles) was used to obtain information on the first corrosion layer. The unsatisfactory

results obtained are presented in appendix 2. The characterisation of the corroded layer was not

fully attained, because elements like sulphur could not be quantified by PIXE. Analytical data

processing difficulties when quantifying light elements in high atomic number matrixes were

reported by Röhrs et al. [32]. The RBS data were also inconclusive. Simulation of the corrosion

layer RBS spectra suggested the presence of a nanometric layer composed of Au, Ag and S, whose

composition could not be estimated, because the error induced by the high surface roughness is

too high. Unreliable results obtained when simulated RBS spectra in the case of porous layers

composed of low-Z elements and interfaces with noticeable roughness, were pointed out by

Darque-Ceretti et al. [30]. Therefore, PIXE and RBS were not selected to characterise the corrosion

layers.

The surface elemental analysis was thus assessed by XRF and by EDS. The advantage of these

techniques is their complementary spatial resolution and depth of analysis. These parameters

depend, however, on the experimental conditions; for example, in the case of XRF are important

the nature of the X-ray tube anode and its operation conditions (filament current, power delivered

to the anode, and anode voltage).

Table 2.3 shows the main characteristics of the analytical techniques selected for this work,

followed by a briefly description of their basic principles and the equipment used.


27

Table 2.3 – Characteristics of the analytical techniques for corrosion layer and substrate characterisation selected for this work.

UV-Vis Ellipsometry SEM EDS XRF XRD Primary beam

UV and visible light

Visible light Electron Electron X-ray X-ray

Operating conditions

DR* 220-1400 nm

Multiangle measurements (60°; 65°; 70°) In situ 70°

5-30 kV 10-25 kV 45 kV/50 μA 50 kV/300 μA

40 kV/ 30 mA

Detected signal

UV and visible light

Light polarisation shift

Electron: SE; BSE

X-ray X-ray X-ray

Information Colour analyses

Optical constants (n, k) Film thickness

Surface morphology

Elemental composition

Elemental composition

Crystalline phases

Information depth

5 - 10 nm 0.1 nm – 1 μm ― ≈ 0.5 μm 9 -60 μm ** 0.1-10 μm ***

Lateral resolution

N/A 1 mm 1 -5 nm 1-10 nm 25 μm 0.5 cm

30 μm

* Diffuse reflectance (DR);** Troalen et al. [33]; *** Kerber et al. [34]

2.2.1 Ultraviolet-visible spectrophotometry

Since the 1970s that UV-Vis spectrophotometry is applied to studies in the field of cultural

heritage. This non-invasive technique has been mainly used to monitor colour changes and to

identify pigments on different painted surfaces [35]. Very few references to studies of ancient

metals, including corrosion, are however available [36,37]. The optical properties of the corroded

layer are highly dependent on the nature of the corrosion products and on the surface

morphology, namely particles size and shape. As UV-Vis is sensitive to these features, it was used

in this work to characterise the corroded surfaces.

UV-Vis spectrophotometry is based on the measurement of the reflected or transmitted incident

UV and visible light by a surface. According to the characteristics of the samples (e.g. specular or

non-specular surfaces), different spectrophotometric parameters can be considered. As above

mentioned, measurement of diffuse reflection is more adequate when rough surfaces are under

study and, for this reason, chosen to be used in this work for colour analysis of the corroded

samples of Ag, Cu, sterling silver and Au alloys.


28

Diffuse reflectance is measured with an integrating sphere, a hollow optical device made from a

diffuse white reflective material or coated with that material, that collects all the reflected

radiation by a surface [22]. Several models can be used to describe diffuse reflectance by non-

specular surface, the most widely used was proposed by Kubelka and Munk, and it is based on the

light propagation in dull colored layer that is parallel to a plane substrate. The model assumes the

sample as an isotropic, uniform, non-fluorescence material. The sample is modelled as a plane

layer with semi-finite thickness but infinite width and length. The incident or reflected radiation

intensities are assumed to be perfectly diffuse and are assigned only to two directions: upwards

and downwards. The proposed model relates the reflectance of a sample, Rꚙ, with the absorption,

α, and scattering, s, coefficients, which are assumed to be independent of the layer thickness.

Hence, the diffuse reflectance, Rꚙ, is given as:

𝑅𝑅∞ = 1 + 𝛼𝛼𝑠𝑠− �𝛼𝛼

𝑠𝑠�2 + 𝛼𝛼

𝑠𝑠� (2.1)

This equation is solved for α/s yielding the known Kubelka-Munk transform function, that allows

to calculate the absorption from the diffuse reflectance data:

𝛼𝛼𝑠𝑠

= (1−𝑅𝑅∞)2

2𝑅𝑅∞ (2.2)

In this work, the colour of corroded samples was analysed by diffuse reflectance and the

absorption calculated by using the Kubellka-Munk function.

Equipment: The diffuse reflectance spectra (DRS) were obtained using a UV-2450 PC

spectrophotometer with an ISR-2600 integrating sphere, from Shimadzu, available at the

Interfacial Electrochemistry Group (IEG) (Centro de Química e Bioquímica), University of Lisbon. A

deuterium (for the UV spectral range) and a halogen (for visible and near infra-red spectral range)

lamps were used as light sources. The DRS were recorded in the 220−1400 nm wavelength range

with a 2 nm width slit at medium scan speed, using BaSO4 as reference. Ag, Cu, and sterling silver

were used as baseline references.


29

2.2.2 Ellipsometry

Ellipsometry is an optical technique used to determine the optical properties (refraction index, n

and extinction coefficient, k) of different materials and the thickness of thin films. This technique

was first used in the 1880s, and in 1933 appeared the first references to the use of ellipsometry

for corrosion studies estimating the corrosion film thickness developed on iron and nickel in acid

and basic solutions [39,40]. Data on the optical properties of the films may be used to identify

their composition and, in some cases, to infer their stoichiometry. The high sensitivity of

ellipsometry allows to estimate the film thickness from a few angstroms to several micrometers

(depending on the k value) and to identify different layers with distinct optical properties [39-41].

There are few references to the application of ellipsometry in the field of cultural heritage. The

very few studies available are devoted to studies of varnishes ageing [42,43] and to the

assessment of museum exhibition conditions by placing Ag coupons in different locations of the

museum. The corroded layer resulting from the atmospheric corrosion was characterised by

ellipsometry [44].

In this work, ellipsometry was used to determine the optical constants and the corrosion layer

thickness of corroded samples of Ag, Cu, sterling silver and gold alloys. In-situ ellipsometry was

used for the corrosion study of a gold alloy sample. The corrosion process was followed over time,

it was determined the optical parameters and the layer thickness were calculated at different

corrosion stages.

Ellipsometry is based on the measurement of the change in polarisation state of a polarised

collimated light beam, due its interaction with the surface [45]. Upon reflection, it occurs a change

of both phase and amplitude of the orthogonal components of the electric (or magnetic) field of

the light. The azimuthal angle amplitude (Ψ) and phase shift (Δ) parameters are defined by the

fundamental equation of ellipsometry:

tanΨ . 𝑒𝑒𝑖𝑖Δ = 𝑅𝑅𝑝𝑝𝑅𝑅𝑠𝑠

(2.3)

where Rp and Rs correspond to the Fresnel reflection coefficients of the sample, for parallel (p)

and perpendicular (s) light components, respectively, and i is the unit imaginary number.

The Fresnel reflection coefficients depend on the optical properties of the sample, in particularly,

on its refractive index and hence Ψ and Δ are a function of the refractive index of the system


30

components, of the radiation wavelength and of the incident angle. When a medium is optical

absorbent, the Fresnel coefficients are complex quantities and hence also will be the index of

refraction (ñ = n – ik, where n is the real part of the refractive index, i = √-1 and k the extinction

coefficient). Considering a given architecture of the probed system, the ellipsometric parameters

Ψ and ∆ can be then mathematically correlated with the optical properties of the system

components and with the thickness of the phases where light refraction occurs. As referred,

ellipsometry can also be used in-situ, for example, for corrosion studies. In this case three different

mediums (corrosion solution, corrosion layer and substrate) must be considered. Ψ and ∆ are a

function of the refractive index of the corrosion solution, of the substrate complex refractive

index, of thickness and complex refractive index of the corrosion layer on the incident angle of the

light. The Fresnel reflection coefficients will result from the reflections and the transmission in and

between the interfaces corrosion solution/corrosion layer and corrosion layer/substrate [45].

Equipment: The ellipsometric measurements were performed using a Sentech SE400

Ellipsometer, from the IEG (Centro de Química e Bioquímica), University of Lisbon, operating in

PSA or PCSA mode (polariser P, compensator C, sample S and rotating analyser A) fitted with a He-

Ne laser (λ = 632.8 nm). For the ex-situ analysis of Ag, Cu and sterling silver samples, Ψ and ∆ were

acquired wherever possible at different angles of incidence (φ = 60, 65, 70°). The thickness of the

corrosion layers was calculated assuming a two-layer model. For the Cu samples the corroded

layer composition was estimated by using the Bruggeman Effective Medium Approximation (BEM)

[46]. In-situ analysis of a gold alloy sample was carried out at an incident angle of 70° and using a

specific cell with two transparent side windows placed perpendicularly to the light axis, allowing

the passage of the beam and then its reflection (figure 2.3). The acquisition and the signal

treatment were carried out using the instrument software. The thickness of the corrosion layers

and optical constants were estimated using a three-layer model.


31

Figure 2.3 – In-situ ellipsometric measurements during the corrosion of a gold alloy sample (A) and the electrochemical cell used for the in-situ measurements (B) [47].

2.2.3 X-ray fluorescence spectrometry

XRF has been largely applied in the field of cultural heritage since 1950, when E.T. Hall used this

technique for the analysis of archaeological objects. Since then, this technique has evolved due to

new equipment development [3,9]. As mentioned, in the field of cultural heritage XRF has been

applied to the study of gold objects, along with other analytical techniques for the analysis of the

material composition and for the characterisation of certain fabrication techniques [2]. In this

work, XRF was used to identify the composition of the gold alloys and for the elemental

characterisation of the corrosion layer.

XRF is based on the interaction of a photon beam, usually X-rays, with a material. The incident

radiation produces ionisation of the atom in its inner-shells by photoelectrical effect. In addition

to the photoelectric effect, the interaction of a photon beam with a material, particularly when

constituted by low atomic number elements, also produces the emission of Auger electrons

(Auger effect) from a less tightly bound state [48,49].

The ionised atom returns to the ground state through transitions of electrons from outer- to inner-

shells (electrons rearrangement). The differences in binding energies of the shells involving

electrons transfers (excess of energy) are emitted, frequently as X-rays. Because each chemical

element has particular binding energies levels, the released set of X-rays is characteristic of the

element. The electron transitions result in the emission of X-rays whose energies correspond to


32

the energies differences between those shells. The X-rays emitted are characteristic of each

chemical element of the material, allowing their identification.

When an X-ray beam passes through matter, its intensity is reduced due to scattering or to

absorption in the material. The attenuation of the beam intensity (I0) passing through a material

of thickness d, and density ρ follows the Lambert-Beer law:

𝐼𝐼 = 𝐼𝐼0𝑒𝑒−𝜇𝜇𝜇𝜇𝜇𝜇 (2.4)

where µ is the mass attenuation coefficient (cm2 g-1). If the material is a complex matrix M,

consisting of a mixture of several chemical elements, µ(Μ) is given by

𝜇𝜇 = ∑ 𝑊𝑊𝑖𝑖𝜇𝜇𝑖𝑖𝑛𝑛𝑖𝑖=1 (2.5)

where µ I is the mass attenuation coefficient of the ith element and wi its mass fraction in the

considered sample. µ has an important role in the quantitative analyses. Both the exciting primary

radiation and the fluorescence radiation are attenuated in the sample. To relate the observed

fluorescence intensity to the concentration of the elements in the matrix, this attenuation must

be taken in consideration [49]. The absorption of radiation in the matter is the cumulative effect

of several types of photon-matter interaction processes that occur in parallel. Accordingly, in X-

ray range the mass attenuation coefficient µ I of the element i is given by:

𝜇𝜇 = 𝜏𝜏𝑖𝑖 + 𝜎𝜎𝑖𝑖 (2.6)

where τ I is the cross-section for the photoelectric ionization and σ I the cross-section for scattering

interactions.

Equipment: In this work, a portable and a stationary XRF equipment from the LIBPhys-

Universidade Nova de Lisboa were used. The portable equipment was used for in-situ analysis

(figure 2.4) when the objects could not be moved from the museums. The stationary µXRF was

used to analyse some selected objects that could be moved to the laboratory and the corroded


33

samples. This equipment has higher spatial resolution than the portable one, allowing the analysis

of small areas, fundamental for the characterisation of the different colours of the corroded

surfaces. In addition, its sensitivity to detect low-Z elements, important feature for the

characterisation of corroded layers, can be increased by using the vacuum option. The accuracy

of the obtained quantitative results in both systems was validated by the analysis of home-made

gold standards.

The portable equipment comprises an X-ray generator ECLIPSE IV from Oxford Instruments (45

kV, 50 µA, 2.25 W max) with a Rh anode and a Be window 250 μm thick and an Amptek XR-100

silicon drift detector (SDD) thermoelectrically cooled detector with a 25 mm2 detection area

collimated down to 17 mm2, 500 µm thick, and a 12.5 µm Be window. The energy resolution is

140 eV at Mn Kα and the acquisition system is an Amptek PX5. For collimating the beam, an acrylic

support with a 2 mm pinhole in Ta was used and a spot size of 0.5 cm on the sample is obtained.

The system components are placed in an aluminium structure in 90° geometry. The elemental

quantification was carried out with the WinAxil software v.4.5.3 that uses fundamental

parameters.

The stationary μXRF is a M4 Tornado equipment from Bruker (50 kV, 300 µA), comprising a Rh X-

ray source with a poly-capillary lens offering a spot size down to 25 µm at a working distance of

10 mm, coupled to a XFlash ® SDD detector, with a 30 mm2 sensitive area and energy resolution

of 142 eV at Mn Kα. The elemental quantification was performed by using ESPIRIT standardless

quantification software, the ZAF matrix correction model, based on analytical expressions for

atomic number (Z), X-ray yield, self-absorption (A) and secondary fluorescence enhancement (F).

Figure 2.4 – Portable XRF during the analysis of a gold necklace (NMA-Au47) from the collections of the National Museum of Archaeology.


34

2.2.4 Scanning electron microscopy with energy dispersive X-ray spectrometry

Scanning electron microscopy (SEM) is based on the interaction of electrons with matter, which

produces the emission of secondary electrons (SE) and backscattered electrons (BSE) that can be

used to obtain high-resolution images of the samples surface. The crystalline structure of the

sample can be studied when diffracted backscattered electrons are collected (EDSB) and its

chemical composition determined by energy dispersive X-ray spectrometry (EDS) when X-ray are

collected. This versatile technique provides information on the samples morphology, topography,

and contrast in composition and in phase.

SEM has been largely applied to the study of goldwork when details related to the objects

fabrication techniques and wear marks are searched as well as to assess corroded surfaces

[33,50,51]. The morphology and the chemical composition of an object can be determined by EDS

enhances the analytical potentialities by allowing a non-invasive elemental microanalysis linked

to a microscopic image of the surface sample or object [51]. SEM-EDS was used in this work for

the characterisation of the corroded samples. A few corroded gold objects could also be moved

to the SEM laboratory for analysis.

As referred, SEM-EDS is based on the interaction of a primary electron beam accelerated to a high-

voltage (normally between 1‒30 kV) with the atoms of the sample. An electromagnetic lenses

system focuses the beam on the sample. The interaction results in the emission of SE, BSE and

EDSB.

SE and BSE are used to acquire images. SE are ejected from the exteriors layers of the atoms as a

result of inelastic collisions with the incident beam. SE have low energy and those collected are

produced close to the sample surface, providing a strong topographic contrast. BSE are primary

electrons that undergo elastic collisions with the atoms nuclei of the sample. These collisions (one

or multiple) result in inelastic processes with a minor energy loss. The original direction of the

electron beam is lost and the electrons are randomly scattered through the material. The signal

intensity increases with the increase of the atomic number, providing composition contrasts of

heterogeneous surfaces [52]. The characteristic emitted X-rays produced by non-elastic collisions

are used for EDS analysis, accordingly to the principles of X-ray spectrometry mentioned in 2.2.3.

Their detection identifies the chemical elements of the sample.

Equipment: The corroded objects and the corroded samples were carried out at the Laboratório

Nacional de Engenharia e Geologia (LNEG) with a SEM-EDS Philips XL 30 FEG (field emission gun)

with a Schottky emitter cathode (this emitter combines the high brightness and low energy spread


35

of the emitted electrons and the high stability of the emission current) (figure 2.5). The FEG is

capable of attaining resolutions of 2 nm at 30 kV and of 5 nm at 1 kV. The equipment was operated

with acceleration voltage from 10 to 15 kV. Elemental composition was obtained with an EDS

(EDAX) system equipped with a Si(Li) detector with a 10 mm2 detection and a 3 μm super ultra-

thin window (SUTW), allowing detection of light elements. The energy resolution is 135 eV at Mn

Kα.

An acceleration voltage ranging from 15 to 30 kV was applied accordingly to the elements to be

analysed (O, S, Cu, Ag, Au). The X-ray spectra were collected in spot mode analysis with 200 s and

300 s acquisition time. The quantitative results were obtained by EDAX software, using the ZAF

matrix correction model and confirmed using a range of home-made gold standards certified by

other techniques.

Figure 2.5 – SEM-EDS equipment (A) and detail of the SEM chamber prior to the exam and analysis of brooch Ophelia (CGM-1138) from the René Lalique collection of the Calouste Gulbenkian Museum (B).

2.2.5 X-ray diffraction

X-ray diffraction (XRD) is based on the constructive interference of an incident monochromatic X-

ray beam and a crystalline material, which acts as a three-dimensional diffraction grating for X-ray

wavelengths similar to the spacing of planes in a crystal lattice (1-100 Å). This technique is used

for the identification and characterisation of crystalline materials based on their diffraction

patterns [53].

XRD has been largely applied in the field of cultural heritage to identify crystalline phases present

in corrosion layers [54]. The use of XRD for corrosion studies presents the advantage of analysing


36

a mixture of crystalline phases, and hence it can identify different corrosion products when the

corroded layers are composed of distinct compounds [55]. In this work, the crystalline corrosion

products were identified by this technique. Analysis were carried out on objects that could be

moved to the XRD laboratory and on corroded samples.

As referred, the three-dimensional structure of crystalline materials is defined by regular,

repeating planes of atoms that form a crystal lattice. When a monochromatic X-ray beam interacts

with these crystallographic planes, part of the beam is diffracted, and the distances between the

planes can be measured by applying Bragg's Law:

𝑛𝑛𝑛𝑛 = 2𝑑𝑑 sin𝜃𝜃𝑛𝑛 (2.7)

in which n is the order of reflection, λ the wavelength of the incident monochromatic radiation, d

the interplanar spacing between the crystallographic planes ([hlk] Miller indices) and θn the angle

of reflection with those planes.

The most usual equipment configuration is the θ:2θ. In this geometry, the tube is fixed and the

sample or object rotates at θ°/min and the detector rotates at 2θ °/min. The intensity of the

reflected radiation is measured throughout the 2θ scan and the result is typically presented in a

diffractogram with intensity as a function of 2θ . From the diffractogram, data can be obtained on

position, intensity, width and half-height of the peaks parameters related with the composition

and structure of the analysed material. The comparison of the d-interplanar set with standard

reference patterns allows for identification of the material. The reference data is available as

Power Diffraction File (PDF) supplied by the International Centre for Diffraction Data [55].

Equipment: Three different equipment were used for this work. For the study of corroded objects

and for the corroded sterling silver samples it was used two Bruker-AXS D8 Discover

diffractometers in the θ:2θ configuration from the Institute for Plasmas and Nuclear Fusion

(Instituto Superior Técnico) and from the Laboratório José de Figueiredo (Lisbon) (figure 2.6) and

from the. Analysis were carried out with a radiation Cu Kα1 (λ= 0.15405 nm) working at 40 kV/40

mA. The beam was collimated with a Göbel mirror and an asymmetric two-bounce Ge (2 2 0)

monochromator. The secondary beam passes through an anti-scattering (0.2 mm) slit and is


37

detected with a scintillation detector. Data was collected with a 0.02° step size and an acquisition

time of 600 s deg-1.

For the other Cu, Ag and gold alloys corroded samples study, it was used a Philips X-ray

diffractometer (PW 1730) from Centro de Química e Bioquímica, University of Lisbon, with

automatic data acquisition (APD Philips v3.6B) with a Cu Kα1 radiation working at 40 kV/30 mA.

Data were collected with a 0.02° step size and an acquisition time of 200 s deg-1.

The aim was to analyse the corrosion products and so a selection of different ranges of 2θ was

made to avoid the main diffraction peaks from the alloys. X-ray diffraction pattern files were

obtained from the Pearson's Crystal Database (PCD).

Figure 2.6 – Analysis of a torc fragment from Serpa (NMA-Au293) of the National Museum of Archaeology with the D8 Discover diffractometer from the Laboratório José de Figueiredo, Lisbon.


38

2.3 References

[1] M.F. Guerra, An overview on the ancient goldsmith’s skill and the circulation of gold in the past: the role of x-ray based techniques, X-ray Spectrom. 37 (2008) 317‒327. [2] M.F. Guerra, T. Calligaro, Gold cultural heritage objects: a review of studies of provenance and manufacturing technologies, Meas. Sci. Technol. (2003) 1527‒1537. [3] M.F. Guerra Role of radiation physics in the study and authentication of ancient gold work, Radiat. Phys. Chem. 95 (2014) 356‒361. [4] A. Buccolieri, A. Castellano, E. Degl’Innocenti, R. Cesareo, R. Casciaro, G. Buccolieri, EDXRF analysis of gold jewelry from the Archaeological Museum of Taranto, Italy, X-Ray Spectrom. (2017) version of record online. [5] G. Demortier, Analysis of gold jewellery artifacts, Gold Bull. 17 (1984) 27‒38. [6] I. Nakai, Y. Abe, Portable X-ray powder diffractometer for the analysis of art and archaeological materials, Appl. Phys. A 106 (2012) 279‒293. [7] V. Leusch, M. Brauns, E. Pernicka, Precise and accurate analysis of gold alloys: Varna, the earliest gold of mankind – a case study in Recent Advances in Laser Ablation ICP-MS for Archaeology (L. Dussubieux, M. Golitko, B. Gratuze, eds.) Springer-Verlag, Berlin Heidelberg, 2016, pp. 95‒114. [8] R.J. Chapman, R.C. Leake, R.A. warner, M.C. Cahill, N.R. Moles, C.A. Shell, J.J. Taylor, Microchemical characterisation of natural gold and artefact as a tool for provenancing prehistoric gold artefacts: a case study in Ireland, Appl. Geochem. 21 (2006) 904‒918. [9] M. Madariaga, Analytical chemistry in the field of cultural heritage, Anal. Methods 7 (2015) 4848‒4876. [10] M. Guerra, M. Manso, S. Longelin, S. Pessanha, M.L. Carvalho, Performance of three different Si X-ray detectors for portable spectrometers in cultural heritage applications, J. Inst. 7 (2012) C10004. [11] S. Pessanha, A. Guilherme, M.L. Carvalho, Comparison of matrix effects on portable and spectrometers for cultural heritage samples, Appl. Phys. A- Matter 97 (2009) 497‒505. [12] M. Uda, A. Ishizaki, R. Satoh, K. Ohada, Y. Nakajima, D. Yamashita, K. Ohashi, Y. Sakuraba, A. Shimono, D. Kojima, Portable X-ray diffractometer equipped with XRF for archaeometry, Nucl. Instrum. Meth. B 239 (2005) 77‒84. [13] C. Liang, C. Yang, N. Huang, Investigating the tarnish and corrosion mechanisms of Chinese gold coins, Surf. Interface Anal. 43 (2011) 763‒769. [14] M. Griesser, R. Traum, K.E. Mayerhofer, K. Piplits, R. Denk, H. Winter, Brown spot corrosion on historic gold coins and medals, Surface Eng. 21 (2005) 385-392. [15] G. Gusmano, R. Montanari, S. Kaciulis, G. Montesperelli, R. Denk, “Gold corrosion”: red stains on a gold Austrian Ducat, Appl. Phys. A 79 (2004) 205‒211. [16] D.M. Bastidas, E. Cano, A.G. González, S. Fajardo, R. Lleras-Pérez, E. Campo-Montero, F.J. Belzunce-Varela, J.M. Bastidas, An XPS study of tarnishing of a gold mask from a pre-Columbian culture, Corros. Sci. 50 (2008) 1785‒1788. [17] J.H. Frantz, D. Schorsch, Egyptian red gold, Archeomaterials 4 (1990) 133‒152. [18] A. Bogner, P.-H. Jouneau, G. Thollet, D. Basset, C. Gauthier, A history of scanning electron microscopy developments: Towards “wet-STEM” imaging, Micron 38 (2007) 390‒401. [19] M. Troyon, Electron sources, Ultramicroscopy 28 (1989) 207‒208. [20] M.E. Nadal, D. Wyble, C.J. Zarobila, Color and appereance in Spectrophotometry: Accurate measurements of optical properties of materials (T.A. Germer, J.C. Zwinkels, B.K. Tsai eds.) Elsevier, Amsterdam (2014), pp. 367‒404. [21] R.M. German, M.M. Guzowski, D.C. Wright, The colour of gold-silver-copper alloys, Gold Bull. 13 (1980) 113‒116. [22] A. Höpe, Diffuse reflectance and transmittance in Spectrophotometry: Accurate measurements of optical properties of materials (T.A. Germer, J.C. Zwinkels, B.K. Tsai eds.) Elsevier, Amsterdam (2014), pp. 180‒216.


39

[23] V. Goossens, J. Wielant, S. van Gils, R. Finsy, H. Terryn, Optical properties of thin iron oxide films on steel, Surf. Interface Anal. 38 (2006) 489‒493. [24] P.J.D. Whiteside, J.A. Chininis, H.K. Hunt, Techniques and challenges for characterizing metal thin films with applications in photonics, Coatings 6 (2016) C6030035 (26 pages). [25] D. Tahir, S. Tougaard, Electronic and optical properties of Cu, CuO and Cu2O studied by electron spectroscopy, J. Phys. Condens. Matter 24 (2012) 175002 (8 pages). [26] I. Martina, R. Wiesnger, M. Schreiner, Micro-Raman investigations of early stage silver corrosion products occurring in sulfur containing atmospheres, J. Raman Spectrosc. 44 (2013) 770‒775. [27] L. Beck, L. Pichon, B. Moignard, T. Guillou, P. Walter, IBA techniques: Examples of useful combinations for the characterisation of cultural heritage materials, Nucl. Instrum. Meth. B 269 (2011) 2999‒3005. [28] C. Jeynes, J.L. Colaux, Thin film depth profiling by ion beam analysis, Analyst 141 (2016) 5944‒5985. [29] T. Calligaro, J.-C- Dran, J. Salomon, Ion beam microanalysis in Non-destructive microanalysis of Cultural Heritage materials vol XLII (K. Janssens, R. Van Grieken eds.) Elsevier, Netherlands (2004) pp. 227‒276. [30] E. Darque-Ceretti, D. Hélary, M. Aucouturier, An investigation of gold/ceramic and gold/glass interfaces, Gold Bull. 35 (2002) 118‒129. [31] J. Salomon, J.-C. Dran, T. Guillou, B. Moignard, L. Pichon, P. Walter, F. Mathis, Ion-beam analysis for cultural heritage on the AGLAE facility: impact of PIXE/RBS combination, Appl. Phys. A 92 (2008) 43‒50. [32] S. Röhrs, T. Calligaro, F. Mathis, I. Ortega-Feliu, J. Salomon, P. Walter, Exploring advantages of 4He-PIXE analysis for layered objects in cultural heritage, Nucl. Instrum. Meth. B 249 (2006) 604‒607. [33] L. Troalen, J. Tate, M.F. Guerra, Goldwork in Ancient Egypt: workshop practices at Querneh in the 2nd Intermediate Period, J. Archaeol. Sci. 50 (2014) 219‒226. [34] S.J. Kerber, T.L. Barr, G.P. Mann, W.A. Brantley, E. Papazoglou, J.C. Mitchell, The complementary nature of X-ray photoelectron spectroscopy and angle-resolved X-ray diffraction Part I: Background and theory, J. Mater. Eng. Perform. 7 (1998) 329‒333. [35] P. Ricciardi, UV-visible-NIR reflectance spectrophotometry in cultural heritage: background paper, Anal. Methods 8 (2016) 5894‒5896. [36] P. Storme, O. Schalm, R. Wiesinger, The sulfidation process of sterling in different corrosive environments: impact of the process on the surface films formed and consequences for the conservation-restoration community, Herit. Sci. 3 (2015) 25‒40. [37] I. Farbman, O. Levi, S. Efrima, Optical response of concentrated colloids of coinage metals in the ultraviolet, visible and infrared regions, J. Chem. Phys. 96 (1992) 6477‒6485. [38] J.A.N.T. Soares, Introduction to optical characterization of metals in Practical Materials characterization (M. Sardela ed.) Springer, Verlag Berlin Heidelberg (2014), pp. 43‒92. [39] P. Hayfield, Ellipsometry as an aid in studying metallic corrosion problems, Surf. Sci. 56 (1976) 488‒507. [40] L. Tronstad, The investigation of thin surface films on metals by means of reflected polarized light, Trans. Faraday Soc. 29 (1933) 502‒514. [41] J. Kruger, Use of ellipsometry in the study of corrosion, Corrosion 22 (1966) 88‒97. [42] K. Polikreti, C. Christofides, Spectroscopic ellipsometry as a tool for the optical characterization and ageing studies of varnishes used in Post-Byzantine icon reconstructions, J. Cult. Herit. 7 (2006) 30‒36. [43] C. Christofides, B. Castellon, A. Othonos, K. Polikreti, C. de Deyne, Fine art painting characterization by spectroscopic ellipsometry preliminary measurements on varnish layers, Thin Solid Films 207‒212 (2004) 455‒456. [44] H. Águas, R.J.C. Silva, M. Viegas, L. Pereira, E. Fortunato, R. Martins, Study of environmental degradation of silver surface, Phys. Stat. Sol. C 5 (2008) 1215‒1218. [45] R.M.A. Azzam, N.M. Bashara, Ellipsometry and polarized light, Elsevier, Amsterdam, 1987. [46] S. Berthier, Optique des milieux composites, Polytechnica, Paris, 1993. [47] R.S. Sampaio, Redução do oxigénio molecular através de eléctrodos modificados por moléculas bio-inspiradas. Master thesis (Mestrado Integrado em Engenharia da Energia), (2016) Faculty of Sciences, University of Lisbon.


40

[48] R. Jenkins, X-Ray fluorescence spectrometry (J.D. Winefordner ed.) John Wiley & Sons, Inc, Toronto (1999). [49] R. Jenkins, R.W. Gould, R.W. Gedck, Quantitative X-ray spectrometry, Marcel Dekker, New York & Basel (1981). [50] M.F. Guerra, G. Demortier, M.L. Vitobello, S. Bobomulloev, D. Bagault, T. Borel I. Mirsaidov, Analytical study of the manufacturing techniques of Kushan gold jewellery, ArcheoSciences 33 (2009) 177‒186. [51] A. Adriaens, M.G. Dowsett, Electron microscopy and its role in cultural heritage studies, in Non-destructive microanalysis of Cultural Heritage materials vol XLII (K. Janssens, R. Van Grieken eds.) Elsevier, Netherlands (2004) pp. 73‒128. [52] J.I. Goldstein, D.E. Newbury, P. Echlin, D.C. Joy, C.E. Fiori, E. Lifshin, Advanced scanning electron microscopy and X-ray microanalysis, 2nd edition, Plenum Publishing, New York, 1987. [53] A. Chauhan, P. Chauhan, Power XRD technique and its application in science and technology, J. Anal. Bioanal. Tech. 5 (2014) 212‒216. [54] D. Neff, S. Reguer, P. Dillmann, Analytical techniques for the study of corrosion of metallic heritage artefacts: from micrometer to nanometer scales in Corrosion and conservation of cultural heritage metallic artefacts (P. Dillmann, D. Watkinson, E. Angelini, A. Adriaens eds.) Woodhead Publishing, Cambridge 2013. [55] Y. Waseda, E. Matsubara, K. Shinoda, X-ray diffraction crystallography, Springer Verlag Berlin Heidelberg 2011.

Chapter 3

Atmospheric corrosion of gold alloy objects:

case studies

Chapter 3 – Atmospheric corrosion of gold alloy objects: case studies

43

3. Atmospheric corrosion of gold alloy objects: case studies

3.1 Introduction

This chapter focuses on three case studies, selected to evaluate the applicability of the analytical

techniques chosen in Chapter 2 for the study of corroded gold objects exhibited and stored in

museums. The objects are subjected to atmospheric corrosion, and their corrosion mechanisms

should be characterised based and related to three main issues: 1) assessment of the

environmental conditions; 2) identification of the alloy composition and the fabrication

techniques employed and 3) identification of the corrosion products.

Different parameters can influence the atmospheric corrosion of those objects [1,2]. The

museums location (climate characteristics) and the nature of the materials entering in the

composition of the exhibition and storage rooms and of the showcases, namely the type of wood,

fibres, etc. have a major impact on atmospheric corrosion processes [3,4].

In the case of gold objects, the alloy composition and the thermo-mechanical techniques used by

the goldsmith, when shaping, decorating and finishing the object, also influence the atmospheric

corrosion development [5,6]. It should also be reminded, as referred in Chapter 1, that the

characteristics of the provenance environment (e.g. burial conditions) might also influence the

corrosion processes [7].

Knowledge of the environmental conditions, the nature of the materials, and the thermo-

mechanical techniques used during the object fabrication is thus required to define an analytical

protocol for the characterisation of corroded layers formed on gold alloys. This first assessment

with the analytical techniques selected in Chapter 2 of three case studies, illustrating different

situations in national and foreigner museums, should evidence the advantages, disadvantages and

limitations of each technique.

The environmental influence on the atmospheric corrosion of gold objects was assessed based on

the environmental conditions study of the Treasure room, at the National Museum of Archaeology

(NMA) in Lisbon (Portugal). The gold objects exhibited in this room show accentuated corroded

surfaces suggesting the presence of atmospheric conditions favourable to the corrosion

development. The possibility to carry out the environmental study of this room and move three

corroded objects to a laboratory for characterisation, allowed to relate the corrosion products

composition to the exhibition conditions (including temperature, humidity, pollutants and


44

materials used for the showcases) and to identify the main pollutants responsible for the high

corrosion rate.

The influence of the alloy composition and fabrication techniques on the corrosion development

was also approached for the three studied objects of the NMA. However, to enlarge the corpus of

gold objects and include those that are composed of different materials and made by employing

distinct manufacturing techniques, it was studied a group of jewels created by René Lalique in the

collection of the Calouste Gulbenkian Museum (CGM) in Lisbon (Portugal). These jewels are

complex mountings of parts made from gold alloys, silver alloys, enamel and precious stones. The

study in-situ of a group of corroded jewels, and the possibility to characterise the corroded layer

of one object by moving them to the laboratory, allowed to assess the influence of the fabrication

techniques on the corrosion development and to identify the difficulties of the corrosion

characterisation caused by the presence of multiple materials.

As mentioned in Chapter 1, the corrosion of gold objects is visually characterised by a surface

colour change, which can acquire a multi-hued reddish colour. The corroded surface colour can

be influenced by the corrosion products formed at the surface and by the corrosion layer

thickness. To relate the corrosion products and the corrosion layer thickness with the surface

colour, the corroded surface morphology should be assessed, the corrosion products identified

and the corrosion layer thickness measured. This issue could not be approached for the objects in

the NMA and in the CGM due to the limited number of objects that could be moved for analysis.

Their dimensions and shape also disable the use of some selected analytical techniques, like the

SEM. To approach this issue, three Egyptian gold foils from the National Museums of Scotland and

from the Garstang Museum of Archaeology, University of Liverpool were studied. The dimensions

of the gold foils allowed to characterise the corroded layer by stationary techniques and hence to

gather detailed utmost information on the corrosion layer formation.


45

3.2 Exhibition and storage environment assessment: the National Museum

of Archaeology, Portugal

3.2.1 The “Treasures of Portuguese Archaeology” collection conservation assessment

The “Treasures of Portuguese Archaeology” exhibited in the so-called Treasure room of the

National Museum of Archaeology (NMA) (figure 3.1), in Lisbon, is composed by circa 700 objects

that give an overview of the jewellery evolution in the present Portuguese territory, since the

beginnings of metallurgy until the Early Middle Ages [8]. This permanent exhibition, inaugurated

in 1980, was object of renewal in 2000.

Figure 3.1 – The Treasure room of the NMA.

As the gold objects displayed colour alterations, the showcase fabrics and the room carpet were

replaced. In this renewal, it was also included the conservation treatment of some gold objects

that consisted of a surface cleaning with ultra-thin sodium carbonate. However, a few years later,

the surface colour of the majority of the gold objects significantly changed. A reddish colour with

a multihued effect, as depicted in figure 3.2A, became visible at the surface. Between 2009 and

2012, this corrosion process continuously grew, accompanied by an important surface colour

alteration. Figures 3.2B and 3.2C show the evolution of an Iron Age torc surface colour over three

years.


46

Figure 3.2 – Iron Age torc from Codeçais, Bragança (NMA - Au 1139) and the detail of the terminal illustrating the corroded surface in 2009 (A) and 2012 (B and C).

All the showcases containing gold objects in the exhibition room were opened for the collection

conservation assessment. The location of each object in the showcase was identified, the

fabrication techniques described, the alloy composition determined, and the conservation

condition itself (presence of fissures, fractures, etc.) examined [9]. The data obtained showed that:

i) the object location in the showcase is not related to the corrosion increase;

ii) an accentuated corrosion development is observed in areas subjected to thermal processes,

such as hard-soldering, or where thermo-mechanical deformations were applied. These processes

occurred during both manufacturing and previous restoration procedures.

Several parameters contribute to the corrosion development. To further identification of the

causes of degradation, the climate and the materials of the Treasure room were studied and the

corrosion products on the gold alloy characterised.

3.2.2 Climate and materials characterisation

The NMA is located at the Jerónimos Monastery. This historical building, located in the western

area of Lisbon, is implemented at 400 m from the north shore of the Tagus River, at 8 km from the

river mouth. The city of Lisbon has a Mediterranean macroclimate, characterised by a summer

with high temperatures and dry air, and with the annual precipitation concentrated between

October and April. This classic Mediterranean climate is, however, modified by the regional

topography, particularly by the nearness of the Atlantic Ocean and the Tagus river. The average

of the monthly temperature varies regularly throughout the year, being the average annual

temperature (T) of 17 °C. The annual relative air humidity (RH) average is high, ranging between


47

72 % and 79 %, which turns Lisbon with RH levels climate similar to the moderated tropical regions,

as, for example, the city of São Paulo, Brazil [10]. Generally, during winter and autumn, the RH

average values are above 80 %, while during spring and summer they range between 60 % and 70

%. These very high RH levels raise the risk of metal degradation.

The high RH average annual values led to the characterisation of the environmental behaviour of

the Treasure room to verify the influence of the outdoor climate. This room was designed as a box

container at the northeast gallery of the Jerónimos Monastery, on the ground floor, with an air

conditioning system. This system only allows the temperature control, and it is programmed to

maintain the room temperature at 19 °C.

Figures 3.3A and 3.3B show the typical behaviour of the NMA indoor and outdoor RH and T, during

approximately one month, when the museum was opened to visitors. The Treasure room is

permanently open (available for visitors) during the museum opening hours: 8 hours per day, 6

days a week. Due to the air exchange, equivalent indoor and outdoor RH fluctuations could be

observed. Inside the Treasure room, the RH reached 80 %. The temperature fluctuation is less

pronounced, ranging from 17 °C to 22 °C. The maximum seasonal variations in T and RH are

respectively of 4 °C and 26 % during autumn/winter.

Figure 3.3 – RH and T measurements for one month taken inside (A) and outside (B) the Treasure room.

Figure 3.4 shows the RH and T values registered inside the Treasure room when the room was

temporarily closed for visitors. During this period, there is no variation in T (19 °C) and RH (75 %),

which allows assuming that the permanently opened access door is the main factor of the

accentuated RH and T variation in figures 3.3A and 3.3B.


48

Figure 3.4 – RH and T measurements for one month taken inside the Treasure room in a period during which the room was temporarily closed.

Fluorescent lamps lighten the room and the showcases. Inside each showcase, there are 2 to 4

fluorescent lamps, depending on the showcase dimension. The UV emission values are below

45µW/lumen, and the illumination values range between 80 and 400 lux, in function of the

distance from the light source. Most of the values are under 150 lux, which is the recommend limit

in the case of metallic objects [11]. Nevertheless, the measured values might accelerate the cover

fabrics degradation [12], increasing the emanation of volatile organic compounds.

The main pollutant sources of degradation are the exhibition materials, which during their

degradation process release volatile organic compounds that gradually induce corrosion on the

displayed objects [13]. The Treasure room is a wood and pressed wood (sometimes painted) box

with elements in steel, built inside the stone historical building. Besides wood and pressed wood,

the showcases also contain aluminium, glass and two types of covering fabrics, one blue and

another green (figure 3.1).

The principal atmospheric pollutants reported for museum context, NO2, SO2, COS, CS2, H2S,

CH3COOH, CH2O, C2H6S and Cl, were identified and quantified by diffusion tubes. These diffusion

tubes have the advantage of not requiring a mechanism for pumping the air, giving the

information in-situ and in a short-time frame [14]. Table 3.1 gives for each gas the amounts

measured inside and outside a showcase of the Treasure room. Table 3.1 also shows the detection

limits that are, for some pollutants, higher than the values reported for indoor pollutants in

museum contexts [15-18].


49

Table 3.1 – Gaseous pollutants concentrations in μg g-1 measured at 55 % RH and 18 °C T inside a showcase from the Treasure room by using diffusion tubes.

Inside the showcase, SO2, CH3COOH and CH2O are present at much higher concentrations than

those reported by Bastidas et al. [19] for a showcase in the Gold Museum in Bogota, Colombia,

where corroded gold alloy objects were exhibited. These authors have measured 0.005 µg.g-1 of

SO2 and 0.11 µg.g-1 of volatile organic compounds.

The detection of SO2, CH3COOH and CH2O only inside the showcase suggests that the source of

pollutants is related to the showcases materials. To identify the pollutants source, qualitative

corrosivity tests, known as Oddy tests, were carried out for the materials used in the exhibition

room: the pressed wood, the wood, the two fabrics, and the glues [20]. The test consists of placing

samples of the materials to be tested in a sealed borosilicate glass flask, together with pure Ag,

pure Cu and pure lead (Pb) coupons with deionised water to create a high RH atmosphere. The

flasks were then placed in an stove at 60°C for 28 days. At the end of this period the surface of

coupons is characterised to search for corrosion products [21]. Each metal is used to detect the

presence of different pollutants. Ag is used to detect reduced sulphur compounds and carbonyl

sulphides, Pb to detect organic acids and aldehydes and Cu to detect chloride, oxide, and sulphur

compounds [16].

Ag, Cu and Pb coupons were also placed inside and outside the showcases to identify the

environmental pollutants. Inside, the coupons were placed directly on the fabrics, close to the

lightening sources, and outside, they were placed in the exhibition room close to the exit of the

air conditioning system.

The only Cu coupon presenting corrosion, identified by XRD as cuprite, Cu2O, was the one placed

outside the museum, being impossible to withdrawn information on pollutants source based on


50

this coupon. XRD analysis of the Pb coupons placed inside the room identified lead formate,

Pb(CHO2)2 and hydrocerussite Pb3(CO3)2(OH)2. These compounds are related to the presence of

volatile organic compounds, like acetic acid identified by using diffusion tubes, which can be

attributed to the degradation of wood and pressed wood. Analysis of the Ag coupons placed inside

the showcases identified acanthite, Ag2S, which can be explained by the presence of wool fabrics.

The degradation of cysteine-containing protein causes the release of reduced sulphur compounds,

like carbonyl sulphide, sulphur oxide and carbon disulphide, the later identified by using diffusion

tubes [14]. In fact, the morphology of the showcase fabrics fibres observed by SEM shows that

both contain wool. The green fabric is made of wool only, and the blue fabric contains, as shown

in figure 3.5, wool and polyester fibres. The Ag coupons placed outside the showcase and close to

conditioning outlet system revealed the presence of chlorargyrite, AgCl. This compound is

expected in museum environments located near the sea. It should be emphasised that sulphur

based corrosion products are dominant in indoor environments (contrary to the chlorine-based

dominant corrosion products in outdoor environments) [22].

Figure 3.5 – SEM image of the blue fabric showing the wool fibres with scales characteristics of the wool fibres morphologies mixed with polyester fibres, exhibiting a flatter surface [23].

Considering the results obtained for the pollutants and the environmental conditions of the

Treasure room, it was possible to suggest that the sulphur released by the wool fabrics inside the

showcases and the high RH inside the room are the main cause of the gold objects atmospheric

corrosion. The reaction between the gaseous sulphur and the gold alloy, promoted by the

presence of high RH, allows the formation of sulphur-based corrosion products. These corrosion

products can thus be composed of S, O, Ag, Au and Cu. Analysis of the corroded layers should


51

provide the identification of the corrosion products responsible for the surface colour alteration

observed for the gold objects.

3.2.3 Characterisation of the corroded layer

Three objects showing accentuated corrosion could be moved for analysis. Their small dimensions

allowed the characterisation of their corroded surface by XRD and SEM-EDS and also by XRF for

determination of their alloy composition.

Figure 3.6 shows the Bronze Age torc from Serpa, Beja, the Bronze Age fragment of torc from

Alentejo but with unknown specific provenance, and the Iron Age earring from Cabeço de

Vaiamonte, Monforte, selected for the study. The colour of their corroded surfaces is distinct. The

earring shows a more accentuated corrosion with a multi-colour effect, particularly in the main

ring. Both torcs show a corroded red colour surface. Like the torc from Codeçais in figure 3.2, the

corrosion of the torc from Serpa is more developed on the terminals. The fragment of torc from

Alentejo shows an increase of the red colour in the decorated areas.

Figure 3.6 – Bronze Age torc (MNA- Au 293) from Serpa (A); fragment of a Bronze Age torc (MNA-Au 283) from Alentejo (B) and Iron Age earring (MNA-Au 574) from Cabeço de Vaiamonte, Monforte (C).

Table 3.2 presents the elemental composition of the three objects determined by portable XRF.

The earring is made from a gold alloy containing 4 wt% Cu and the highest Ag content, which could

justify a higher corrosion rate and a more accentuated colour change.


52

Table 3.2 – Composition of the base-alloy obtained by XRF for the two Bronze Age fragment torcs and the Iron Age earring.

In the recessed decorated areas of the torc fragment from Serpa the presence of acanthite (Ag2S)

and petrovskaite (AuAgS) could be identified by XRD. The corroded areas of the earring showed

instead the presence of Ag2S and uytenbogaardtite (Ag3AuS2). As suggested above, the presence

of two distinct Au-Ag-S compounds on the torc from Serpa and on the earring could be attributed

to the use of different alloys. However, data is insufficient to infer on the influence of the alloy

composition on the corrosion development.

It should be emphasised that both the torc and the earring were subjected to a severe mechanical

deformation during fabrication. By observation with SEM, the presence of corrosion products with

different morphologies could be revealed. Figure 3.7 shows the surface of the torc fragment from

Alentejo, where agglomerates of small rounded particles with a more complex structure could be

evidenced in the areas where corrosion is more developed. This morphology suggests an evolution

of the surface structure through time. The surface structure changes as corrosion products grow,

and the formed film becomes more uniform. Further surface changes could be observed. The

presence on the same corrosion products with different morphologies suggests a continuous

corrosion process that results in the formation of different products corresponding to distinct

phases of corrosion [24]. However, it was impossible to establish a direct relation between the

manufacturing techniques and the surface corrosion.


53

Figure 3.7 – SEM image of the corroded surface of the torc from Alentejo (NMA-Au283) composed of small rounded particles. Areas with a more developed corrosion show a structure that can be related to distinct corrosion products.

Figure 3.8 illustrates the EDS spectra obtained for corroded and non-corroded areas of torc from

Alentejo. In the corroded areas, it can be seen an enhancement of the right tail of the Au-Mα peak

caused by the presence of the S-Kα peak. The peak of Ag is higher for the corroded areas,

suggesting the formation of Ag-based corrosion products. The EDS spectrum of the corroded areas

also reveals the presence of Cl and Na. These elements can be related to the presence of NaCl

particles. NaCl has been reported for atmospheric corrosion in coastal marine atmospheres [25],

as seawater contains 3.5 wt% of this compound. Cl can also be related to the presence of AgCl, a

Ag atmospheric corrosion product. In fact, this compound was identified for the Ag coupon placed

close to conditioning outlet system of the exhibition room. NaCl and AgCl were not identified by

XRD, maybe because too small amounts are present on the object surface.

Table 3.3 gives the concentrations obtained by EDS for the torc from Alentejo. The above

mentioned possible increase of the Ag content in the corroded area could be quantified for that

area. The increase of this element is in accordance with the XRD results, confirming that Ag plays

an active role in the corrosion of gold alloys.


54

Figure 3.8 – EDS spectra of the fragment torc from Alentejo (MNA-Au 283) obtained in areas with and without visible corrosion. The presence of higher Ag and S contents in the corroded areas indicates the possible formation of S-based and Ag-based corrosion products.

Table 3.3 – Composition of the corrosion layer obtained by EDS compared to the Au-base alloy of the fragment torc from Alentejo (MNA-Au283).

3.2.4 Final remarks

The environmental assessment of the Treasure room at NMA was important to identify the

atmospheric corrosion pollutants of the gold objects and their sources. It was shown that the high

variation of RH inside the exhibition room, and the presence of SO2, CH3COOH and CH2O, gaseous

pollutants influence the corrosion development on gold surfaces, and promote the formation of

S-based corrosion products. The exhibition materials (wood, press wood and fabrics) were

identified as sources of pollutants. The presence of those compounds in concentrations higher


55

than those reported for indoor museum environments may also justify the accentuated corrosion

development on the gold objects from this collection.

Three objects were moved to the laboratory to be analysed by SEM-EDS and XRD. The corroded

surface morphology of the fragment of torc from Alentejo studied by SEM showed the presence

of agglomerates of small rounded particles, with a more complex structure in the areas where

corrosion is more developed. The relation between the manufacturing techniques and the surface

corrosion was also searched. The objects that were subjected to severe mechanical stress, like the

fragment of torc from Alentejo, presented an accentuated corrosion, but it was impossible to

establish for this object a relation between the manufacturing techniques and the surface

corrosion. Ag2S and AuAgS were the corrosion products identified on the earring and on the

fragment of torc from Serpa, confirming the presence of S-based compounds. The results indicate

that Ag plays a major role in the corrosion process of gold alloys, but data was insufficient to infer

on the influence of the alloy composition on the corrosion development.


56

3.3 Composite objects: the Jewellery of René Lalique

3.3.1 The Jewellery of René Lalique collection

The Calouste Gulbenkian Museum (CGM) in Lisbon exhibits over eighty René Lalique’s creations,

made between 1899 and 1927. The jewellery of René Lalique, representative of the art disruption

of the early 20th century [26,27], is characterised by a profusion of colour, transparency and bright

obtained using gold and silver alloys combined with plique-à-jour enamels, stones and other

materials (glass, ivory, horn, etc.) [28].

Several gold jewellery pieces currently show a colour alteration, a red-hued surface, caused by

atmospheric corrosion (figures 3.9A and 3.9B). Lalique’s jewellery pieces are often metallic plates

supporting transparent coloured materials. The artist searched for light reflection, transmission,

refraction, scattering and absorption effects, enhanced by the colour of the gold substrate. It is

suggested that he obtained nuances of gold colour by varying the composition of the alloys and

by applying specific treatments to their surface [28,29].

Figure 3.9 – Details of red corroded surfaces of the pendants Peacocks on prunus (CGM-1203) (A) and Nymph in a tree (CGM-1165) (B).

The alteration of the jewels colour has a direct impact on their perception. Conservation

treatments should be carried out to restore their original aspect, or if the original aspect cannot

be again attained, preventive measurements should be defined. For any of the situations, it is

necessary the identification of the corrosion products. The corrosion products growing on the


57

objects surfaces depend, in addition to environmental parameters as referred in the previous

section, on the gold alloy composition and the possible surface treatments applied to the object.

It is also expected that in a composite object, the alteration of one material influences the

alteration of another material. Considering all these possibilities, a group of corroded objects, was

selected to be studied in-situ. The conservation condition of the corroded surfaces was assessed,

the fabrication techniques were characterised, and the alloy compositions were determined by

using portable XRF. One brooch of smaller dimensions was moved for SEM-EDS analysis to

characterise the corrosion layer.

3.3.2 The gold alloy composition and fabrication techniques

Figure 3.10 shows two ternary diagrams representing the Au, Ag, and Cu concentrations obtained

by portable XRF for a group of brooches and pendants that give indication on the melting

temperatures of the alloys and their colour. Data show that René Lalique used Au-Ag-Cu alloys

with Au contents that range from 75 to 84 wt%, situated in a colour region that corresponds to

yellow (based on McDonald and Sistare [29]). The melting points of these alloys ranges from 900

°C to about 1000°C. The melting point diversity can be justified by the objects complexity, made

by hard-soldering several parts. In spite of Ag contents ranging between 13 and 25 wt% and Cu

contents ranging between 1 e 12 wt% in the alloys, it was observed no relation between the alloy

composition and the corrosion development.

Figure 3.10 - Ternary diagrams (Au,Ag,Cu) in wt% of the pendants and brooches analysed by XRF, with the colour (A) and the melting temperatures (B) adapted from McDonald and Sistare [29].


58

The main technique used by Lalique to form his jewellery creations is casting, but several objects

present casting defects. Observation under the stereomicroscope of the bracelet pair ensnared

showed the presence of dendritic structures, which result from the heterogeneous nucleation of

the metal (figure 3.11).

Figure 3.11 – Dendritic structure at the surface of the bracelet Pair ensnared (CGM-1180), a result of the metal solidification process.

In the case of the brooch Ophelia that was moved for study, the observation under the SEM (figure

3.12) revealed the presence of a microstructured morphology that can be due to a high porosity

defect related to metal shrinkage during solidification [30]. In addition, in other areas the presence

of small rounded particles could be seen on the corroded areas (figure 3.13A), similar to those

observed on the surface of the torc fragment from Alentejo studied in the previous section (cf.

figure 3.7).

Figure 3.12 – SEM image of brooch Ophelia (CGM-1138) showing the microstructure characteristic of porosity defects.


59

The areas of brooch Ophelia with more accentuated corrosion show the formation of a uniform

corrosion layer (figure 3.13B) that developed in a preferential direction, along the file marks from

the surface finishing process, as it can be seen in figure 3.14. The corrosion develops preferentially

on rougher surface areas, which have more active sites for the particles nucleation.

The casting defects cited above increase the surface irregularity, modifying the light scattering

and decreasing the reflectivity [31], which influences the colour and shine perception of the

object. This surface colour alteration due to casting can be misinterpreted as surface corrosion.

Figure 3.13 – SEM images of corroded areas of brooch Ophelia (CGM-1138). Different morphologies were observed: small rounded particles corresponding to the nucleation of corrosion products (A) and a uniform corrosion film (B).

Figure 3.14 – SEM image of the file tool marks on brooch Ophelia (CGM-1138).


60

When the composition of corroded and non-corroded areas analysed by EDS are compared (table

3.4), it is observed an increase of the Ag content and the presence of amounts of S, Cl and O,

suggesting the formation of Ag- or Ag-Au- sulphides, and of Ag chlorides or oxides.

Table 3.4 also includes the results obtained by XRF for the brooch Ophelia base alloy, revealing a

difference between the surface and the substrate compositions. With the selected experimental

conditions and for this type of gold alloys, the information depth for Ag is nearly 30 μm, while for

EDS it under 0.5 μm. This confirms, as suggested by the morphology, that on the surface there is

an increase of the Ag content and a decrease of the Cu content.

Table 3.4 – Composition normalised to 100 wt% of the corrosion areas and of the Au-based alloy obtained by EDS and XRF.

Some jewels in the collection show cracks that are localised in regions subjected to important

thermo-mechanical processes or submitted to high temperatures (like when joining by hard-

soldering, hammering to bend, etc.) as illustrated in figure 3.15A. In these regions, it can also be

observed a colour change, as observed for the NMA earring (discussed in page 51). In this case the

colour change can be explained by the presence of either corrosion products or a different

microstructure with a distinct reflectivity (figure 3.15B).

Composition in wt% Au Au Ag Ag Cu S Cl O Lα Mα Kα Lα Kα Kα Kα Kα Base alloy XRF 81 16 3 EDS 81 17 2 Corroded areas Particles 15 54 2 2 11 16 45 38 2 2 7 6 Film 65 16 2 2 - 16 68 18 1 - - 13


61

Figure 3.15 – Cracks in the gold alloys of pendants The Abduction (CGM-1173) (A) and Peacocks on prunus (CGM-1203) (B).

The presence of these cracks can accelerate the degradation of other materials applied on the

gold substrate. An example of this situation is when enamels are applied on the gold plate, over a

silver foil. The gold plate cracking produces cracks on the enamel surface. In turn, when the

enamels cracks, the silver foil is exposed to humidity, oxygen and gaseous pollutants that induce

its corrosion. The corrosion of silver renders a black colour to the surface, contributing to the

misperception of the jewels, as shown in figure 3.16 for pendant Peacock on prunus. As for the

silver foil, the gold alloy substrate in contact with degraded enamels can also undergo corrosion.

However, its characterisation is highly complex due to the layered structure - gold substrate; silver

foil; enamel, disabling the access to the possible corroded gold surface.

Figure 3.16 – Corrosion of the silver foil on the gold plate under the enamels of the pendant Peacocks on prunus (CGM-1203), due to enamel cracks.


62

3.3.3 Final remarks

The selected group of corroded gold jewels studied in-situ by XRF showed that René Lalique used

gold alloys with Ag contents ranging between 13 and 25 wt% and Cu contents ranging between 1

and 12 wt%. Although, different Ag and Cu contents could result in the formation of distinct

corrosion products, it was observed no relation between the alloy composition and the corrosion

development.

The study of the brooch Ophelia by SEM-EDS showed the influence of the fabrication techniques

on the corrosion development. The presence of casting defects can increase the surface

irregularity, modifying the light scattering and decreasing the reflectivity, influencing the colour

and shine perception of the object. The EDS results showed an accentuated increase of the Ag

content and a decrease of the Cu content on the corroded regions and the presence of S, Cl and

O, suggesting the formation of Ag- or Ag-Au- sulphides, and of Ag chlorides or oxides. The presence

of Cu-based corrosion products was not detected.

The presence of different constituent materials on the jewels contributes to their degradation.

The cracks on enamels can promote the corrosion of silver foils and of the gold alloy substrate.

However, due to the layered structure of these objects, it was not possible to characterise their

corroded surfaces.


63

3.4 Ancient Egypt gold jewellery objects

3.4.1 The corrosion of Egyptian gold objects

Corroded surfaces with different colours have been ascribed to distinct thicknesses of the

corrosion layer of gold and silver alloys [32,33]. In the case studies presented in the previous

sections, although the surface of corroded objects shows different colours, their study was not

possible once the objects showing the multi-hued effect are too big, disabling the characterisation

by SEM-EDS. In this section two corroded Egyptian foils and one bead, of small dimensions that

could be study by SEM-EDS, µXRF and XRD were selected to research the relation between the

surface colour, the corrosion products composition and the corrosion thickness.

Corrosion of Egyptian gold objects has been currently referred since 1926, when this phenomenon

was reported for the first time by Alfred Lucas [34]. Frantz and Schorsch [35] and Rifai and El

Hadiddi [36] refer that the red colouration of Egyptian gold alloys is caused either intentionally or

by corrosion. Colour is an important characteristic of the Egyptian gold jewellery [37], but it is also

an important characteristic used to visually identify the presence of corrosion. The reddish colour

of gold, characteristic of the corroded gold alloys, can be misread as intentional. Frantz and

Schorsch [35], underline that the red colouration only appears on specific areas of the objects

such as, for example, part of the gilded face of the coffin of Nephthys from the Metropolitan

Museum of Art in New York. These same authors attributed the red colour surface to colloidal

transformations of the gold alloy and to the formation of sulphide- and chlorine-based corrosion

products of copper, silver and gold, namely Ag-Au-S complexes. Although the authors mentioned

above focused their studies on the surface colour, none of them searched the influence of the

nature of the developed corrosion products on neither the corrosion layer thickness nor the gold

alloy composition.

The three selected objects for this work includes two gold foils and one bead that come from

archaeological excavations and were subjected to burial corrosion before the exhibition in

museums (table 3.5). The two gold foils from the Garstang Museum of Archaeology (GMA)

(University of Liverpool) collections were found at the tombs of the so-called “North Cemetery”

at Abydos, Upper Egypt and were probably attached to a wooden artefact. This site is thought to

have contained burials of individuals from the Middle Kingdom (mid-11th Dynasty– 13th Dynasty,

c. 2055-1650 BC) [38]. The bead from the National Museums of Scotland (NMS) collection is from

a necklace found in an intact tomb of a young girl at the cemetery of Harageh, in an area between

the Nile and the city of Fayum, Middle Egypt. The intact burial dating to the Middle Kingdom (mid-


64

late 12th Dynasty, c.1875-1795), contained several beads, jewellery and other findings like steatite

and turquoise scarabs [39].

Table 3.5 Egyptian gold objects selected for surface corrosion characterisation with indication of the provenance, date and local of conservation.

3.4.2 Characterisation of the corroded surface

3.4.2.1 Observation

Two distinct surface colour alterations could be observed with naked eye: i) an heterogeneous

coloration with a colour variation going from dark yellow, violet and blue until red (figure 3.16A)

and ii) an homogeneous red coloration (figure 3.16B). The foils show the two coloration types,

whereas the bead only present the homogeneous surface coloration.

Figure 3.17 –Fragment foil (GMA-432-25A) from Abydos excavations showing a heterogeneous corroded surface with colours varying from dark yellow, red and blue (A); and bead from Harageh (NMS-A1914.1096) with homogeneous corrosion with a red colour surface (B).

Object Ref. nº Provenance Date Local of conservation Gold foils GMA-432-25A

GMA-432-25B Abydos Middle Kingdom

(c. 2055-1650 BC) Garstang Museum of Archaeology (GMA) (University of Liverpool)

Bead NMS-A1914.1096

Harageh Middle Kingdom (from one tomb dated from the 12th dynasty (c. 1875–1795 BC)

National Museums Scotland (NMS)


65

3.4.2.2 Analysis

The elemental composition of the different colours observed on the corroded surfaces was

estimated by μXRF. The equipment used has a beam spot size of 25 µm, allowing to analyse each

area with a different colour and to to relate it with the surface composition. Both heterogeneously

and homogeneously coloured areas were analysed for comparison.

Data obtained for the foil (GMA-432-25A) with heterogeneous colour show an increase of the Ag

content in the corroded areas without any significant change on the S or Cl contents, suggesting a

relation between the colour of corrosion products and the Ag content at the surface, as shown in

figure 3.18. This increase of the Ag content indicates the formation of an Ag-based corrosion

product, which is in accordance with the results published by the few authors who published data

on corroded Egyptian gold objects [35,36].

Figure 3.18 – Au and Ag contents obtained by µXRF for the different corrosion colours of the Abydos foil (GMA-432-25A).

Data obtained for the bead and for the foil (GMA-432-25B) with homogeneous red coloration

revealed an increase of the S content, as shown in figure 3.19, and only a slight increase of the Ag

content.


66

Figure 3.19 – Au and S contents obtained by μXRF for the corroded and non-corroded areas of the Abydos foil (GMA-432-35B) and the Harageh bead (NMSA1914.1096) with homogenous corrosion.

The corroded surfaces were observed under the SEM-EDS and it could be verified that the

homogeneous and the heterogeneous colorations correspond to distinct morphologies. The

surface morphology of the heterogeneous colour alteration is a porous structure layer, with the

accentuated corrosion areas being constituted by thin and solid tubes (figure 3.20A) that are

morphologically different according to the distinct corrosion colour. Figure 3.20B illustrates the

red corrosion area constituted by a homogeneous layer and figure 3.20C shows the blue corrosion

area where three distinct layers can be observed. The first layer (1) is morphologically similar to

the red corroded surface, the second layer (2) is characterised by agglomerated corrosion

products, and the third (3) is characterised by thin tubes randomly distributed. This last layer

corresponds to areas with accentuated corrosion. Data obtained by EDS also show an increase of

the Ag content that varies according to the distinct corrosion layers (table 3.6). The Ag content,

which for the base alloy is 10 wt%, ranges from 54 wt% to 70 wt% for the blue corroded area and

it attains 70 wt% in the red corroded area. This foil was analysed by XRD, but no corrosion

compound could be identified, maybe due to the presence of too small quantities of those

compounds or to a non-crystalline structure.


67

Figure 3.20 – SEM images for the heterogeneous colour corrosion surfaces on Abydos foil (GMA-432-25A). Morphology of the corrosion products featuring thin tubes (A); Surface morphology of the red corroded area (B) and of the blue area (C). The blue area is composed of three distinct layers: 1) one nearer the substrate, morphologically similar to the red corroded areas, 2) a second with corrosion products featuring agglomerates and 3) a third with corrosion products featuring thin tubes. Table 3.6 – Composition of the corrosion layer obtained by EDS and composition of the Au-base alloy obtained by XRF for the Abydos foil (GMA-432-25A) presenting a heterogeneous colour surface.

The morphology of the foil (GMA-432-25B) with homogeneous coloration shows aggregates of

small round particles forming a nanoporous layer, similar to the one previously described for the

red corrosion areas of the heterogeneous colour corrosion (figure 3.21A). The areas with more


68

accentuated corrosion present hollow polycrystalline tubular formations randomly oriented on

the top of the porous layer, also indicating a formation of a layer-by-layer corrosion (figure 3.21B).

The corrosion products observed on the bead (figure 3.21C) are mixed with silicates from the

burial context, which disabled the development of tubes. However, they present the same type

of surface morphology with two layers, the first composed of aggregates of small round particles.

Figure 3.21 – SEM images for the homogeneous colour corrosion surface of the Abydos foil (GMA- 432-25B) (A); a detail of the polycrystalline tubular formations of the corrosion products (B) and homogeneous colour corrosion on the bead from Harageh (NMS-A1914.1096) composed by corrosion products mixed with silicates from burial (C).

The EDS analyses of the first layer of the foil (GMA-432B) show an increase of the Ag contents and

an increase of the S contents while for the bead an accentuated increase of the Ag content could

be revealed (table 3.7). The area with a higher concentration of tubular formations on the foil

shows an increase of the S and Ag contents, respectively 9 wt% and higher than 20 wt%.


69

Table 3.7 – Composition of the corrosion layer obtained by EDS and composition of the Au-base alloy obtained by μXRF for the Abydos foil (GMA-432-25B) and for the Harageh bead (NMS-1914.1096) presenting a homogeneous colour surface.

3.4.3 Final remarks

The study of the two Egyptian corroded gold foils and one bead by µXRF, SEM-EDS and XRD

allowed relating the surface colour to its morphology, composition and to the thickness of the

corroded layer.

The elemental analysis of the different coloured corroded areas by µXRF showed a relation

between the surface colour and the Ag and S contents. The SEM-EDS results suggest that the

colour is also related to the presence of different corrosion products, with distinct compositions,

developed in a layer-by-layer structure. The results also show that the gradual progress of

corrosion modifies the structure of these layers. With time, each layer becomes more uniform and

compact. The nucleation of new particles on the top of each layer indicates that corrosion is a

continuous process. However, it was impossible to identify any corrosion compounds by XRD,

perhaps because present in small quantities or as less crystalline compounds.

The thickness of the Abydos foils could not be measured due to the highly absorbing corrosion

layer and the surface roughness, which induce light scattering disabling the use of ellipsometry.


70

3.5 Conclusions

The aim of studying corroded gold objects was to evaluate whether the chosen analytical

techniques could be applied to the characterisation of the corroded surfaces based on: the

assessment of the environmental conditions, the identification of the alloy composition, the

description of the fabrication techniques, and the identification of the corrosion products.

In this work, a portable XRF was used for in-situ analysis and µXRF, SEM-EDS and XRD for ex-situ

analyses of the objects that could be moved to the laboratory and the coupons used for the

environmental assessment of the Treasure room of the NMA.

The Treasure room, as a case study of an exhibition room with objects presenting an accentuated

corroded surface due to atmospheric corrosion, allowed to infer on the main pollutants for indoor

corrosion of gold objects and to identify their sources. The environmental assessment identified

the high variation of humidity inside the exhibition room and the presence of S and Cl as elements

that influenced the corrosion development on the gold surfaces. Nevertheless, and as expected

for indoor corrosion, the identified corrosion products are composed of S-based compounds.

The study of the jewels created by René Lalique showed the influence of the fabrication

techniques on the corrosion development. It was shown that the surface colour alteration could

be due either to the formation of corrosion products or to a change in the light reflectivity. The

latter phenomenon is caused by a modification of the surface morphology, resulting from

problems related to the fabrication techniques employed and surface finishes. Data obtained for

the Egyptian foils suggest, however, that the colour is also related to the presence of different

corrosion products, with distinct compositions, developed in a layer-by-layer structure. However,

it was not possible to establish the composition of each layer and determine their thickness.

Although it was shown in the three case studies that Ag plays an important role on the corrosion

of gold alloys, since the identified corrosion products are mainly composed of Ag-based

compounds, it was impossible to infer on the influence of the alloy composition, namely the Ag

and Cu contents, on the corrosion mechanisms.

The three case studies allowed identifying different situations that should be addressed to define

an analytical protocol for the study of the corrosion mechanisms of the gold alloys.

The prehistoric gold objects from the NMA and the jewels made by René Lalique were studied in-

situ and ex-situ. With the XRF portable equipment used in this work, it was possible to identify the

alloy composition, but it was impossible to characterise the elemental composition of the


71

corroded layer, since the penetration depth is higher than the corrosion layer thickness. In fact,

the characterisation of the corrosion layer of the objects was only possible because a group of

objects was analysed ex-situ by SEM-EDS and by XRD. This situation, justified by security reasons,

puts in evidence the analytical limitations based on restrictions of moving the objects.

Considering the results obtained ex-situ, and including those on the Egyptian foils analysed by

µXRF, SEM-EDS and XRD, several comments can be addressed. The µXRF, with a higher spatial

resolution than the portable XRF, allowed the elemental analysis of the different colours of the

corroded surfaces and showed that Ag and S contents are different for each of the corroded

colours.

The SEM was fundamental to study the questions related to the fabrication techniques. The study

of the brooch Ophelia showed a microstructured morphology resulting from a fabrication defect,

demonstrating the importance of high resolution imaging. In the case of the NMA objects, the

brooch Ophelia and the Egyptian foils, it was identified by SEM different corrosion morphologies

and the development of a layer-by-layer structure. With EDS it was possible to analyse each of the

corroded layers and infer their composition. However, the quantification of the S contents was

complex. Although its presence is visible in the spectra, its quantification by EDS is difficult due to

the right tail of the intense peak corresponding to the Au-M lines that are close to the S-Kα line.

The identification of the crystalline compounds could only be obtained by XRD for some objects.

Maybe the corrosion products are present in small quantities or have low crystallinity.

Another identified analytical constraint was related to the determination of the corrosion layer

thickness by ellipsometry. The surface roughness of a Egyptian foil scattered the light, disabling

the use of this technique.


72

3.6. References

[1] J. Tidblad, Atmospheric corrosion of metals in 2010-2039 and 2070-2099, Atmos. Environ. 55 (2012) 1‒6. [2] Schieweck, T. Salthammer, Indoor air quality in passive-type museum showcases, J. Cultural Heritage 12 (2011) 205‒213. [3] C.D. Grzywacz, Monitoring for gaseous pollutants in Museum Environments, Getty publications, Los Angeles, 2006. [4] P. Brimblecombe, D. Shooter, A. Kaur, Wool and reduced sulphur gases in museums, Studies in Conservation 37 (1992) 53–60. [5] C. Cason, L. Pezzato, M. Breda, F. Furlan, M. Dabalà, Effect of microstructure and residual stresses, generated from different annealing and deformation processes on the corrosion and mechanical properties of gold welding alloy wire, Gold Bull. 48 (2015) 135–145. [6] S. Viennot, M. Lissac, G. Malquarti, F. Dalard, B. Grosgogeat, Influence of casting procedures on the corrosion resistance of clinical dental alloys containing palladium, Acta Biomaterialia 2 (2006) 321‒330. [7] M. Kibblewhite, G. Tóth, T. Hermann, Predicting the preservation of cultural artefacts and buried materials in soil, Sci. Total Environ. 529 (2015) 249‒263. [8] R. ParreIra, C. Vaz Pinto, Tesouro da Arqueologia Portuguesa no Museu Nacional de Arqueologia e Etnologia, Secretaria de Estado da Cultura, Instituto Português do Património Cultural, Lisboa, 1980. [9] M. Tissot, M. Lemos, Levantamento do estado de conservação das joias arcaicas em ouro da Sala do Tesouro in A Ourivesaria pré-Histórica do Ocidente Peninsular Atlântico Compreender para Conservar (M.F. Guerra, I. Tissot eds.) AuCORRE, Lisboa (2014) pp. 86–89. [10] I.V. Aoki, T.C. Diamantino, M. J. F. Marques, I.F. Vasques, M.E.D. Taqueda, O climatograma de Lisboa segundo uma norma ETS, Corros. Prot. Mater. 26 (2007) 40‒45. [11] T.T. Schaeffer, Effects of light on materials in collections: Data on photoflash and related sources, Research in Conservation, Getty Conservation Institute, Los Angeles 2001. [12] P. Dellaportas, E. Papageorgiou, G. Panagiaris, Museums factors affecting the ageing process of organic materials: review on experimental designs and the INVENVORG project as a pilot study, Heritage Science 2 (2014) 2‒13. [13] J. Tétreault, Display Materials: The Good, the Bad and the Ugly, Scottish Society for Conservation and Restoration (SSCR) Edinburgh, 1994. [14] P.B. Hatchfield, Testing for pollutants in Pollutants in the Museum Environment: Practical strategies for problem solving in design, exhibition and storage, Archetype, London (2002) pp.43‒54. [15] D. Camuffo, R. Van Grieken, H.-J. Busse, G. Sturaro, A. Valentino, A. Bernardi, N. Blades, D. Shooter, K. Gysels, F. Deutsch, M. Wieser, O. Kim, U. Ulrych, Environmental monitoring in four European museums, Atmospheric Environment 35 (2001) S127‒S140. [16] J. Tétreault, E. Cano, M. Van Boomel, D. Scott, M.-G. Barthés-Labrousse, L. Minel, L. Robbiola, Corrosion of copper and lead by formaldehyde, formic and acetic acid vapours, Studies in Conservation 48 (2003) 231‒250. [17] D. Stirling, G. Noble, N. Tennent, R. Kennedy, Sulfur: from industry to museum, SSCR Journal 11 (2000) 12‒13. [18] P. Brimblecome, The composition of museum atmospheres, Atmospheric Environment 24B (1989) 1‒8. [19] D.M. Bastidas, E. Cano, A.G. González, S. Fajardo, R. Lleras-Pérez, E. Campo-Montero, F.J. Belzunce-Varela, J.M. Bastidas, An XPS study of tarnishing of a gold mask from a pre-Columbian culture, Corros. Sci. 50 (2008) 1785‒1788. [20] J.A. Bamberger, E.G. Howe, G. Wheeler, A variant Oddy test procedure for evaluating materials used in storage and display cases, Studies in Conservation 44 (1999) 86‒90.


73

[21] M.J. Samide, G. D. Smith, Analysis and quantification of volatile organic compounds emitted from plastic used in museum construction by evolved gas analysis-gas chromatography-mass spectrometry, J. Chromatograf. A 1426 (2015) 201‒208. [22] H. Lin, G.S. Frankel, W.H. Abbott, Analysis of Ag corrosion products, Journal of the electrochemical society 160 (2013) C345‒C355. [23] M. Tissot, M. Lemos, M. Dubus, Conservação preventiva – Sala do Tesouro do Museu Nacional de Arqueologia in A Ourivesaria pré-Histórica do Ocidente Peninsular Atlântico Compreender para Conservar (M.F. Guerra, I. Tissot eds.) AuCORRE, Lisboa (2014) pp. 80–85. [24] J.P. Randin, P. Ramoni, J.P Renaud, Tarnishing of the AuAgCu alloys. Effect of the composition, Werkst. Korros. 43 (1992) 115–123. [25] Y. Yoon, J.D. Angel, D.C. Hansen, Atmospheric corrosion of silver in outdoor environments and modified accelerated corrosion chambers, Corrosion 72 (2016) 1424‒1432. [26] Y. Brunhammer (cord.), René Lalique : bijoux d'exception, 1890-1912, Musée du Luxembourg, Paris, France, 2007. [27] H. Vever, French jewelry of the nineteenth, K. Purcell (trad.), Thames and Hudson, London, 2001. [28] S. Barten, Materials and techniques in the jewelry of René Lalique in The jewels of Lalique (Y. Brunhammer ed.) Flammarion, New York, 1998. [29] A.S. McDonald, G.H Sistare, The metallurgy of some carat gold jewellery alloys. Part I – coloured gold alloys, Gold Bull. 11 (1978) 128–131. [30] M.A. Grande, D. Ugues, D. Pezzini, Controls and quality demands in the jewellery investment casting process, Metalli Preziosi 1 (2004) 55‒60. [31] K. Lee, J.T. Lue, Annealing effect on the optical properties of gold films deposited on silicone substrates, Appl. Optics, 27 (1988) 1210–1213. [32] J.P. Randin, Electrochemical assessment of the tarnish resistance of decorative gold alloys, Surf. Coat. Tech. 34 (1988) 253‒275. [33] P. Homem, I. Fonseca, J. Cavalheiro, O escurecimento do altar de prata da Sé do Porto: um caso de corrosão atmosférica, Corros. Prot. Mater. 27 (2008) 82‒86. [34] A. Lucas, Ancient Egyptian Materials and Industries, Edward Arnold Publishers Ltd., 1926. [35] J.H. Frantz, D. Schorsch, Egyptian red gold, Archeomaterials 4 (1990) 133–152. [36] M.M Rifai, N.M.N El Hadidi, Investigation and analysis of three gilded wood samples from the tomb of Tutankhamun in Decorated Surfaces on Ancient Egyptian Objects, Technology, Deterioration and Conservation (J. Dawson, C. Rozeik, M.M. Wright dds.) Archetype Publications, London (2010) pp.16‒24. [37] L. Troalen, J. Tate, M.F. Guerra, Goldwork in Ancient Egypt: workshop practices at Querneh in the 2nd Intermediate Period, J. Archaeol. Sci. 50 (2014) 219‒226. [38] S.R. Snape, Mortuary assemblages from Abydos, Vol.1, unpublished PhD thesis, University of Liverpool, 1987. [39] L. Troalen, I. Tissot, M. Maitland, M.F. Guerra, Jewellery of a young Egyptian girl: Middle Kingdom gold work from Haraga tomb 72, Historical Metallurgy 49 (2015) 75‒62.


74

Chapter 4

Characterisation of corroded binary and ternary

alloys in sulphide environments

Chapter 4 – Characterisation of corroded binary and ternary alloys in sulphide environments

77

4. Characterisation of corroded binary and ternary alloys in

sulphide environments

4.1 Introduction

As pointed out in Chapter 3, in the field of cultural heritage the development of an analytical

protocol for the study of corroded gold objects is, due to security issues, quite restricted

regarding moving the objects to laboratories. To overcome this situation, it is possible to

fabricate samples from the same gold alloys and with the same surface finishing to be subjected

to corrosion in natural or artificial environments. There are other advantages to the use of

samples, since it becomes possible to isolate the parameters that influence the corrosion

development and study their effect one by one: like the composition of the alloys, the corrosion

environment and the time of corrosion. In this way, the detailed study of the corrosion products

that grow on the gold samples can be achieved and the results extrapolated to describe the

phenomena occurring for corroded gold objects.

To gather further information on the corrosion of gold alloys in atmospheric environments, it

was decided to approach this question by making samples whose characteristics were chosen

based on data obtained for the three case studies described in Chapter 3. The main questions

to be disclosed, pointed out in the conclusions of Chapter 3, are:

i) to determine the role of the constitutive elements on the corrosion of gold alloys. It

was verified that the silver content of the gold alloy influences its corrosion process,

but the role of copper could not be assessed.

ii) to determine the influence of the corrosion products morphology on the colour of

the corroded surfaces. It was proposed that besides the layer thickness, the

presence of corrosion products with different morphologies could influence the

colour of the corroded surface.

To approach these issues, 12 samples made from gold alloys with distinct elemental

compositions were fabricated and subjected to accelerated corrosion tests. In order to search

for the role of each element of the alloy and to facilitate the development of the analytical

protocol for the study of gold alloys, 11 samples made from silver, copper and silver-copper

alloys were also fabricated and subjected to accelerated corrosion at the same conditions. The

silver-copper alloy chosen for this study was sterling silver, an alloy commonly used in jewellery

and silverware fabrication composed of 92.5 wt% Ag and 7.5 wt% Cu.


78

The silver, the copper and the gold samples were acquired from a local jewellery supplier. Those

made from binary and ternary alloys were fabricated by a goldsmith by melting the metals. The

surface finish achieved by the goldsmith reproduces the routine surface roughness of gold

objects. The samples were then artificially corroded by immersion in a sulphide containing

solution. The corrosive atmosphere was chosen based on the three case studies of Chapter 3:

sulphur is the expected pollutant for indoor environments, as shown in the case of the Treasure

room study, sulphur-based compounds were identified as the main corrosion products of the

corroded objects analysed.

A corroded surface displays colours that change with the corrosion time increase [1]. The

duration of the corrosion tests was therefore set in function of the surface colours observed,

particularly on the Egyptian foils.

Another parameter that was taken into account was the corrosion rate of the gold alloys, which

is slower than those of silver, copper and silver-copper alloys [2]. The methodological approach

was based on the characterisation of the different colours of the corroded surfaces of silver,

copper and sterling silver, obtained in short-time corrosion tests, to be compared to those

obtained for gold alloys corroded during longer times and in a less concentrated corrosion

solution. Then, the gold alloy which presented the most accentuated colour change was studied

under the same analytical conditions as the silver, copper and sterling silver corroded samples

using the analytical techniques described in Chapter 2. All the samples were characterised

before and after the corrosion tests.


79

4.2 Sample preparation

4.2.1 Silver, copper and sterling silver samples

A set of 11 samples of Ag, Cu and sterling silver were cut from standard sheets acquired from a

local jewellery supplier (Lima & Teixeira, Lisbon). The samples, with 1.5 cm2 area and 1 mm

thickness, were grinded with SiC-paper of different grits up to 1000 mesh. To remove grinding

residues, the samples were washed with deionised water in an ultrasonic bath. It was not

intended to replicate a perfect grinded surface. The objective of this surface finishing was to

obtain a rough surface similar to those found on objects in the field of cultural heritage.

The elemental analysis of the Ag, Cu and sterling silver samples was carried out by μXRF before

the corrosion test, to check the surface homogeneity and to prove the elemental composition.

The compositions obtained in twelve areas show the samples homogeneity for the metals and

the alloy, despite an Ag content 0.5 wt% higher than expected for sterling silver (table 4.1),

justified by the strict requirements to obtain the legal hallmark1.

Table 4.1 – Elemental composition obtained by µXRF, normalised to 100 % of the Ag, Cu and sterling silver samples used in this work.

4.2.3 Gold alloy samples

A set of 12 samples of Au alloys was fabricated for this work. After melting together the sheets

of single metals, the alloy was poured and then laminated to a 0.3 mm thickness sheet to be cut

into samples with a 0.5 cm2 area. Table 4.2 shows the elemental composition of the eight ternary

gold alloys (Au-Ag-Cu) and of the four binary gold alloys (Au-Ag and Au-Cu) obtained by μXRF.

1 The legal hallmark states that sterling silver should have at least 925 parts of Ag. Consequently, the jewellery suppliers slightly increase the Ag content to assure that the final Ag alloy has the required 92.5 wt%.

Composition in wt% Ag Cu kα kα

Samples Ag 100.0 ± 0.0 - Cu - 100.0 ± 0.0 Sterling silver 93.1 ± 0.1 6.9 ± 0.1


80

The Au alloys samples were fabricated by a goldsmith with a finish that reproduces as much as

possible the normal handcraft work.

Table 4.2 – Elemental composition obtained by μXRF, normalised to 100 % of the binary (B) and ternary (T) Au alloys fabricated for this work.

4.3 Corrosion test method and conditions

Several corrosion tests to reproduce naturally corroded Ag and Au alloys have been developed

during the last decades [3]. In general, they are divided in gaseous or liquid phase methods. The

gaseous methods are sub-divided in exposure to gas or vapours tests and flowing gas tests, being

the latter the most commonly used [3]. In this method, air containing from 20 µg.g-1 down to

0.1 µg.g-1 of one or more gases is passed at a known flow rate, temperature and humidity over

the samples [3,4]. The most used gases are sulphur containing pollutants such as H2S, and SO2

[5,6]. The liquid phase methods are less used, being the most widely applied the “Tuccillo-

Nielsen Wheel” developed in 1971 for the study of corroded gold alloys used in dentistry [7].

This test is carried out using 0.1 M to 0.26 M Na2S aqueous solutions [6,8], with a rotating

apparatus where the samples are repetitively dipped in the solution and removed to open air

Composition in wt% Au Ag Cu Lα Kα Kα

Au alloy ID B1 99.1 ± 0.1 0.1 ± 0.1 0.8 ± 0.1 B2 98.1 ± 0.2 0.1 ± 0.1 1.8 ± 0.1 B3 94.5 ± 0.2 5.4 ± 0.1 0.1 ± 0.1 B4 89.4 ± 0.4 10.5 ± 0.4 0.1 ± 0.1 T1 94.9 ± 0.2 2.8 ± 0.1 2.3 ± 0.1 T2 89.7 ± 0.5 9.6 ± 0.5 0.7 ± 0.0 T3 88.3 ± 0.2 6.7 ± 0.1 5.0 ± 0.3 T4 82.3 ± 0.6 15.6 ± 0.7 2.1 ± 0.1 T5 80.6 ± 0.1 11.9 ± 0.2 7.5 ± 0.1 T6 75.9 ± 0.3 16.9 ± 0.2 7.2 ± 0.3 T7 73.6 ± 0.2 22.9 ± 0.2 3.5 ± 0.1 T8 71.1 ± 1.0 24.5 ± 1.0 4.4 ± 0.2


81

[3]. Other methods developed consist in the immersion of the sample in an opened or in a closed

container, with or without forced convection and controlled temperature [3].

Different accelerated corrosion methods were compared for long-time duration tests. This

comparison showed that the liquid-phase methods carried out in open-air condition lead, due

to air oxidation, to the reduction of the sulphide [8] and to the creation of galvanic effects, which

do not occur in the gaseous-phase methods [6]. However, the corrosion rate in liquid-phase

methods is a linear function of time, and the duration of the accelerated tests required to reach

a colour change similar to those of naturally corroded gold samples is much lower than for

gaseous-phase tests, in a proportion of 1 to 40 accordingly to Randin [6]. Although accelerated

corrosion tests in liquid-phase are not fully representative of the naturally corroded surfaces,

these easily implemented methods should give, considering the above-mentioned, useful

information when short-time experiments are carried out. For this work, it was chosen to carry

out the corrosion study using liquid-phase methods in Na2S aqueous solutions.

The accelerated corrosion tests of Ag, Cu and sterling silver samples were carried out in an open

container; the samples were immersed in a 0.1 M Na2S aqueous solution for different times (1

to 7 min, and then 10, 15, 30, 60, 120, 240, 480 and 1020 min). The replicating conditions were,

for all the experiments, close to those corresponding to a normal atmosphere [9], i.e., solutions

kept in equilibrium with the normal atmosphere (no forced convection) at room temperature

(22 ± 2)°C. Under these conditions, the oxygen concentration in the oxidising medium is about

2.9x10-4 M [10]. The local oxygen concentration is maintained only by the diffusion of the

dissolved gas through the electrolytic medium. Several replicate corrosion tests were done

under the same conditions to check whether the surface changes for each immersion time were

visually similar.

The gold alloys samples were immersed for 12 months in a closed container with a 50 mM Na2S

aqueous solution. After 30 days of immersion, the samples were removed for a first

characterisation and then re-immersed again in the containers. The containers were kept at

room temperature (22 ± 5°C) and light-protected. Alloy T6 was also subjected to accelerated

corrosion by immersion circa 35 hours in an open container with a 0.1 M Na2S solution, at the

same conditions as the Ag, Cu and sterling silver samples.

After immersion, samples were removed, washed with deionised water, and dried in an N2

stream. The corroded samples characterisation was performed ex-situ.


82

4.4 Corroded surface characterisation

4.4.1 Silver

Artificially aged Ag surfaces present colour variations that depend on the immersion time. Figure

4.1 shows the colour variation that goes from light brown (1 and 3 min) to blue (5 and 7 min),

blue-grey (15, 30, 60 and 120 min) and grey (240 min onwards). Contrary to what has been

suggested, the colours of the corroded Ag surfaces are different from those reported for

corroded sterling silver [1].

Figure 4.1 – Surface colour of the corroded Ag samples immersed during 1, 3, 5, 7, 15, 30, 60, 120, 240, 480 and 1020 min in a 0.1M Na2S aqueous solution.

Figure 4.2A presents the UV-Vis absorption spectra for the samples immersed during 1, 3, 5 and

7 min. All these samples show an absorption band increase around 370 nm that can be ascribed

to the presence of Ag nanoparticles [11]. The sample immersed during 1 min has a broad band

with a maximum absorption at 484 nm, which can be attributed to the nucleation of amorphous


83

Ag2S with a particle size inferior to 10 nm [12]. The sample immersed during 3 min has a well-

defined absorption band with a maximum at 528 nm, corresponding to an estimated band gap

energy of 1.2 eV, value reported for the Ag2O [13].

As shown in figure 4.2B, the UV-Vis absorption spectra of the samples with immersion times

longer than 15 min have a broad band between 370 and 480 nm, similar to the one observed

for the sample immersed during 1 min. The UV-Vis spectra of samples immersed during 5, 7 and

15 minutes show another band, which broadens with the immersion time increase and shifts

towards higher wavelengths with a maximum absorbance at 646 nm (5 min), 689 nm (7 min)

and 820 nm (15 min), respectively. This band can be ascribed to the formation of Ag2S particles,

whose size increases with the immersion time [12]. The band energy gap estimated for the 7

min sample is 1.1 eV that could correspond to the presence of Ag2S particles [14].

The UV-Vis spectra of the samples immersed during 60 and 120 minutes show a slight bump at

566 nm followed by a wide band with a maximum absorbance at 679 nm (60 min) and 839 nm

(120 min). At higher wavelengths, it is observed a slight increase with a maximum absorbance

at 1141 nm (60 min) and at 1123 nm (120 min). Figure 4.2C shows the data obtained for the

samples with immersion times of 240, 480 and 1020 minutes. The absorption spectrum of the

240 min sample has a small bump at 569 nm followed by a wide band with an absorbance

decrease at around 1000 nm. The sample immersed during 480 min has a wide band with a

maximum of absorbance at 708 nm. The 1020 min sample has a broad band with a slight bump

at 575 nm followed by an absorbance decrease and then by another accentuated decrease at

1150 nm. The bands broadening can be related to the corrosion layer increase.

The UV-Vis absorption data suggests the coexistence of Ag2S and Ag2O particles with distinct

sizes. When the samples are immersed during short periods of time (1 to 3 minutes), there is a

possible nucleation of small particles which size increases with the immersion time, followed by

the nucleation of new particles in a continuous corrosion process.


84

Figure 4.2 - UV-Vis absorption spectra obtained for the Ag samples corroded during 1, 3, 5 and 7 min (A); 15, 30, 60 and 120 min (B); 240, 480 and 1020 min (C) in a 0.1M Na2S aqueous solution.

The surface morphology and the elemental composition of the compounds formed during the

corrosion process were investigated by SEM-EDS. Figure 4.3A shows the bare Ag surface before

accelerated corrosion. The grid marks are visible on the surfaces of samples immersed until 120


85

min. The surface of the samples immersed during 1, 3 and 5 minutes are characterised by the

presence of small near-spherical particles (figures 4.3B and 4.3C). The 7 and 15 minutes

corroded surfaces present a uniform layer, composed of spherical particles whose sizes increase

with time (figures 4.3D and 4.3E). The layer has localised hollows resulting from the sulphidation

process [15]. The corrosion layer of the sample immersed during 30 min, figure 4.3F, appears

thicker and containing particles with well-defined boundaries. From 240 min onwards, the shape

of the formed structures resembles crystallites. Figure 4.3H shows Ag protrusions caused by the

electron beam acceleration voltage as reported by Morales-Masis et al. [15].

Figure 4.3 - SEM images of Ag (A) and corroded Ag by immersion in a 0.1 M Na2S aqueous solution during 3 min (B), 5 min (C), 7 min (D), 15 min (E), 30 min (F), 60 min (G), 240 min (H) and 1020 min (I) (scale bar is 2 µm for A to H and 5 µm for I).

Table 4.3 summarises the EDS data obtained for Ag corroded samples. With the immersion time

increase, the Ag content decreases and the S and O values increase. The increase of S content is

more pronounced than the increase of O. Keast et al. [16] attribute the increase of S to an Ag2S

corrosion layer thickening. For samples immersed more than 120 min, the Ag/S ratio in at. % is

consistent with the presence of Ag2S.


86

Table 4.3 - Elemental composition by EDS (wt%), normalised to 100 %, of Ag corroded samples.

XRD analysis allowed identifying α-Ag2S for all the samples except for the one immersed during

1 min. As shown in figure 4.4, the intensity of the different Ag2S diffraction peaks increases with

the immersion time, which can be justified either by an increase of the compound quantity or

by an increase of the degree of crystallisation of the corrosion products.

Figure 4.4 - X-ray diffractograms obtained for the 1, 3, 5, 15, 60, 240 and 1020 min corroded pure Ag. The diffraction peaks correspond to Ag2S ( PCD file nº00-014-0072).

Composition in wt%

Ag S O

Lα Kα Kα

Sample ID

1 min 97.7 1.5 0.8 3 min 97.7 1.2 1.2 5 min 97.6 1.5 0.6 7 min 97.3 1.5 0.9 15 min 97.8 1.5 0.7 30 min 96.7 3.0 0.3 60 min 91.2 7.2 1.6 120 min 88.2 10.7 1.2 240 min 88.3 10.4 1.3 480 min 87.2 11.5 1.3 1020 min 86.9 11.8 1.3


87

For immersion times inferior to 240 min, the strongest diffraction peaks are distinct from those

reported for α-Ag2S (PCD file nº 00-014-0072). Figure 4.5 shows the diffractogram obtained for

the 240 min sample. The highest intensity peak corresponds to the (103) diffraction plane of

Ag2S instead of the expected (121). This difference can be justified by a distinct crystal

orientation acquired during the growing process [15]. The presence of crystalline Ag2O was not

identified by XRD.

Figure 4.5 - X-ray diffractogram obtained for the 240 min corroded Ag sample with the identification of the main diffraction peaks of Ag and Ag2S. Ag (PCD file nº 00-004-0783); Ag2S (PCD file nº 00-014-0072).

The Ag corrosion layer thicknesses were assessed by ex-situ ellipsometry. For each sample, the

ellipsometric parameters, azimuthal angle (Ψ) and phase shift (Δ) were measured at 65° and 70°

and the data fitted with a two-layer model in a multiangle approach. This model considers one

homogeneous and isotropic layer on a semi-infinite homogeneous substrate and allows the

estimation of the thickness of the corrosion layer, being known the optical properties of the

substrate. The obtained dielectric constants, refractive index (n) and extinction coefficient (k)

are effective values (in the restrict sense of the optical outcome) and not necessarily

characteristic of a single compound. Table 4.4 presents the complex refractive index and the

computed thickness of the corrosion films for the samples immersed until 30 min. The thickness


88

values obtained are higher than those reported for Ag2S layers formed in natural environments

[17,18] and by artificial corrosion [19,20].

Table 4.4 – Complex refractive index (ñ) of the corrosion layer formed on Ag samples and their thicknesses (d).

4.4.2 Copper

Figure 4.6 shows the colour variation of the Cu corroded surfaces with the immersion time. Alike

the colour variation reported for pure Cu corroded in Na2S solutions [21,22], the surface of

samples change from violet (1 min) to blue-green (3 min), green (5 and 7 min), violet (15 min),

dark violet (30 min), blue-grey (60 to 240 min) and then grey-black (480 and 1020 min). A violet

tone is observed for samples immersed during 1 and 15 minutes while those immersed for 3, 5

and 7 minutes have green tones.

The UV-Vis absorption spectrum shown in figure 4.7, obtained for the sample corroded during

1 min, has a well-defined band. This band has a maximum absorbance at 560 nm corresponding

to a 1.78 eV band gap energy, which can be ascribed to the presence of Cu2O [23] or to the

presence of CuxS particles with x=1, i.e. covellite [24,25]. For the sample immersed during 3 min,

two bands are observed: one in the same wavelength range as in the previous sample, but with

lower intensity, and a well-defined band with a maximum at 670 nm. The latter band may

correspond either to an increase of the particles size or to the formation of another

stoichiometric CuxS compound, with 1<X<2, which is supported by the calculated band gap

energy at 1.4 eV [24,25]. For samples immersed during 5 and 7 minutes, the second band shifts

towards higher wavelengths attaining a maximum of absorbance at 856 nm and 848 nm,

respectively. These bands correspond to 1.1 eV and 1.2 eV band gap energies, the latter value is

reported for the formation of both Cu2S [25] and Cu2O particles [26].

Sample ID ñ d (nm)

1 min 3.0 - 0.1i 118 ± 0.6 3 min 3.0 - 0.1i 126 ± 1.3 5 min 3.0 - 0.4i 127 ± 2.6 7 min 3.0 - 0.4i 129 ± 0.5 15 min 3.0 - 0.4i 214 ± 0.5 30 min 3.0 - 0.3i 222 ± 0.4


89

Figure 4.6 – Surface colour of the corroded Cu samples immersed during 1, 3, 5, 7, 15, 30, 60, 120, 240, 480 and 1020 min in a 0.1M Na2S aqueous solution.

This band shifts to lower wavelengths for the sample immersed during 15 min. The maximum

absorbance reached, around 560 nm, is identical to that of the sample immersed during 1 min.

This shift can be related to the nucleation of new particles of distinct size or stoichiometry, but

similar to those formed on the 1 min sample surface. Since the optical properties also depend

on the size and shape of the formed particles [25], this result supports the colour similarity cited

above. For the samples immersed during 30 min onwards, the bands broadening should be

related to the corrosion layer thickening.


90

Figure 4.7 - UV-Vis absorption spectra for the Cu samples corroded during 1, 3, 5 and 7 minutes (A); 15, 30 and 60 min (B); 120, 240 and 1020 min (C) in a 0.1M Na2S aqueous solution.

SEM observation of the sample immersed during 3 min (figure 4.8B) revealed the presence of a

uniform layer composed of small round particles. The layer morphology reminds the one

reported for Cu2S films [22]. The corrosion layer thickens with the immersion time increase and,


91

from an immersion time of 5 min onwards, it covers the grit marks (figure 4.8C). Round particles

of different sizes can be seen in the corrosion layer developed on the sample immersed 15 min

(figure 4.8E). For longer immersion times, the particles change to an undefined shape that is

characteristic of the layer growth mechanism [27-29]. On samples immersed during times longer

than 60 min, the particles have either a round or a laminar shape (figures 4.8G – 4.8I). The latter

can be ascribed to Cu2S [28-30].

Figure 4.8 - SEM images of Cu (A) and corroded Cu by immersion in a 0.1 M Na2S aqueous solution during 3 min (B), 5 min (C), 7 min (D), 15 min (E), 30 min (F), 60 min (G), 240 min (H) and 1020 min (I). (scale bar is 2 µm for A to H and 5 µm for I).

Table 4.5 summarises the EDS data obtained for the corroded samples. With the immersion time

increase, the S and O contents increase and the Cu content decreases. The accentuated increase

of S on the 60 min sample surface can be related to the formation of corrosion products with

sulphur. The increase of O is more pronounced in the case of samples immersed during 240 and

1020 min, suggesting the formation of Cu2O.


92

The presence of Cu2S was identified by XRD for all the samples with immersion times longer

than 15 min. The presence of CuS expected from the UV-Vis results could not be confirmed by

XRD, probably due to the small amount in the samples.

Table 4.5 - Elemental composition by EDS (wt%), normalised to 100 %, of Cu corroded samples.

As shown in figure 4.9, Cu2O was identified by XRD for the 120, 240 and 1020 min samples,

suggesting the formation of two different compounds for immersion times longer than 60 min.

The highest intensity peaks of Cu2S are different from those reported in the PCD file nº00-033-

0490 showing that it grows with a distinct crystal orientation. The formation of Cu2O could result

from the Cu2S oxidation. With the corrosion layer thickening, cracks and defects appear,

allowing S to react with Cu, to form Cu2S.

Composition in wt%

Cu S O

Kα Kα Kα

Sample ID

1 min 97.3 1.2 1.5 3 min 96.7 1.5 1.8 5 min 93.6 4.9 1.5 7 min 94.5 3.5 2.1 15 min 95.1 3.2 1.6 30 min 92.1 5.5 2.4 60 min 82.8 14.8 2.4 120 min 88.4 8.0 3.6 240 min 87.7 4.6 7.7 480 min 86.3 10.2 3.4 1020 min 83.8 7.1 9.1


93

Figure 4.9 - X-ray diffractograms obtained for the samples immersed 60, 120, 240 and 1020 min with the identification of main diffraction peaks of Cu2S and Cu2O. Cu2S (PCD file nº00-033-0490); Cu2O (PCD file nº00-077-0199).

The corrosion layer thickness was also assessed for the Cu samples by ex-situ ellipsometry using

a multiangle analysis (70° and 60°) and the two-layer model. Table 4.6 presents the complex

refractive index and the estimated corrosion film thickness for the samples corroded until 60

min. The values obtained for the sample immersed during 1 min are in accordance with those

reported for Cu2S films with violet colour surface [22].

Table 4.6 – Complex refractive index (ñ) of the corrosion layer formed on Cu samples and their thicknesses (d).


94

Other information, such as the corrosion layer structural composition, can be retrieved from the

dielectric constants. The composition was estimated by modelling the corrosion film structure

using the Bruggeman Effective Medium Approximation (BEMA) theory [31]. In this model, the

corrosion layer structure is considered as a mixture of known compounds [32]. Based on the

elemental composition obtained by EDS (table 4.5) that identified the presence of S and O, the

corrosion layer could be modelled as a combination of Cu2S and Cu2O, using reported n and k

values [33,34]. For samples corroded until 5 min, a mixture of Cu2O/Cu2S with relative

proportions of 85±5/15±5 vol% was found. For longer immersion times (until 60 min), it prevails

the Cu2S component with a concentration ranging from 70 to 80 vol%. The latter result is in

agreement with the XRD data.

The 15 min sample could not be fitted by the BEMA model. The similar violet surface colour and

the similar UV-Vis spectra profiles for samples corroded during 1 and 15 minutes (figure 4.6)

suggested the presence of similar species on the surface of these samples. However, the

estimated corrosion layer for the 15 min sample is, as expected, higher than the thickness of the

1 min corrosion layer. For this reason, the existence of a multi-layered corrosion film was

considered, and data was fitted using the three-layer model. The results of this theoretical

model indicate the presence of a 160 nm inner-layer composed of 80 vol% of Cu2O and 20 vol%

Cu2S and a 230 nm outer layer composed exclusively of Cu2O. These values are in agreement

with the calculated corrosion layer thickness (332 nm by the multiangle ellipsometry approach)

and with the UV-Vis data.

4.4.3 Sterling silver

Figure 4.10 displays the colour of the sterling silver surfaces after corrosion. With the immersion

time increase the sterling silver samples colour change from light yellow (1-2 min) to orange (3

min), red (5 min), and then to violet (7 min), blue- grey (15 min), blue (30 min) and grey (60 min).

From 60 minutes onwards the surface maintains the grey colour becoming darker as the

immersion time increases. When the colour of pure Ag is compared with the colour of sterling

silver, it can be seen that in the early stages of corrosion the colour variation of pure Cu and

sterling silver are alike, while for longer immersion times the colour variation of sterling silver

approaches that of pure Ag. This suggests for each corrosion stage the formation of similar

species for the sterling silver and pure metals.


95

Figure 4.10 - Colour evolution of the corroded sterling silver samples for different immersion times (1, 3, 5, 7, 15, 30, 60, 120, 240, 480 and 1020 min) in a 0.1M Na2S aqueous solution.

Figure 4.11 shows the UV-Vis absorption spectra of the samples obtained after immersion times

of 1, 3, 5 and 7 min. The data indicate that the colour variation is associated with the corrosion

development. It is possible to establish several groups depending on the immersion times used.

For all the samples, an increase on the absorption bands starting at about 400 nm is observed.

The samples immersed during 1 and 3 min, which had developed a yellow to yellow-orange

corroded layer, are characterised by an absorption band with a maximum absorbance around

342 nm, for the sample immersed 1 min, and 500 nm, for the sample immersed during 3 min.

These two UV-Vis spectra profiles suggest the formation of identical species on the surface of

these samples. The red shift of the 342 nm band for the sample immersed 3 min can suggest an

increase in size of the formed species [11,35].

The UV-Vis spectra obtained for the samples immersed for 5 and 7 min, which display a dark red

and a light blue surface colour, show two absorption bands at 430 nm and at 800 nm for the 5

min sample, and at 430 nm and 860 nm for the 7 min sample. The presence of these two bands,

which are distinct from those observed for the 1 and 3 min samples, could be caused by a


96

difference on the Ag/Cu ratio [11]. This may be explained by the formation of distinct Cu- and

Ag-based corrosion products.

The UV-Vis spectra obtained for the samples immersed for 30 and 60 min, shown in figure 4.11B,

present one absorption band at 642 nm for the sample immersed 30 min and at 584 nm for the

sample immersed 60 min. These absorption bands shift towards shorter wavelengths getting

closer to the ones measured for the samples immersed 1 and 3 min. This phenomenon may be

caused by the formation of new corrosion products that are similar to those observed for the

first corrosion stage. The UV-Vis spectrum obtained for the sample immersed for 15 min, which

has a dark blue grey colour, shows an absorption band with a maximum around 560 nm. Figure

4.11C present the UV-Vis spectra obtained for the samples immersed 120, 240, 480 and 1020

min. The 120 min sample shows an absorbance increase at 330 nm with a maximum around 720

nm. This band broadens with the immersion time increase, which can be related to the corrosion

layer thickness. The sample with the highest immersion time, 1020 min, absorbs radiation in the

400 -1180 nm range.

In order to confirm the presence of Ag2S on the samples surfaces, which should be the main

corrosion product according to the literature, the band gap energy of this compound was

calculated for all the UV-Vis spectra. According to Chattopadhyay & Roy [14] depending on the

particle size, the band gap energy for Ag2S ranges from 0.72 to 1.1 eV. For the 7 min sample, the

energy band gap was estimated to 0.9 eV, close to the published value for Ag2S thin films (0.96

eV) [14]. However, it must be noted that this value is very close to the band gap energy reported

for AgCuS [36]. Contrary to what was expected [35], it was not possible, from the band gap

energy, to confirm the presence of Ag2S in the samples immersed during 1 and 3 min.

Nevertheless, by comparison with published UV-Vis spectra, the two absorption bands (318 nm

and 506 nm) of the 3 min sample can be attributed to the formation of Cu2O [37].

Although it was not, therefore, possible to identify the corrosion products by UV-Vis

spectrophotometry, the results point to a relation between the observed colours and the

compounds formed during the different corrosion stages.


97

Figure 4.11 - UV-Vis absorption spectra for the sterling silver samples corroded during 1, 3, 5 and 7 min (A); 15, 30 and 60 min (B); 120, 240, 480 and 1020 min (C) in a 0.1M Na2S aqueous solution.

In order to evaluate the corrosion layer thickness and to obtain information on the dielectric

constants changes across the film, the artificially aged sterling silver samples were analysed by


98

ex-situ spectroscopic ellipsometry. The highly absorber corrosion layer hindered the analysis of

samples corroded with immersion times higher than 60 min, as the reflected radiation intensity

was too low to be detected.

Figure 4.12 presents the experimental values obtained for the Ψ and Δ of the samples. In order

to estimate the formed film thickness, a regression analysis was applied. The experimental data

were fitted to simulate Ψ vs. Δ values generated for a theoretical two-layer model that

approaches the sample structure. The complex index of refraction (ñ=n-ik, n being the real part

of the refractive index and k the extinction coefficient) of the corrosion product, as well as its

thickness, are computed based on the best fit of the theoretical and experimental data [38]. The

two-layer model (one homogeneous layer coating on a semi-infinite homogeneous substrate)

employed to describe the samples was found to be appropriate to evaluate n, k and the

thickness of the samples exposed to the corroding medium for up to 3 minutes. It is evident

from figure 4.12 that there is a change of the “snail-shape” behaviour of the Δ vs. Ψ evolution

characteristic of a film thickening, after the 3 min sample. This observation must correspond to

the development of a new phase with distinct properties. Specimens with higher immersion

times in Na2S solution will require more complex physical models to be employed, but the

correct definition of such a model cannot be deduced from discrete ex-situ measurements.

Indeed, the observed differences of the corrosion layers suggest that they may consist of

combinations of several chemical compounds, whose proportions may change throughout the

film thickness, and also from one sample to another.

Figure 4.12 - Experimental values of Ψ vs. Δ collected by ellipsometry for sterling silver samples corroded until 60 min immersion time. The angle of incidence was fixed at 70 ⁰.


99

Table 4.7 presents the complex refractive index and the estimated thickness of the corrosion

films for the samples immersed from 1 to 3 min. It is worth noting the feasibility of the multiangle

analysis approach to compute the ellipsometric data assuming the formation of a uniform

deposit. This is only possible for isotropic layers, which is another characteristic of the very thin

corrosion films studied in this work (at least for up to 3 min immersion time). The values

obtained are within the range of those reported for silver alloys corrosion layers containing Ag2S

[39,40]. It should be emphasised that the small thickness of the corrosion layer grown at the

early stages of artificial ageing already presents an accentuated colour alteration. Contrary to

what has been reported [40], this work indicates that a 100 nm thick film of corroded silver is

not black; the sample aged for 3 min is orange, and the corroded layer is 166 nm thick.

Table 4.7 - Complex refractive index (ñ) of the corroded sterling silver alloy samples after different immersion times and calculated layer thickness (d).

All the samples were studied by SEM-EDS to check whether the colour evolution can be

associated with morphological changes of the species (size of the particles) formed over time or

to surface composition changes (modification of the species formed).

It must be noted that sterling silver samples were abraded only to a level corresponding to a

used-wear jewellery. This explains the sharp tool marks on the surface even under low

magnification (figure 4.13A). No additional grinding was carried out in order to keep the samples

in conditions as similar as possible to ancient objects. For this reason, few morphological

differences could be observed for each sample.

In spite of the tool marks on the surface of the samples artificially aged for 1 to 5 min, a few

remaining materials from the grinding process could still be found, particularly in rougher areas.

Under observation with adequate magnification, it is possible to distinguish, on the rough

surface close to the tool marks, the nucleation of the corrosion products (figure 4.13D).

Moreover, the nucleation of corrosion products inside some surface scratches has already been

reported for the formation of AgCl on silver wires, where it is stated that the scratched bottom

Sample ID ñ d (nm) 1 min 1.8-0i 34 2 min 1.8-0.5i 41 3 min 0.6-0.8i 166


100

facilitates heterogeneous nucleation [41]. Figure 4.13E shows a particles size increase on the

samples until the seventh minute of immersion. After this period, one porous film is observed

on the surfaces (figure 4.13F). This data could suggest that the corrosion mechanism is distinct

for the early stages [41].

Figure 4.13 - SEM images of the samples prepared by immersion in a 0.1 M Na2S aqueous solution at different times: 0 min, bare silver (A), 1 min (B), 3 min (C), 5 min (D), 7 min (E) and 15 min (F).

The micrographs of the corroded surfaces immersed for 60 minutes onwards (figure 4.14) show

overlapping corrosion layers with different morphologies.

The EDS analyses of the surface crystals (table 4.8) show an increase of Cu, S and O when

compared to the results obtained on the surface of samples exposed to 1 to 7 min immersion

times. The ratio Ag/S provided by EDS analysis in at.% is consistent with the presence of Ag2S.


101

Table 4.8 - Elemental composition by EDS (wt%), normalised to 100 %, of the set of artificially corroded sterling silver samples.

For Ag-Cu alloys, the formation of both Ag-Cu-S complexes and Ag- and Cu-based species is

expected. The continuous formation of a film, which grows with time and that has a layer-by-

layer morphology, is observed particularly for higher immersion times (figure 4.14B). This layer

microstructure, previously reported for silver chloride films [42] and silver sulphide films [39], is

also suggested by the UV-Vis results, which indicate the formation with time of several different

corrosion products. It should be emphasised that, although the layer-by-layer morphology is

observed from 120 min onwards, there is a size increase of the corrosion products crystals. This

increase occurs in the crystals of the top layer but also on the intermediary corrosion layers,

suggesting there is a thickening of the intermediary layers with time.

Composition in wt%

Ag Cu S O

La Ka Ka Ka Sample ID

1 min 90.1 8.5 0.5 0.9 2 min 88.5 8.0 1.8 1.7 3 min 85.6 12.6 0.7 1.1 4 min 89.5 8.1 1.3 1.2 5 min 87.6 9.0 1.9 1.4 7 min 92.1 5.3 1.2 1.4

15 min 78.4 12.6 5.8 3.3 30 min 68.5 20.4 8.7 2.4 60 min 76.3 10.2 11.0 1.4

120 min 74.7 14.5 9.7 1.1 240 min 79.0 6.7 12.8 1.6 480 min 78.9 6.7 12.8 1.7

1020 min 85.4 0.5 12.9 1.3


102

Figure 4.14 – SEM images of the samples prepared by immersion in a 0.1 M Na2S aqueous solution at: 60 min (A), 120 min (B), 240 min (C), 480 min (D) and 1020 min (E).

An attempt to determine the nature of the corrosion products formed on the surface of the

sterling silver alloys was made using XRD. Figure 4.15 shows the diffractograms obtained for

selected samples and table 4.9 presents the compounds that could be so far identified.

Crystalline Cu2O was only identified for the samples immersed until 30 minutes together with

Ag-S corrosion products, either Ag2S or AgCuS. These two latter compounds show a similar

crystallographic structure and their detection is in agreement with previous reports [43]. The

low intensity of the diffracted signal (mainly related to the thicknesses of the corrosion and to

the signal contribution of the substrate) prevents an accurate identification of these two

compounds.


103

Figure 4.15 - X-ray diffractograms obtained for the 2, 4, 5, 15 and 60 min corroded Ag-Cu alloy. ● Ag (PCD file nº 00-004-0783); Cu (PCD file nº 00 002 1225) diffraction peaks.

Table 4.9 - Crystalline elements and compounds identified by XRD on the surface of the corroded samples.

Only from 30 min onwards, it is possible to identify the presence of both Ag2S and AgCuS as

corrosion products. For longer ageing times, the presence of Ag-Cu-S complexes prevails. The

Figure 4.16 diffractogram includes the diffraction peaks of jalpaite (Ag3CuS2) and shows that

this compound is the main corrosion product of the sample aged for 60 min. EDS results are in

Sample ID Chemical Compounds 0 min Ag Cu 1 min Ag Cu Cu2O Ag2S/AgCuS 2 min Ag Cu Cu2O Ag2S/AgCuS 3 min Ag Cu Cu2O Ag2S/AgCuS 4 min Ag Cu Cu2O Ag2S/AgCuS 5 min Ag Cu Cu2O Ag2S/AgCuS 7 min Ag Cu Cu2O Ag2S/AgCuS

15 min Ag Cu Cu2O AgCuS 30 min Ag Cu Cu2O 60 min Ag Cu Ag3CuS2

120 min Ag Cu Ag3CuS2 240 min Ag Cu Ag3CuS2 480 min Ag Cu Ag3CuS2 Ag2S

1020 min Ag Cu Ag3CuS2 Ag2S


104

accordance with this observation, as the at.% of Ag, Cu and S are consistent with the Ag3CuS2

stoichiometry. Some small peaks in this diffractogram, like the weak peak at 16.62o, could not

be identified. This peak completely disappears for the samples immersed for longer times (120

min onwards). It is interesting to point out that for long immersion times (higher than 30 min),

and contrary to what is referred by several authors [44], it is AgCuS that is the most prevalent

corrosion product instead of Ag2S and copper sulphides (Cu2S and CuS). Ag2S was identified on

samples immersed between 1 and 7 min and on samples immersed for 480 min and 1020 min.

This compound was not detected on samples immersed for 15, 30, 60, 120 and 240 min. This

could be justified by the presence of either non-crystalline species (concentrations below

detection limits) or by a multilayer structure composed of several corrosion products, but not

including Ag2S.

Figure 4.16 - X-ray diffractogram obtained for the 60 min corroded Ag-Cu alloy with the identification of the Ag, Cu and Ag3CuS2 patterns. ● Ag (PCD file nº 00-004-0783); Cu (PCD file nº 00 002 1225), □ Ag3CuS2

(PCD file nº 00 012 0207) diffraction peaks.

The XRD results obtained for the analysed samples suggest the presence of different corrosion

products, whose nature depends on the time of immersion. However, as XRD cannot resolve

information as a function of depth at this level, it was impossible to determine whether the

layer-by-layer corrosion process, identified by SEM, corresponds to the formation of distinct

corrosion products.


105

Ag and Cu corrosion versus sterling silver corrosion

The colour variation of corroded Ag and Cu samples depends on the corrosion film thickness.

This film is mainly composed of Ag2S, in the case of Ag, and Cu2S in the case of Cu. The colour

variation of sterling silver is, instead, due not only to the film thickness but also to the formation

of different compounds. At the early stages of corrosion, based on the colour variation, there is

a significant influence of the Cu on the corrosion mechanism of sterling silver. Nevertheless,

these data on the pure elements provide evidence on the predominant role of Cu, even when

this element is present at low concentrations (7.5 wt% Cu in the sterling silver).

The morphology of the corroded surfaces of the pure elements and the alloy were compared. In

the case of Ag, it was shown the nucleation and growth of the same species, but with sizes and

shapes that change with the immersion time. Cu presents identical morphologies until 60 min

of immersion time, but for longer times it was observed the formation of particles with laminar

and round shape. At the early stages of corrosion, the surface morphology of the corroded

sterling silver has similarities with those of Ag and Cu. For longer corrosion times, the

morphology of the sterling silver corrosion evidently approaches the morphology of the Ag

corrosion. The corrosion layer was found to be thicker for pure elements. When all the samples

immersed during 1 min are compared, the estimated corrosion layer thickness for the pure

elements is about 90 nm higher than for sterling silver.

Ag is the main constituent element of the sterling silver alloy. It would be expected a major

influence of this element on the corrosion mechanisms of the alloy. However, data (colour and

morphology) suggest that the corrosion mechanism of sterling silver is predominantly influenced

by Cu at the early stages of corrosion, and by Ag for longer times. It should be highlighted that

sterling silver develops a layer-by-layer corrosion structure. This type of structure was not

observed for any of the pure metals, suggesting a distinct corrosion mechanism.

4.4.4 Gold-silver-copper alloys

Contrary to what was observed for Ag, Cu and sterling silver, the corroded surface of the Au

alloys did not show any colour variation. After 1 month immersion in a 50 mM Na2S solution,

the samples surface became duller, but no significant colour change was observed. It was only

after 12 months immersion that samples B4, T4, T5, T6, T7 and T8 became clearly red. These

samples, as shown in table 4.2, are all made from alloys containing more than 10 wt% Ag and

copper contents inferior to 7.5 wt%, corresponding to less than 90 wt% Au.


106

Sample T6, with the most accentuated colour alteration, was selected for further

characterisation. Figure 4.17 presents the reflectance spectra obtained for this sample before

and after corrosion. For comparison purposes, the reflectance spectra of the pure metals were

added to the plot. The spectrum obtained for the sample before corrosion shows a reflectance

increase between 320 and 630 nm, which is characteristic of the gold alloys [45]. Pure Au has a

steep reflectance increase between 500 and 600 nm, but the presence of Ag in the gold alloy

shifts the reflectance curve to lower wavelengths, close to the reflectivity region of Ag (λ = 325

nm) [46]. It is not perceived in the reflectance spectrum the influence of Cu. This is justified by

the fact that the Cu and the Au reflectivity curves have the same behaviour [47].

The spectrum obtained after corrosion of sample T6 shows a reflectance decrease that can be

linked to the surface alteration. This reflectance decrease corresponds to an absorption

between 320 and 600 nm, which can be ascribed to the presence of corrosion products.

Unfortunately, the characterisation of AuAgS, corrosion product identified by UV-Vis is not

referred in the literature. Thus, it is impossible to infer its presence on our sample. Based on the

results obtained for Ag, Cu and sterling silver samples it could also be expected the presence of

Ag- and Cu-based compounds, such as Ag2O, Ag2S, Cu2O and Cu2S. Accordingly to the results

obtained for the pure metals and for the sterling silver corrosion, and depending on the particles

shape and size, the absorption bands of these corrosion products vary in the range of

approximately 400 to 560 nm. Their presence can contribute to the band broadening and the

reflectance decrease, however, it is impossible to identify, based on the absorption values,

specific corrosion products.

Figure 4.17 – UV-Vis diffuse reflectance spectra of sample T6 before (A) and after corrosion test (B) and of pure metals used to obtained the alloy Ag, Cu and Au.


107

The optical properties and the corrosion layer thickness evolution during the corrosion test of

sample T6 was investigated by in situ ellipsometry. By fitting the experimental ellipsometric

parameters, Ψ and Δ, using a three-phase model which considers the corrosion solution,

corrosion layer and gold alloy substrate, it is possible to estimate the complex index of refraction

(ñ = n – ik, where n is the refractive index and k the extinction coefficient) and the thickness of

the corrosion layer [48]. Figure 4.18 displays the experimental data and the obtained theoretical

curves. The fitting was achieved considering the existence of two distinct layers. Thus, the

corrosion film is composed of a 60 nm thickness inner layer and a 20 nm outer layer, resulting

in an 80 nm total thickness of a corrosion film formed during 2090 minutes. The n values of the

two corrosion layers are the same, but the outer layer k value is higher than the inner layer k

value. Considering the wavelength of the laser used (λ = 632.8 nm), the data suggest for the

outer layer a higher absorption in the IR region [38], which is in accordance with the red surface

colour observed after the corrosion test. The estimated corrosion layer of T6 is significantly

lower than the calculated thickness of pure silver and pure copper corroded at the same

conditions. After 1 minute immersion, a corrosion film was already formed on the surface of

pure silver and of pure copper, with a 118 nm and 129 nm thickness, respectively.

Figure 4.18 – Experimental values of Δ vs. Ψ () acquired during the corrosion of T6 sample in a 0.1M Na2S solution and the obtained theoretical curves for the two-layer corrosion film: Inner layer (—) and outer layer (—).


108

All the samples were observed under the SEM. Figure 4.19 shows the surface of T6 before

corrosion. A nanoporous microstructure is observed in localised areas (figure 4.19B). This type

of structure is characteristic of a dealloying process and, in the case of gold alloys, to the

formation of gold atoms islands resulting from the interfacial phase separation [49]. This

structure can be due to annealing defects that lead to the formation of gold-rich surfaces,

observed in Chapter 3 for the Egyptian gold bead from the Harageh site. The surface defects can

be the source of higher corrosion development.

Figure 4.19 – SEM images of the T6 sample before corrosion.

The presence of corrosion products was detected on the surface of the corroded samples. As

shown in figures 4.20A and 4.20B, after 1 month corrosion, the T6 sample presents corrosion

products with a preferential nucleation near the nanoporous structure areas [50]. The corrosion

products are composed of spherical particles with different sizes, some presenting a well-

defined crystalline shape, morphologically similar to Ag2S (figure 4.20B). With the corrosion time

increase, the corroded layer thickens, becoming more regular. On the top of that corrosion layer,

corrosion products with different morphologies, like thin laminar shapes and tubular structures,

can be observed. The former is morphologically similar to copper sulphides and the latter to Au-

Ag-S corrosion products. Figures 4.20C and 4.20D also reveal the corrosion film cracking. Figures

4.20E and 4.20F display the morphology of the T6 corrosion film grown after circa 35 hours.

Spherical particles, like those formed at the corroded surface after 1 month (figure 4.20B), are

observed. It is possible to verify the presence of different layers, which confirms the results

obtained by ellipsometry.


109

Figure 4.20 – SEM images of the T6 sample surface after immersion during 1 month (A and B) and 12 months in a 50 mM Na2S solution (C and D) and during 35 hours in a 0.1 M Na2S solution (E and F).

The corroded surfaces of B4, T5, T7 and T8 were also observed under the SEM. Spherical

particles and, in localised areas, tubular formations could be observed. However, like for T6,

their distribution and shape is highly dependent on the surface finish.

The surface finish deeply influences the corrosion layer morphology. For example, it is possible

to observe for B4 several areas containing corrosion products of distinct morphologies, as shown

in figure 4.21. It was impossible to directly relate the presence of distinct corrosion products

with the gold alloy composition. Nevertheless, the morphologies observed are similar to those

reported for corrosion products that grew on naturally corroded gold alloys of the same


110

composition. For example, sample B4, made from an alloy containing 90 wt% Au, 10.6 wt% Ag

and 0.1 wt% Cu, show corrosion products morphologically similar to those observed for the

Abydos foil, discussed in Chapter 3, whose composition is 89.2 wt% Au, 10 wt% Ag and 0.7 wt%

Cu.

Figure 4.21 – SEM images of B4 surface sample after immersion during 12 months in a 50 mM Na2S solution.

Data obtained for the gold samples seem to suggest that the corrosion products morphology

depends not only on the composition of the alloy but also on the fabrication techniques, in this

case the surface finish. Although it is not clear in what way these parameters contribute to the

modification of the corrosion products, the different morphologies of those corrosion products

influence the final colour of the corroded surface.

In order to identify the corrosion products, the samples were analysed by EDS and XRD. EDS

data, displayed in table 4.10, was complex to quantify, due to the presence of the Au-Mα close

to the S-Kα lines. The layer thicknesses and the compounds morphologies, varying in one same

sample from point to point, prevented the direct comparison of the spectra obtained in the

different spots to estimate, at least, a relative composition. Nonetheless, it is possible to infer

that the Ag content increases and the Cu content decreases in the corroded areas. The S and the

O contents are variable, fact that can be related to the presence of distinct corrosion products.

The decrease of the Cu content tends to be slightly higher when the corrosion time increases,

suggesting the possible formation of Cu-based products at early corrosion stages, followed by

the formation of Ag-based compounds. This suggestion corroborates the ellipsometry data,

which showed the formation of a 100 nm in average corrosion layer. It is difficult to analyse such

thin layers by EDS. The penetration depth into the sample is too high, in spite of the high atomic


111

number of gold and of the alloy constituents, because this analysis reflects the surface

composition of the first 0.5 μm [51].

Table 4.10 – Composition of the corroded layer obtained by EDS and composition of the Au-base alloy obtained by µXRF for the Au alloy samples after immersion during 1 and 12 months in 50 mM Na2S solution and for the T6 sample also in immersion during 35h in a 0.1 M Na2S solution.

Composition in wt% Au Ag Cu S O Mα Lα Lα Kα Kα B4 (89 wt% Au; 11 wt% Ag) 1 month Plane area 88 10 - 1 1 82 15 - 3 0 12 months 83 14 - 0 2 84 12 - 1 2 Nanostructure 65 27 - 6 1 T5 (81 wt% Au; 12 wt% Ag; 7 wt% Cu) 1 month 72 16 4 0 9 23 53 5 14 6 83 13 4 0 1 12 months 87 12 0 0 1 83 14 2 0 1 56 37 2 1 4 57 37 1 2 3 T6 (76 wt% Au; 17 wt% Ag; 7 wt% Cu) 1 month Crystal 15 66 2 14 3 28 54 2 10 7 12 months 71 16 2 0 12 70 20 1 2 7 59 26 2 8 5 62 24 2 8 3 35 h in 0.1 M Na2S 66 29 3 1 0 66 30 3 0 1 53 40 3 3 1 54 39 3 2 1 55 38 3 2 1 50 42 3 3 2 T7 (74 wt% Au; 23 wt% Ag; 3 wt% Cu) 1 month 52 33 4 9 2 12 months 58 27 2 7 6 T8 (71 wt% Au; 25 wt% Ag; 4 wt% Cu) 1 month 73 24 3 0 1 72 26 1 0 1 12 months Plane area 68 28 3 0 1 62 25 2 3 8


112

Figure 4.22 shows the XRD diffractogram obtained for T6 after corrosion. The significant

broadening of the diffraction peaks, the shift in the centroids position and the long tails were

reported as characteristic features of nanostructured metals with plastic deformation [52],

confirming the presence of defects due to failed annealing observed under SEM (figure 4.19).

The diffractogram allowed to identify the presence of AuAgS (petrovskaite). This corrosion

product is also present at the surface of B4, T4 and T8. Possibly because present at very small

amounts when the immersion times are short, the presence of Cu-based corrosion products

could not be identified by XRD. In addition, and contrarily to what has been reported [53], the

presence of Ag2S was not identified, as well. It should be referred that the morphology of the

corrosion products composed of spherical particles could indicate the presence of that

compound. In fact, one crystal identified on the surface of T6, which can be observed in figure

4.20B highlighted with a red circle, is composed of 15 wt% Au; 66 wt% Ag; 2 wt% Cu, 14 wt% S

and 3 wt% O, which could correspond to a crystal of Ag2S.

Figure 4.22 – X-ray diffractograms obtained for T6 sample after corrosion in a 50 mM Na2S solution during 12 months with the identification of the main peaks of Au (PCD file nº00 004 0784), Ag (PCD file º 00 004 0783), Cu (PCD file nº 00 002 1225) and AuAgS (PCD file nº 00 038 0396).


113

Ag, Cu and sterling silver corrosion versus gold alloys corrosion

Corroded Au alloys samples have a red colour surface. Like for the Abydos foils (cf. Chapter 3),

gold alloys of distinct compositions developed corrosion with a uniform colour with no multi-

hued effect. Based on the results obtained for Ag, Cu and sterling silver corroded samples, it

could be expected the formation of different colours accordingly to the corrosion layer thickness

and to the different corrosion products formed. However, it was impossible to relate for the

gold alloys the thickness of the corrosion layer to the surface colours and the formation of

different corrosion products.

UV-Vis spectrophotometry data obtained for corroded and non-corroded Au alloys indicate a

reflectance decrease corresponding to an absorption increase in the range of 230 to 600 nm for

the corroded Au alloy samples. This absorption increase can be attributed to the presence of

corrosion products at the surface, but it was impossible to identify specific compounds.

The surface morphology observed by SEM shows for early corrosion stages the nucleation of

small near-spherical particles similar to those observed for the pure metals. The surface is

nevertheless more heterogeneous, with distinct particles of different sizes and shapes, showing

the complexity of the formed corrosion products. It was shown that the corrosion develops in a

layer-by-layer structure, equally observed for the sterling silver and for the corroded foils from

Abydos and Harageh (cf. Chapter 3). This layer-by-layer structure was confirmed by the results

obtained by in-situ ellipsometry. For an Au alloy (76 wt% Au; 17 wt% Ag; 7 wt% Cu) corroded

during 35h in a 0.1 M Na2S solution, the corroded surfae consists in a two-layer film composed

of a 60 nm thick inner layer and a 20 nm thick outer layer, which is significantly thinner than the

corrosion layer formed on Ag, on Cu and on sterling silver.

Contrary to pure metals and sterling silver, the elemental identification and quantification of the

corroded layer by EDS was complex due to the presence of the Au-Mα line close to the S-Kα line

and to the high effective penetration depth when considering the thickness of the two-layer

film.

Nevertheless, it could be suggested the possible formation at the early corrosion stages of Cu-

based products, followed by the formation of Ag-based products, similarly to what was observed

for the sterling silver corrosion process. The presence of AuAgS was identified by XRD.


114

4.5 Conclusions

The characterisation of the corroded layers of Ag, Cu, sterling silver and Au alloys samples by

immersion in a sulphide containing solution, allowed:

i) to infer on the role of the constitutive elements on the corrosion process of Au alloys,

ii) to estimate the corrosion thickness of Au alloys,

iii) to relate the colour of the corroded layer to the morphology of the corrosion products.

These parameters were not disclosed with the characterisation of the corroded gold objects,

presented in Chapter 3, and are important to establish the application limits, advantages and

disadvantages of the selected analytical techniques.

Observation under the SEM of the surface morphology of the corroded samples allowed to

suggest that the colour of the Ag and Cu corroded surfaces depends on the corrosion layer

thickness. The identification by XRD revealed that the corrosion layer is mainly composed of

Ag2S and Cu2S, respectively for the corroded Ag and Cu surfaces. The observation of the

corroded sterling silver surface also showed a colour variation, which is related to the formation

of distinct corrosion products with different morphologies, in a layer-by-layer structure. At early

corrosion stages, there is a prevalence of the formation of Cu-based compounds followed by the

formation of Ag-based compounds. Corroded Au alloys also showed the formation of a layer-by-

layer structure, suggesting that distinct compounds grow on their surfaces, as observed for the

foils from Abydos and Harageh. However, changes on the surface colour are not evident to see.

EDS analysis of the corroded Ag, Cu and sterling silver samples allowed to quantify the elemental

composition of the corrosion layer. For the sterling silver, the quantitative elemental

composition information on each layer of the layer-by-layer structure was achieved. However,

elemental composition information on Au alloys was more complex due to the presence of the

Au-Mα lines close to the S-Kα lines.

The corrosion layer developed on one Au alloy corroded during 35 h in a 0.1M Na2S solution was

estimated to be 80 nm, much thinner than those developed on Ag, Cu and sterling silver samples

that attain an average thickness of 140 nm after immersion during 3 min in that same solution.

The estimated thickness of the corrosion layer confirmed the data obtained for corroded gold

objects (cf. Chapter 3), that suggested the presence of such thin corrosion layers that their

elemental characterisation with the selected analytical techniques, EDS and XRF, could not be

carried out. The effective penetration depth of these two techniques is much higher than the

thickness of the corrosion layer.


115

The structural analysis by XRD of the corroded surface could provide the identification of the

compounds present on the Ag and Cu samples. However, for sterling silver and gold alloys, with

a layer-by-layer corrosion structure, the XRD equipment available does not have the required

spatial resolution (that should go down to 1 μm) to identify the compounds of each layer. In

addition, when the corrosion products are present in low quantities and/or have less crystallised

phases, their identification by XRD is difficult.


116

4.6 References

[1] T. Baše, Z. Bastl, V. Havránek, K. Lang, J. Bould, M.G.S. Londesborough, J. Macháček, J. Plešek, Carborane-thiol-silver-interactions. A comparative study of the molecular protection of silver surfaces, Surf. Coat. Technol. 204 (2010) 2639–2646. [2] P.P. Corso Jr., R.M. German, H.D. Simmons Jr., Tarnish evaluation of gold-based dental alloys, J. Dent. Res. 64 (1985) 856–859. [3] A.T. Kuhn, G.H. Kelsall, Methods for the testing of tarnish, Br. Corros. J. 18 (1983) 168–173. [4] H. Kim, Corrosion process of silver in environments containing 0.1 ppm H2S and 1.2 ppm NO2, Mater. Corros. 54 (2003) 243–250. [5] R. Wiesinger, R. Grayburn, M. Dowsett, P.-J. Sabbe, P. Thompson, A. Adriaens, M. Schreiner, In situ time-lapse synchroton radiation X-ray diffraction of silver corrosion, J. Anal. At. Spectrom. 30 (2015) 694–701. [6] J.P. Randin, J.P. Renaud, P. Ramonu, Tarnishing of AuAgCu alloys. Choice of an accelerated test, Werkst. Korros. 43 (1992) 48–55. [7] J.J. Tuccillo, J.P. Nielsen, Observation of onset of sulphide tarnish on gold-base alloys, J. Prosth. Dent. 25 (1971) 629–637. [8] M.S. Chana, A critique of the Tuccillo-Nielsen wheel method for tarnish testing of dental alloys, J. Dentistry 12 (1984) 314–318. [9] C. Kleber, R. Wiesinger, J. Schnöller, U. Hilfrich, H. Hutter, M. Schreiner, Initial oxidation of silver surfaces by S2- and S4+ species, Corros. Sci. 50 (2008) 1112–1121. [10] M. Huang, Y. Shao, X. Sun, H. Chen, B. Liu, S. Dong, Alternate assemblies of platinum nanoparticles and metalloporphyrins as tunable electrocatalysis for dioxygen reduction, Langmuir 21 (2005) 323–329. [11] M. Valodkar, S. Modi, A. Pal, S. Thakore, Synthesis and anti-bacterial activity of Cu, Ag and Cu-Ag alloy nanoparticles: a green approach, Mater. Res. Bull. 46 (2011) 384–389. [12] R. Chen, N.T. Nuhfer, L. Moussa, H.R. Morris, P.M. Whitmore, Silver sulfide nanoparticle assembly obtained by reacting an assembled silver nanoparticle template with hydrogen sulphide gas, Nanotechnology 19 (2008) 1–11. [13] X. Wang, S. Li, H. Yu, J. Yu, S. Liu, Ag2O as a new visible-light photocatalyst: self-stability and high photocatalytic activity, Chem. Eur. J. 17 (2011) 7777–7780. [14] P. Chattopadhyay, S. Guha Roy, Simple method of controlling the resistivity of silver sulphide thin film prepared by chemical conversion technique, Indian J. Pure Ap. Phy. 51 (2013) 98–103. [15] M. Morales-Masis, S.J. van der Molen, W.T. Fu, M.B. Hesselberth, J.M. van Ruitenbeek, Conductance switching in Ag2S devices fabricated by in situ sulfurization, Nanotechnology 20 (2009) 095710–095716. [16] V.J. Keast, T.A. Myles, N. Shahcheraghi, M.B. Cortie, Corrosion processes of triangular silver nanoparticles compared to bulk silver, J. Nanopart. Res. 18 (2016) 45–55. [17] H. Gil, C.P. Buitrago, A. Echavarría, Characterization of atmospheric corrosion products formed on silver in tropical-mountain environments, J. Solid State Electrochem. 19 (2015) 1817–1825. [18] H. Águas, R.J.C. Silva, M. Viegas, L. Pereira, E. Fortunato, R. Martins, Study of environmental degradation of silver surface, Phys. Stat. Sol. (c) 5 (2008) 1215–1218. [19] Y.B. Amor, E. Sutter, H. Takenouti, B. Tribollet, M. Boinet, R. Faure, J. Balencie, G. Durieu, Electrochemical study of the tarnish layer of silver deposited on glass, Electrochim. Acta 131 (2014) 89–95. [20] T. Tamai, Ellipsometric analysis for growth of Ag2S film and effect of oil film on corrosion resistance of Ag contact surface, IEEE T. Compon. Hybr. 12 (1989) 43–47. [21] S. Nakayama, T. Umemoto, Quantitative analysis of copper sulphides by a new electrochemical method, Sci. Technical Review 74 (2012) 8–14. [22] I. Grozdanov, M. Najdoski, Optical and electrical properties of copper sulphide films of variable composition, J. Solid State Chem. 114 (1995) 469–475.


117

[23] Z. Wu, M. Shao, W. Zhang, Y. Ni, Large-scale synthesis of uniform Cu2O stellar crystals via microwave-assisted route, J. Cryst. Growth 260 (2003) 490–493. [24] M. Senthilkumar, S.M. Babu, Synthesis and characterization of hexagonal faceted copper sulphide (Cu1.8S) nanodisks, Mat. Sci. Semicon. Proc. 40 (2015) 203–208. [25] K. Itoh, T. Kuzuya, K. Sumiyama, Morphology and composition-controls of CuxS nanocrystals using alkyl-amine ligands, Mater. Trans. 47 (2006) 1953–1956. [26] M.R. Johan, M.S. M. Suan, N.L. Hawari, H.A. Ching, Annealing effects on the properties of copper oxide thin films prepared by chemical deposition, Int. J. Electrochem. Sci. 6 (2011) 6094–6104. [27] M. Ramya, S. Ganesan, Influence of thickness and temperature on the properties of Cu2S thin films, Iran. J. Sci. Technol. 37 (2013) 293–300. [28] T. Martino, R. Partovi-Nia, J. Chen, Z. Qin, D. W. Shoesmith, Mechanisms of film growth on copper in aqueous solutions containing sulphide and chloride under voltammetric conditions, Electrochim. Acta 127 (2014) 439–447. [29] J. Chen, Z. Qin, D.W. Shoesmith, Copper corrosion in aqueous sulphide solutions under nuclear waste repository conditions, Mater. Res. Soc. Symp. Proc. 1475 (2012) 465–470. [30] T. Palomar, B.R. Barat, E. García, E. Cano, A comparative study of cleaning methods for tarnished silver, J. Cult. Herit. 17 (2016) 20–26. [31] S. Berthier, Optique des milieux composites, Polytechnica, Paris, 1993. [32] A. Szekeres, S. Bakalova, S. Grigorescu, A. Cziraki, G. Socol, C. Ristoscu, I.N. Mihailescu, AIN:Cr thin films synthesized by pulsed laser deposition: Studies by X-ray diffraction and spectroscopic ellipsometry, App. Surf. Sci. 255 (2009) 5271–5274. [33] N. K. Abbas, N. J. Ghdeeb, The effect of thickness on the optical properties of Cu2S thin films, Iraqi J. Phys. 13 (2015) 121–127. [34] E.D. Palik, Handbook of Optical Constants of Solids, vol. II, Academic Press, Boston, 1991. [35] S. Baset, H. Akbari, H. Zeynali, M. Shafie, Dig. J., Size measurement of metal and semiconductor nanoparticles via UV-Vis absorption spectra, Nanomater. Bios. 6 (2011) 709–716. [36] S. N. Guin, J. Pan, A. Bhowmik, D. Sanyal, U. V. Waghmare, K. Biswas, Temperature Dependent reversible p-n-p type conduction switching with colossal change in thermopower of semiconducting AgCuS, J. Am. Chem. Soc. 136 (2014) 12712–12720. [37] Y. Yu, L. Zhang, J. Wang, Z. Yang, M. Long, N. Hu, Y. Zhang, Preparation of hollow porous Cu2O microspheres and photocatalytic activity under visible light irradiation, Nanoscale Res. Lett. 7 (2012) 347-352. [38] L. Abrantes, J. Correia, A.I. Melato, An ellipsometric study of poly(3,4-ethylenedioxythiophene) electrosynthesis – from the initial stages to tick layers formation, J. Electroanal. Chem. 647 (2010) 75–84. [39] R. van Langh, H. A. Ankersmit, I. Joosten, The delamination of silversulphide layers in Proceedings of the International Conference on Metals Conservation, METAL04 (J. Ashton, D. Hallam eds.) Canberra, Australia (2004) pp. 137–141. [40] L. Selwyn, Metals and corrosion. A handbook for the conservation professional, Canadian Conservation Institute, Ottawa, 2004. [41] H. Ha, J. Payer, The effect of silver chloride formation on the kinetics of silver dissolution in chloride solution, Electrochim. Acta 56 (2011) 2781–2791. [42] S. Capelo, P. Homem, J. Cavalheiro, I. Fonseca, Linear sweep voltammetry: a cheap an powerful technique for the identification of the silver tarnish layer constituents, J. Solid State Electr. 17 (2013) 223–234. [43] Y. Asadov, Y. Alyev, K. Jafarov, X-ray diffraction study of compounds in the Ag2S-CuS system, Inorg. Mater. 44 (2008) 460–466. [44] J. Hirunyagird, G. Lothongkum, E. Nisaratanaporn, An improvement in tarnish and corrosion resistance of 94Ag-4Zn-Cu alloys with Sn addition, Adv. Mat. Res. 894 (2014) 138–142.


118

[45] T. Shiraishi, R.J.D. Tilley, An estimation of the reflectivity of some monophasic binary gold alloys, Gold Bull. 47 (2014) 75–82. [46] O. Loebich, The optical properties of gold. A Review of their technical utilisation, Gold Bull. 5 (1972) 2–10. [47] E.F.I. Roberts, K.M. Clarke, The colour characteristics of gold alloys, Gold Bull. 12 (1979) 9–19. [48] R.M.A. Azzam, N.M. Bashara, Ellipsometry and polarised light, 3th ed., Elsevier, Amsterdam, 1996. [49] Y. Ding, Y.-J. Kim, J. Erlebacher, Nanoporous gold leaf: “Ancient Technology”/Advanced material, Adv. Mater. 16 (2004) 1897–1900. [50] A.J. Forty, Micromorphological studies of the corrosion of gold alloys, Gold Bull. 14 (1981) 25–35. [51] L.G. Troalen, J. Tate, M.F. Guerra, Goldwork in Ancient Egypt: workshop pratices at Querneh in the 2nd Intermediate Period, J. Archaeol. Sci. 50 (2014) 219–226. [52] I.V. Alexandrov, R.Z. Valiev, X-ray analysis of bulk nanostructured metals, Mater. Sci. Forum 321-424 (2000) 577–582. [53] D.M. Bastidas, E. Cano, A.G. González, S. Fajardo, R. Lleras-Pérez, E. Campo-Montero, F.J. Belzunce-Varela, J.M. Bastidas, An XPS study of tarnishing of a gold mask from a pre-Columbian culture, Corros. Sci. 50 (2008) 1785–1788.

Chapter 5

Towards the definition of an analytical strategy

Chapter 5 – Towards to a definition of an analytical strategy

121

5. Towards the definition of an analytical strategy

5.1 Introduction

The characterisation of corrosion products on gold objects stored and displayed in museums

requires three main axes of research:

1) assessment of the environmental conditions;

2) identification of the alloy composition and the fabrication techniques employed;

3) identification of the corrosion products and the corrosion layer thickness.

In the area of cultural heritage, the corrosion of metallic objects is approached using equivalent

procedures [1,2]. When dealing with gold objects, the main difficulty of this type of approach lies

on the limitation of the analytical methods providing information on the substrate and on the

morphology and constitution of the corroded layer, which are seldom portable and non-

destructive. In fact, it was possible to estimate in this work the thickness of the corrosion layer

that grows on gold alloys submitted to atmospheric corrosion to less than 100 nm, which very

much restricts the number of techniques applicable to its characterisation. It is frequently difficult

to move gold objects for analysis, and it is difficult to have access to equipment allowing the non-

destructive analysis of thin layers on big objects. The characteristics of the portable equipment

tested so far for the determination of the elemental composition of the gold objects, do not allow

to gather information from the first 100 nm only.

In spite of the difficulties encountered throughout this study, it was possible to define an analytical

strategy that could in the future be applied to the characterisation of corroded gold objects stored

and displayed in museums, built on the strength of the study of three cases – the prehistoric

goldwork in the collection of the National Museum of Archaeology, the jewellery of René Lalique

in the collection of the Gulbenkian Foundation Museum, and the Egyptian gold foils in the

collections of the Garstang Museum of Liverpool and the National Museums Scotland, in the

United Kingdom - as well as on the simulated samples made from binary and ternary gold alloys

fabricated for this particular work to be subjected to artificial corrosion. The analytical study of

both natural and artificial corrosion layers enabled the definition of guidelines for an analytical

strategy, presented and discussed in this chapter.


122

5.2 Exhibition/storage context assessment

Corrosion of gold objects occurs in environments with high levels of relative humidity (RH) and in

the presence of sulphur and chloride. The three case studies revealed that the corrosion products

are, in their majority, composed of sulphur compounds. The source of this pollutant can be either

environmental or due to the materials degradation inside the showcases. However, although

sulphur was identified in these three cases as the main pollutant for the corrosion of gold alloys,

the environmental assessment should always be carried out before any analytical characterisation

of the corroded objects. Each case may present different atmospheric characteristics, influenced

by several parameters like the museum location, the RH and temperature (T) values, etc., and the

presence of other pollutants that must be assessed.

RH and T monitoring should include the environment inside and outside the showcase(s), inside

and outside the exhibition/storage room, and outdoor. The monitoring should be carried out

during a period that allows verifying the RH and T behaviour inside the showcases and the

exhibition/storage room when there are pronounced RH and T outdoor fluctuations.

Pollutants identification and quantification should be carried out inside and outside the

showcases, as well as near the air inlets and outlets. The values should be compared to those

obtained for the indoor and outdoor environments. The main gaseous pollutants that should be

assessed are the NOx, O3, CO, CO2 (outdoor pollutants) and SO2, COS, CS2, H2S, CH3COOH, CH2O,

C2H6S, and Cl (indoor pollutants) [3]. Pollutant measurements can be carried out by active and

passive methods. Active methods involve sampling the air with a pump and posterior analysis. The

identification can be carried out by distinct analytical methods such as mass spectrometry, gas

chromatography, high-performance liquid chromatography and infra-red spectroscopy. These

methods are sensitive and accurate with detection limits in the range of the ρg.g-1 for all the air

components. Nevertheless, they are expensive, and the results are time-consuming [4]. Passive

methods encompass, for example, colorimetric gas detector tubes (diffusion tubes) like those

used in this work. Analysis are carried out in-situ, and the result is obtained in a short period.

However, they identify specific pollutants with detection limits that are normally higher than the

threshold of their potential influence on the corrosion development. This situation was observed

in the Treasure room of the National Museum of Archaeology. Despite the high concentration of

SO2 in the showcases of the exhibition room, revealing a high polluted museological environment,

the values of H2S, another reduced sulphur compound, could not be quantified because present

in quantities lower than 0.2 µg g-1, the detection limit of the employed H2S diffusion tube. In spite

of this, this method allows to carry out a first assessment of the environmental situation for


123

further analysis. Another passive method, also inexpensive and easy to use, that can complement

the data obtained with the diffusion tubes method is the corrosivity test that uses metallic

coupons. This method can identify the source of pollutants, but cannot quantify them.

5.3 Gold alloy composition and fabrication techniques assessment

The gold alloy composition has influence on the corrosion development [5]. Also, it is expected

that the presence of different contents of silver and copper in the alloy promotes the formation

of distinct corrosion products. The case studies presented in Chapter 3 could not identify the

influence of the gold alloy composition on the corrosion development. However, data obtained

by SEM-EDS, XRD and ellipsometry for artificially corroded gold alloys showed the influence of

their elemental composition on the corrosion development rate. The corrosion phenomena seem

to start in gold alloys when the amount of gold is below 90 wt%. The influence on the corrosion

development of the presence of silver in the alloy is higher than that of copper. However, the

corrosion products identified either in the objects and in the artificially corroded gold alloys were

the same, Ag2S and AuAgS.

The corrosion processes also depend on the thermo-mechanical processes to which the object

was subjected. Casting defects such as internal shrinkages and porosities result in a

heterogeneous surface texture. If this characteristic of the surface structure is a major factor

contributing to the corrosion development, it is also responsible for the possible misinterpretation

of a corrosion process. As referred, the change of the surface colour is the main visual hint of the

corrosion process development, when conservation procedures are carried out. This change in the

surface colour can, however, also result from a different surface structure acquired during

fabrication. Some areas of the brooch Ophelia made by René Lalique (discussed in Chapter 3)

showed indeed the formation on its surface of a nanostructure caused only by a casting defect.

Hence, the visual aspect of the brooch changed, because of a different light reflection caused not

only by the corrosion process but also by a nanoporous structure formed during the fabrication of

the object. An intervention without a clear separation of these two phenomena might lead to the

removal of parts that enclose information on manufacturing.

The identification of the gold alloy composition and the fabrication techniques is, therefore,

mandatory to understand the corrosion process of an object. If the elemental composition of the

alloy can be easily determined in-situ by using portable XRF equipment, it is difficult to assess

certain microstructure particularities that are only visible at the micrometer or nanometer level.


124

These studies require the use of SEM. Benchtop SEMs are available, with high and low vacuum

and an EDS system mode, but with a maximum sample size limited to approximately 7 cm

diameter and 5 cm height and only with 3 stages of acceleration voltages till 15 kV for SE images

and 2 stages for BES images.

To overcome this situation, it is possible to make surface replicas from the corroded objects to

study their structure. This method allows to replicate the object surface in-situ and to observe the

replica by SEM. The replica technique is widely used by engineers, and there are references to its

application in the field of cultural heritage [6-9]. Several materials can be used to replicate

surfaces, being the most common the silicone-based ones. However, the use of silicone should be

avoided, since this material may induce surface corrosion due to its composition [10]. Cellulose

acetate sheets are an alternative material that is used in metallographic studies, in crack

inspection and topographic surface research in the space-engineering domain [8]. Surface replicas

obtained using cellulose acetate sheets, allow detecting 25 μm long cracks indicating a high

resolution of the surface imprinting [11].

To evaluate the possibilities offered by this method in the case of corroded metallic alloys studied

in this work, the surface of a naturally corroded sterling silver box without cultural value was

replicated with cellulose acetate sheets (figure 5.1A). SEM images of the thick the corrosion layer

are shown in figure 5.1. Figure 5.1B shows the morphology of the corroded surface obtained

directly on the box, and figure 5.1C the negative replica of the surface imprinted in the cellulose

acetate sheet. Due to the corrosion layer thickness, it was not possible to replicate the silver alloy

substrate. However, the preliminary results suggest that this method is suitable to examine the

surface morphology. The surface replica can thus become a potential method to be used for the

characterisation of objects that cannot be moved from the museum. Nevertheless, further

investigation should be carried out to identify the applications and limits of the cellulose acetate

sheets. Namely, the effects of the replica process on the corroded surface should be assessed, and

it should be evaluated whether the substrate can also be observed.


125

Figure 5.1 – Sterling silver box used for testing the surface replica method carried out with a 30 µm cellulose acetate sheet (A). SEM images of the box corroded layer (B) and of the negative replica of the surface showing the imprinted corrosion products (C). The cellulose acetate sheet was coated with gold during 20 s at 100 mA in a EMITECH K550X sputter coater.

5.4 Corrosion products identification

The results obtained in Chapters 3 and 4 suggest that corrosion mechanisms of gold alloys are

complex, involving the formation of different compounds, whose concentrations depend on the

alloy composition. The corrosion develops in a layer-by-layer structure. Each thin layer has a

distinct composition, morphology and thickness. The analytical limitations described in those

chapters prevent the full understanding of the corroded layer formation and, therefore, of the

corrosion mechanisms.

In fact, the analytical techniques selected for this study cannot provide the composition of each

corrosion layer. Their spatial resolution is not at the appropriated scale, and their depth resolution

is not adapted to the variation of composition observed with depth (layer-by-layer structure). This

is clearly exemplified by the data obtained for the Abydos gold foils. SEM imaging revealed the

presence of three thin layers composed of corrosion products with different morphologies. While

EDS analyses allowed the quantification of the elements present in each layer, this elemental

information is insufficient to identify the corrosion products. The structural composition of the

corrosion layer that could provide information on these compounds, was unsucessfully searched

by XRD. For these foils, the required spatial resolution should go down to circa 1 μm.


126

In brief, the results obtained in Chapters 3 and 4 showed that with mobile analytical techniques it

is impossible to characterise the corroded surface of a gold object.

Therefore, two situations need to be considered: i) the objects can be moved for analysis and the

techniques available in a laboratory can analyse large samples or ii) the object cannot be moved.

To overcome the latter situation and when large samples cannot be analysed, a protocol can be

implemented, based on the creation of samples with the same characteristics of the objects. It

should however be emphasised that the elemental composition is a parameter necessary but

insufficient to approach the corrosion of an original object. The fabrication techniques that

employ thermo-mechanical processes also contribute to the development of corrosion. The

samples should, thus, be fabricated with the same alloy and submitted to equivalent thermo-

mechanical stress.

A set of samples produced as referred can be placed in the exhibition and storage rooms, close to

the objects, to reproduce the corrosion mechanisms that depend on the environmental

parameters. Otherwise, the environmental conditions can be assessed in order to submit the same

set of samples to accelerated corrosion. The corroded samples can then be analysed by any

analytical technique, because they can be moved and destroyed if necessary; the results obtained

can then be extrapolated for the objects. However, this protocol is time-consuming, and full

reproduction of the object is hard to achieve, in particular the stress caused during manufacturing.

In this work, based on the data obtained for objects in exhibition and storage rooms, samples

made from the same alloys were fabricated to be subjected to accelerated corrosion together

with samples made from the single metals that constitute the alloys. No trial was carried out to

submit samples to stress. To accomplish this, it would have been necessary to fully describe the

various aspects of the stress induced by any of the large range of techniques employed by the

goldsmith.

It was indeed possible to move those samples for analysis, but even then, some limitations could

be addressed. UV-Vis was used to analyse the colour of corroded samples. The absorption bands

in the spectra can be ascribed to the presence of compounds (corrosion products) with specific

dimensions, depending on the particles size. However, the analysis of corroded gold samples only

showed a reflectance decrease suggesting that different corrosion products could absorb in the

same wavelength range. The total thickness of the corrosion layer that grows on gold alloys could

be determined by ellipsometry and estimated for one sample to be in the order of 80 nm.

Concerning the elemental composition of the corroded layer, the use of EDS was limited to the

identification of sulphur due to the tail of the intense peak corresponding to the Au- M lines close


127

to the S-K line. When a corrosion layer is too thin (below 80 nm, according to the results) it is

difficult to identify and quantify the sulphur. The structural analysis by XRD in grazing incidence

configuration of the corroded surface would provide the identification of the compounds present

in the analysed areas. However, when several corrosion products are present in the same area,

their identification is complex due to the similarity of corrosion product patterns and peak

intensities. Nonetheless, it was so far possible to identify on the surface of gold alloy samples,

Ag2S and AuAgS compounds. Another limitation of the XRD could be addressed, the spatial

resolution of the equipment used is inappropriate to identify the compounds that constitute each

layer of the layer-by-layer structure that develops on the corroded gold alloys.

When the objects can be moved for analysis, the analytical limitations are similar to those

described for samples, but with addition of the drawbacks related to the size, shape and surface

roughness of the objects. When the objects are too big, they might not fit the SEM sample

chamber. When their surface roughness is accentuated, it is difficult to determine the corroded

layer thickness. The surface roughness induces light scattering disabling the use of ellipsometry,

as it was shown for the Egyptian gold foil from Abydos.


128

5.5 Conclusions

Considering the characteristics of the corrosion layer of gold alloys, it is impossible to characterise

the corrosion of gold object in-situ with the presently available portable equipment.

Consequently, the characterisation of the corrosion layer can only be achieved by using stationary

equipment, with high spatial resolution and more appropriate penetration depth.

When the gold objects cannot be moved to be analysed ex-situ, their corroded surfaces can be

simulated by the production of samples having the same composition. In order to produce correct

replicas, the stress induced by the fabrication techniques employed should also be simulated.

These samples can then be corroded in the same environment as the objects or submitted to

accelerated corrosion. The former can be attained by placing the samples in the showcase or

storage and exhibition rooms, near the objects. When accelerated corrosion is adopted, the

environmental conditions should be assessed and the samples corroded in a similar environment.

The corroded samples can then be analysed by using selected analytical techniques. This approach

is time-consuming, but, considering the results ontained in this work, it gives representative

information on the corroded layer. To be fully representative of the original surfaces, the stress

induced by the fabrication techniques employed should also be considered.

When the objects can be moved, their optical properties, corrosion layer thickness, surface

morphology and composition can be researched by UV-Vis, ellipsometry, SEM-EDS, µXRF and XRD.

However, there are some limitations concerning the selected analytical techniques considered in

this work that are summarised in table 5.1.


129

Table 5.1 – Overall evaluation of the selected analytical techniques for the characterisation of corroded gold alloy surfaces.

Optical properties

UV-Vis Applications: Characterisation of the corroded surface colour. Identification of elements and compounds present at the surface.

Limitations and disadvantages: - The spectrum interpretation can be complex if the surface contains several corrosion products presenting absorption bands in the same wavelength range.

Ellipsometry Applications: Identification of optical constants (n, k) of each corroded layer.

Limitations and disadvantages: - Surface roughness and shape of the object can induce light absorbance and scattering, hindering the measurement.

Morphology

SEM Applications: Characterisation of the morphology of different corrosion products and distinct corrosion layers.

Limitations and disadvantages: - Limitations concerning the dimensions and shape of the object due to the chamber dimensions.

Thickness

Ellipsometry Applications: Determination of the corroded layer thickness.

Limitations and disadvantages: - Surface roughness, dimensions and shape of the object can induce light absorbance and scattering preventing the measurement.

Composition

Elemental XRF Applications: Elemental composition of the gold alloy. Elemental composition of the corroded layer.

Limitations and disadvantages: - Depending on the experimental conditions: nature of the X-ray tube anode and the operation conditions - the identification and quantification of the composition of the corroded layer can be difficult to achieve.

EDS Applications: Elemental composition of each corrosion layer.

Limitations and disadvantages: - If the corrosion layer is too thin (< 80 nm) it is difficult to identify and quantify S due to the penetration depth and to the intense peak tail corresponding to the Au-M lines close to the S-K line.

Structural XRD Applications: Identification of crystalline corrosion products.

Limitations and disadvantages: - Limited database for corrosion products of gold alloys. - Complex identification of the corrosion products when more than one compound is present, because XRD patterns are similar. - not suitable when corrosion products are present in small quantities or with low crystallinity. - The low spatial resolution prevents the identification of the corrosion products present in each corrosion layer.


130

5.6 Outlook

Based on the assessment of three case studies of corroded gold objects in national and

international collections and of a set of artificially corroded binary and ternary gold alloys, the

present work allowed to define an analytical strategy to be applied to the identification of thin

corrosion layers growing on ancient gold objects.

To obtain four types of information - optical properties, corrosion layer thickness, surface

morphology and elemental and structural composition – a set of analytical techniques (UV-Vis,

ellipsometry, SEM-EDS, XRF and XRD) was selected. Their advantages and limitations were

assessed and it was shown that it is impossible to characterise the corroded gold objects in-situ

by using the available portable equipment.

This requires studying the corroded gold objects ex-situ, which is first limited by the possibility to

move them. In addition, the use of certain techniques, like SEM, might be restricted to objects

with small dimensions and in a good state of conservation. It was proposed a method to overcome

this situation, based on the fabrication of samples with identical composition and corroded in the

same environmental conditions of the gold objects. However, it was underlined that when the

stress caused by manufacturing is not considered, a full understanding of the corrosion

development cannot be attained.

This unsolved question should be further investigated. To approach this issue, the production of

gold alloy samples fabricated by a goldsmith with different manufacturing techniques could be

envisaged. For a first trial, the gold alloy that showed the most accentuated colour alteration after

the corrosion tests carried out in this work (76 wt% Au, 17 wt% Ag and 7 wt% Cu) should be

selected. Hammered, laminated and casted samples could then be fabricated. The influence of

other techniques like soldering on the corrosion development could also be assessed.

Those samples could then be artificially corroded in the same conditions used in this work and

characterised applying the proposed analytical protocol to which other analytical techniques

should be added, in order to obtain further information on each one of the corroded layers that

constitute the identified corroded layer-by-layer structure of the gold alloys. To be able to

characterise each layer, two possible research orientations can be foreseen:

i) to use XRD with high spatial resolution (1 µm) allowing the identification of each layer.

To attain the required low spatial resolution, µXRD with micro and sub-micro beams

using synchrotron radiation could be investigated [12].


131

ii) to monitor the corrosion process in-situ, by immersion of a gold alloy sample in a

sulphide solution. This can be done using analytical techniques with nanoscale

resolution, like, for example, neutron reflectivity, a technique that is being used to

nanoscale control of interfacial processes studies including corrosion studies [13].

The present work also pointed out future challenges in the field of cultural heritage. It was shown

that gold alloy objects undergo atmospheric corrosion when exhibited or stored in environmental

conditions with high humidity values and in the presence of low concentrations of pollutants.

Although preventive measures concerning the exhibition and storage environments conditions

are not new in the field of cultural heritage, in the case of gold objects these procedures are still

not fully implemented as gold objects are so far, consider to be corrosion resistant. The case

studies considered in this work stressed out the need to monitor and to control the exhibition

conditions in order to prevent the corrosion development before carrying out a conservation

treatment. If it is necessary to carry out a conservation treatment to remove the corrosion,

justified by the need to recover the original aspect of the object, it should be emphasised that the

corroded layers formed on gold surfaces are nanometric. However, the currently used

conservation methods applied to the removal of the corrosion products do not guarantee the

elimination of nanolayers. In this case, the challenge is to research for cleaning methods that

ensure the controlled removal of the nanolayers and that do not interfere with the gold substrate,

removing wear marks or other characteristics of the objects manufacturing. Based on the present

knowledge on the conservation materials two directions could be researched: the use of

electrochemical techniques and laser ablation cleaning. Both techniques involve the selective

removal of the corrosion products. Therefore, their application as conservation methodology to

corroded gold objects is a challenge for conservation science. In a first approach, corroded gold

alloys fabricated and corroded following the method applied in this work could be used as samples

to test the applicability of those cleaning techniques. The cleaning effects on the surface could be

assessed by characterising the surface before and after the tests with the analytical protocol

proposed in this work.


132

5.7 References

[1] A. Arafat, M. Na´es, V. Kantarelou, N. Haddad, A. Giakoumaki, V. Argyropoulos, D. Anglos, A.-G. Karydas, Combined in situ micro-XRF, LIBS and SEM-EDS analysis of base metal and corrosion products for Islamic copper alloyed artefacts from Umm Qais museum, Jordan, Journal of Cultural Heritage 14 (2013) 261‒269. [2] M.F. Alberghina, R. Barraco, M. Brai, T. Schillaci, L. Tranchina, Integrated analytical methodologies for the study of corrosion processes in archaeological bronzes, Spectrochim. Acta B 66 (2011) 129‒137. [3] A. Schieweck, B. Lohrengel, N. Siwinski, C. Genning, T. Salthammer, Organic and inorganic pollutants in storage rooms of the Lower Saxony State Museum Hanover, Germany, Atmos. Environ. 39 (2005) 6098‒6108. [4] P. Hatchfield, Pollutants in museum environment: practical strategies for problem solving in design, exhibition and storage, WAAC Newsletter 26 (2004) 10‒22. [5] L.W. Laub, J.W. Stanford, Tarnish and corrosion behaviour of dental gold alloys, Gold Bull. 14 (1981) 13–18. [6] M. Sax, N. Meeks, The manufacture of a small crystal skull purported to be from ancient Mexico, Technical Research Bulletin (2009) 47‒55. [7] R. Bertholon, N. Lacoudre, J. Vasquez, The conservation and restoration of the copper scroll from Qumran, in Copper scroll studies (G.J. Brooke, P.R. Davies eds.) T&T Clark International, London (2002) pp. 12‒24. [8] M.H. Swain, Monitoring small.crack growth by replication method in Small-crack test methods (J.M. Larsen, J.E. Allison eds.) ASTM, Philadelphia (1992) pp. 34‒56. [9] A.J. Gwinnett, L. Gorelick, The change from stone drills to copper drills in Mesopotamia, Expedition 29 (1987) 15–24. [10] J. Tétreault, Display Materials: The Good, the Bad and the Ugly, Scottish Society for Conservation and Restoration (SSCR) Edinburgh, 1994. [11] J.A. Newman, S.A. Willard, S.W. Smith, R.S. Piascik, Replica-based crack inspection, Eng. Fract. Mech. 7 (2009) 898‒910. [12] A.S. Budiman, Synchrotron white-beam X-ray microdiffraction at the Advanced Light Source, Berkeley Lab in Probing crystal plasticity at the nanoscales. Synchrotron X-ray microdiffraction, Springer Singapore, 2015, pp. 15‒35. [13] J.J. Noël, Oxide films and corrosion in Neutron reflectometry, a probe for materials and surfaces, Proceedings of a Technical Meeting, International Atomic Energy Agency, Vienna, 16-20 August (2004) pp. 79‒84.

Appendix 1

Analysis of a gold alloy corrosion layer

by Raman spectroscopy

Appendix 1 – Analysis of a gold alloy corrosion layer by Raman spectroscopy

135

App. 1 Analysis of a gold alloy corrosion layer by Raman

spectroscopy

Raman spectroscopy has been widely used in the field of cultural heritage for the identification

and characterisation of molecular structures such as pigments, textile fibres, minerals,

gemstones, glass, and polymers [1-3]. It has also been applied for the characterisation of patinas

on copper alloys [4] and for corrosion studies of iron, copper and silver [5-8].

Raman spectroscopy is based on the scattering of monochromatic radiation after impinging on

a sample. Scattering may be elastic (Rayleigh scattering) or inelastic (Raman scattering). The

Raman and the Rayleigh scatterings are described in terms of electromagnetic radiation

generated by oscillating electric dipoles in the materials, induced by the incident excitation

radiation [9]. The induced dipole moment occurs as a result of the molecular polarizability. In

Raman scattering, a photon excites a molecule from the ground state to a virtual energy state.

The molecule immediately returns to ground electronic or vibrational states emitting photons.

Thus, due to energy differences between the original and the new state, the wavelength of the

emitted photon shifts, causing a light scattering. If the Raman scattered photon has lower energy

than the incident photons, it results in a Stokes-Raman shift; if the Raman scattered photon has

higher energy than the incident photons, it results in an anti-Stokes-Raman shift. Since the initial

population of the excited state is usually very small, the anti-Stokes lines in the spectrogram are

much weaker than the Stokes lines and sometimes may not be observed. As molecular energy

levels are quantified, the interaction produces discrete lines from which information about the

molecule itself can be obtained. It should be noted that some vibrations are not observed in

Raman spectroscopy. The general selection rule to determine the activity of a vibrational mode

in Raman states that a molecule is active in Raman when there is a change in polarizability during

the vibration.

Raman spectroscopy can be used to identify oxides and sulphides, characteristic compounds of

the corrosion products developed on metals, whose vibration frequencies are close to the IR

wavelength range [10]. The penetration depth, which can attain several tens of nanometer when

a laser with a shorter wavelength is used [11], turns Raman spectroscopy to a powerful

technique to characterise thin films. In addition, it is a non-invasive technique with a high spatial

resolution, and it is available as portable equipment.

Appendix 1 – Characterisation of gold alloy corrosion layer by Raman spectroscopy

136

These characteristics make this technique suitable for the determination of the compounds

present in the corrosion layer of gold alloys. As mentioned, Raman spectroscopy has in fact

been used for the characterisation of corrosion products of copper, iron and silver, but its

application to gold alloys corrosion is not reported in the field of cultural heritage. This situation

is explained by the absence of a reference spectrum database for the chemical compounds that

constitute the corrosion products, which prevents their identification. This same limitation

occurred in the case of corroded silver, but the work developed by Martina et al. [10] who

created a reference database for the corrosion products of silver caused by atmospheric

corrosion, namely Ag2O, AgCl, Ag2S, Ag2SO4, Ag2SO3, Ag2CO3, AgC2H3O2, AgNO3, allowed to

consider this technique as a useful tool for the identification of silver corrosion products.

The identification of the corrosion products of gold alloys when subjected to atmospheric

corrosion is more complex, because the corrosion process has not yet been fully investigated.

Moreover, although sulphur-based compounds are expected to be formed during the corrosion

process, the totality of the possible corrosion compounds are unknown.

Considering the potentiality of Raman spectroscopy to characterise thin films, to assess whether

this technique could integrate the analytical protocol for the study of the corrosion layers of

gold alloys, a first test was carried out on one gold alloy sample. This gold alloy (83 wt% Au, 15

wt% Ag, 2 wt% Cu) was subjected to accelerated corrosion in sulphide containing environment1

.

The Raman spectra of the gold alloy were obtained using the XploRA Confocal Raman

microscope by Horiba France, with a 638 nm laser source and the 50X and 100X objectives. Using

an entrance slit of 100 μm, the scattered light collected by the objectives was dispersed onto

the air-cooled CCD array of an Andor iDus detector by a 1200 lines/mm grating. The incident

power on the sample ranged from 0.1 to 1 mW. The analyses were made in the corroded surface

areas showing the characteristic red colour of the corrosion process and in non-corroded yellow

areas. Spectra obtained for these two areas are compared in figure app1.1.

1 The corrosion test was carried out following the corrosivity test protocol described in Chapter 3 (section 3.2.2).

Appendix 1 – Analysis of a gold alloy corrosion layer by Raman spectroscopy

137

Figure app.1.1 – Raman spectra obtained for the non-corroded (―) and corroded (―) Au alloy sample.

The Au alloy spectrum shows three bands in the range of 107‒306 cm-1. Bands at 107 and 224

cm-1 can be ascribed to Ag vibrational modes, as they are observed in the Raman spectra of

various Ag compounds [12]. The Raman spectrum obtained for the corroded area has the same

profile, but the bands have a higher intensity, which can be due to the presence of corrosion

products and/or an effect of the variable surface roughness, that results from the grinding

process. The intensity increase of the 224 cm-1 band is higher than the intensity increase of the

107 and 306 cm1 bands. The 224 cm-1 band also shows a slight enhancement of the right tail,

which may correspond either to Ag-S stretching modes, that could suggest the presence of Ag2S

[8] or to the presence of Cu2O. This compound has a characteristic strong band at 218 cm-1 and

a broader band at circa 523 cm-1 [13], also visible in the spectrum at 520 cm-1. However, this

data interpretation is, as above-mentioned, strongly limited by the inexistence of an adequate

database of reference Raman spectra for gold alloys and their possible corrosion products.

For this reason, Raman spectroscopy was not selected to integrate our analytical approach to

the corrosion processes of gold alloys. Nevertheless, this first test has shown that Raman

spectroscopy is sensitive to surface alterations and that it can be used, in the future, for the

characterisation of the corrosion products of gold alloys, when the corrosion products

developed on gold alloys submitted to atmospheric corrosion will be fully identified. A Raman

spectra database can then be created based on spectra obtained from highly pure and certified

chemical compounds.

Appendix 1 – Characterisation of gold alloy corrosion layer by Raman spectroscopy

138

References

[1] M. Veneranda, M. Irazola, A. Pitarch, M. Olivares, A. Iturregui, K. Castro, J.M. Madariage, In-situ and laboratory Raman analysis in the field of cultural heritage: the case of a mural painting, J. Raman Spectrosc., 45 (2014) 228-237. [2] P. Colomban, The on-site/remote Raman analysis with mobile instruments: a review of drawbacks and success in cultural heritage studies and other associated fields, J. Raman Spectrosc. 43 (2012) 1529-1535. [3] M.L. Franquelo, A. Duran, L.K. Herrera, M.C. Jimenez de Haro, J.L. Perez-Rodriguez, Comparison between micro-Raman and micro-FTIR spectroscopy techniques for the characterization of pigments from Southern Spain Cultural Heritage. J. Mol. Struct. 924–926 (2009) 404-412. [4] V. Hayez, V. Costa, J. Guillaumen, H. Terryn, A. Hubin, Micros Raman spectrosocpy used for the study of corrosion products on copper alloys: study of the chemical composition of artificial patinas used for restoration purposes, Analyst 130 (2005) 550-556. [5] N. Yucel, A. Kalkanli, E.N. Caner-Saltik, Investigation of atmospheric corrosion layers on historic iron nails by micro-Raman spectrosocopy, J. Raman Spectrosc. 47 (2016) 1486-1493. [6] D. Neff, S. Reguer, L. Bellot-Gurlet, P. Dillmann, R. Bertholon, Structural characterization of corrosion products on archaeological iron: a integrated analytical approach to establish corrosion forms, J. Raman Spectrosc. 35 (2004) 739-745. [7] V. Hayez, J. Guillaume, A. Hubin, H. Terryn, Micro Raman spectroscopy for the study of corrosion products on copper alloys: setting up of a reference database and studying works of art, J. Raman Spectrosc 35 (2004) 732-738. [8] I. Martina, R. Wiesinger, M. Schreiner, Micro-Raman investigations of early stage silver corrosion products occurring in sulfur containing atmospheres, J. Raman Spectrosc. 44 (2013) 770-775. [9] J.A.N.T. Soares, Introduction to optical characterization of metals in Practical Materials characterization, (M. Sardela ed.), Springer (2014), pp. 43-92. [10] I. Martina, R. Wiesinger, D. Jembrih-Simbürger, M. Schreiner, Micro-Raman characterization of silver corrosion products: Instrumental set up and reference database, e-Preservation Science 9 (2012) 1-8. [11] J. Kreisel, M.C. Weber, N. Dix, F. Sánchez, P.A. Thomas, J. Fontcuberta, Probing individual layers in functional oxide multilayers by wavelength-dependent Raman scattering, Adv. Funct. Mater., 22 (2012) 5044-5049. [12] G. Marucci, A. Monno, I.D. van der Werf, Non-invasive micro-Raman spectroscopy for investigation of historical silver salt gelatin photographs, Microchemical Journal, 117 (2014) 220-224. [13] Y. Deng, A.D. Handoko, Y. Du, S. Xi, B.S. Yeo, In situ Raman spectroscopy of copper and copper oxide surfaces during electrochemical oxygen evolution reaction: identification of CuIII oxides as catalytically active species, ACS Catal. 6 (2016) 2473-2481.

Appendix 2

Analysis of a gold alloy corrosion layer

by PIXE and RBS

Appendix 2 – Analysis of a gold alloy corrosion layer by PIXE and RBS

141

App. 2 Analysis of a gold alloy corrosion layer by PIXE and RBS

Particle Induced X-ray Emission (PIXE) and Rutherford Backscattering Spectrometry (RBS) are

two non-invasive ion beam analysis (IBA) techniques currently applied in the field of cultural

heritage for multi-elemental analysis of the near surface (from about 3 to 20 μm according to

the materials analysed) [1].

PIXE is based on the same process as XRF and EDS, but uses an incident beam of charged

particles, that are for the study of gold alloys often protons with an energy ranging from 2 to 3

MeV [2,3]. PIXE is highly sensitive due to the efficient X-ray production and the low background

in particular for low atomic numbers, which decreases the detection limits [1]. The energy of the

incident beam can be varied to improve the determination of selected elements. RBS is based

on the electrostatic interaction between the incident ion, usually protons or alpha particles, and

the nucleus. At a given scattering angle, the energy of the elastically scattered ions is related to

the mass of the nucleus of the sample atom. By measuring the energy loss of the particles after

the interaction with the sample it is possible to determine its structure and composition [4].

Both IBA techniques are available external beams allowing their application to various cultural

heritage objects.

Considering the different depth information provided by PIXE and RBS, their combination could

give in the case of gold alloys the elemental distribution of the elements present at different

depths, from the very first surface to the gold alloy substrate. For this reason, these techniques

were tested to define whether they can usefully integrate the analytical approach that we

proposed to develop for the characterisation of the corrosion layer of gold alloys.

PIXE and RBS analyses were carried out at the van de Graaff accelerator AGLAE – Accélérateur

Grand Louvre d’analyse élémentaire (Centre de Recherche et Restauration des Musées de

France, C2RMF, Palais du Louvre, Paris). The raw data collected was analysed by deconvolution

of the PIXE spectra with the GUPIX software. The accuracy of the obtained quantitative was

validated by the analysis of home-made gold standards. RBS data was analysed with SIMNRA

program [5].

One gold alloy (82 wt% Au, 16 wt% Ag, 2 wt% Cu) sample corroded in sulphide containing

atmosphere1 was used in this test. Although the AGLAE setup allows to carry out simultaneous

1 The corrosion test was carried out following the corrosivity test protocol described in Chapter 3 (section 3.2.2).

Appendix 2 – Characterisation of gold alloy corrosion layer by PIXE and RBS

142

PIXE and RBS measurements, in this first test data were acquired in sequence with acquisition

times ranging from 1 to 15 minutes. PIXE was carried out with a 3 MeV proton beam and RBS

with a 3 MeV alpha particles beam. The beam was extracted in air and for RBS a flow of helium

gas was provided over the sample to minimise particle energy degradation and straggling

induced by the atmosphere [4]. For PIXE measurements a multi–SDD detector set-up [6] was

used and for RBS a surface barrier detector [7].

As the presence of S-based compounds was expected to be detected on the sample surface, the

measurements by PIXE were carried out at different angles of the incident beam (90°, 100°, 120°

and 130°) in relation to the sample to obtain information on the first layers (Figure app2.1).

Figure app2.1 – Variable angle sample holder with a gold sample in front of the extracted proton beam for analysis of the same sample at different angles.

Figure app2.2 shows the two PIXE spectra obtained for the corroded gold alloy using incident

angles of 90° and 100°. As it can be depicted from the spectra, although there is a better

definition of the peaks at lower energies with the grazing incidence, the background in the

energy region of the K-lines of sulphur is still high.

As expected, data obtained by PIXE did not clearly identify the presence of sulphur contained in

the analysed corrosion products. In fact, although PIXE can be used to identify low-Z elements,

the absorption of the low energy characteristic X-rays of these elements into thick samples

restricts the use of PIXE for quantitative determination [8]. In addition, the presence of the Au-

Mα lines close to the S-K lines, as shown in figure app2.3, makes the data deconvolution complex

hindering the quantification of S.


143

Figure app2.2 – PIXE spectra obtained for the corroded Au alloy using incident angles of 90° and 100°.

Figure app.2.3 – PIXE spectra obtained for the non-corroded and corroded Au alloy (3 MeV proton beam, 90° incident angle), showing that is impossible to clearly identify the presence of S, (Kα1 2.308 keV ; Kα2 2.307 keV; Kβ 2.464 keV ), due to the close energies of the Au M-Lines (Mα1 2.142 keV ; Mα2 2.133 keV; Mβ 2.220 keV; Mƴ 2.404 keV).


144

RBS spectra of the corroded gold alloy were analysed using the simulation code SIMNRA. This

simulation produces a distribution of the elements, in atomic concentration, as a series of

discrete layers, whose thickness is expressed in terms of number of atoms per square centimetre

(at.cm-2) [2]. The analysis was conditioned by the lack of knowledge on the corrosion process

and the type of corrosion compounds. Thus, a simple two layer model was chose: an outer layer

composed of AuAgS on the gold alloy substrate. The spectral simulation, shown in figure app.2.4,

indicated for the selected model the presence of a corrosion layer circa 100 nm thick. However,

data is unreliable because it was impossible to define the accuracy of this simulation, as a result

of the error introduced by the high roughness of the sample.

Figure app.2.4 – RBS experimental and simulated spectra for the corroded gold alloy (data obtained with a 3.1 MeV proton incident beam).

Considering the unsatisfactory results obtained with PIXE and RBS for the characterisation of the

corroded gold alloy, it was decided not to select these IBA techniques to integrate the analytical

protocol.


145

References

[1] T. Calligaro, J.-C- Dran, J. Salomon, Ion beam microanalysis in in Non-destructive microanalysis of Cultural Heritage materials vol XLII (K. Janssens, R. Van Grieken eds.) Elsevier, Netherlands (2004) pp. 227‒276. [2] V. Corregidor, L.C. Alves, P.A. Rodrigues, M. Vilarigues, R.C. Silva, The external ion beam facility in Portugal for studying cultural heritage, e-conservation 22 (2011) 41‒52. [3] J. Salomon, J.-C. Dran, T. Guillou, B. Moignard, L. Pichon, P. Walter, F. Mathis, Present and future role of ion beam analysis in the study of cultural heritage materials: the example of the AGLAE facility, Nucl. Instrum. Meth. B 266 (2008) 2273‒2278. [4] E. Darque-Ceretti, D. Hélary, M. Aucouturier, An investigation of gold/ceramic and gold/glass interfaces, Gold Bull. 35 (2002) 118‒129. [5] M. Mayer, SIMNRA, a simulation program for the analysis of NRA, RBS and ERDA, AIP Conf. Proc. 475 (1998) 541‒544. [6] L. Pichon, B. Moignard, Q. Lemasson, C. Pacheco, P. Walter, Development of a multi-detector and a systematic imaging system on the AGLAE external beam, Nucl. Instrum. Meth. B 318 (2014) 27‒31. [7] L. Beck, L. Pichon, B. Moignard, T. Guillou, P. Walter, IBA techniques: examples of useful combinations for the characterisation of cultural heritage materials, Nucl. Instrum. Meth. B 269 (2011) 2999‒3005. [8] G. Demortier, Targeting ion beam analysis techniques for gold artefacts, ArcheoSciences 33 (2009) 29‒38.


146

147

Publications resulting from this work

Published

I. Tissot, O.C. Monteiro, M.A. Barreiros, J. Correia, M.F. Guerra, The influence of the constituent elements on the corrosion mechanisms of silver alloys in sulphide environments: the case of sterling silver, RSC Adv. 7 (2017) 28564-28572.

I. Tissot, O.C. Monteiro, M.A. Barreiros, M.F. Guerra, Silver-copper alloy corrosion mechanisms – New perspectives for the conservation assessment of cultural heritage objects. ICOM-CC 18th Triennial Conference Preprints, Copenhagen, 4-8 September 2017, ed. J. Bridgland, art. 1611. Paris: International Council of Museums.

I. Tissot, O.C. Monteiro, M.A. Barreiros, V. Corregidor, J. Correia, M.F. Guerra, Corrosion of silver alloys in sulphide environments: a multianalytical approach for surface characterisation, RSC Adv. 6 (2016) 51856‒51863.

M.F. Guerra, I. Tissot, Bronze Age and Iron age gold torcs and earrings from the Iberian Atlantic façade: a non-invasive multi-analytical approach to the characterisation of the alloys and the corrosion, X-ray Spectrom. 45 (2016) 5‒13.

L. Troalen, I. Tissot, M. Maitland, M.F. Guerra, Jewellery of a young Egyptian girl: Middle Kingdom gold work from Haraga tomb 72, Historical Metallurgy 49 (2015) 75‒62.

I. Tissot, L.G. Troalen, M. Manso, M. Ponting, M. Radtke, U. Reinholz, M.A. Barreiros, M.L. Carvalho, M.F. Guerra, A multi-analytical approach to gold in Ancient Egypt: studies on provenance and corrosion, Spectrochim. Acta Part B, 108 (2015) 75‒82.

I. Tissot, M. Tissot, M.F. Guerra, Atmospheric corrosion in museum context – the case of the Treasure Room from the National Archaeology Museum, Lisbon, Corros. Prot. Mater. 33 (2014) 73‒77.

I. Tissot, M. Tissot, M. Manso, L.C. Alves, M.A. Barreiros, T. Marcelos, M.L. Carvalho, V. Corregidor, M.F. Guerra, The earrings of Pancas Treasure: Analytical study by X-ray based techniques – a first approach, Nucl. Instrum. Meth. B 306 (2013) 236‒240.

M.F. Guerra, I. Tissot, The role of nuclear microprobes in the study of technology, provenance and corrosion of cultural heritage: The case of gold and silver items, Nucl. Instrum. Meth. B 306 (2013) 227‒231.

A Ourivesaria pré-Histórica do Ocidente Peninsular Atlântico Compreender para Conservar (M.F. Guerra M.F., I. Tissot eds.), AuCORRE, Lisboa (2014) ISBN 978-898-20-4470.

Accepted for publication

I. Tissot, O. C. Monteiro, M. A. Barreiros, M. F. Guerra, Atmospheric corrosion of patinated silverwork: a conservation challenge, Corros. Prot. Mater. (2017)

148

I. Tissot, M.F. Guerra, Composition of Bronze Age gold bracelets from the Portuguese area: in Archaeometallurgy in Europe IV (Ignacio Montero et Alicia Perea, eds.) Collection Bibliotheca Praehistorica Hispana, vol XXXIII, Editorial CSIC, Madrid. (2017)

I. Tissot, B. Comendador, M. F. Guerra Estudo fisico-quimico dun conxunto de objectos de ourivaria protohistorica da coleccion do Museo Provincial de Lugo: in Catalogo da coleccion do Museo Provincial de Lugo (A. Balseiro Garcia ed.) Servicio publicaciones, Deputacion Provincial Lugo (2017)

I. Tissot, M.F. Guerra, Investigating burial and atmospheric corrosion of Egyptian Gold objects in Ancient Egyptian Gold: Archaeology and Science (M.F. Guerra, M. Martinón-Torres, S. Quirke eds.) Part IV, Archaeopress Archaeology, Oxford (2017).

Documents

Analytical strategies for the identification and characterisation of thin … › bitstream › 10451 › 35143 › 1 › ulsd... · 2018-10-31 · Analytical strategies for the identification