The "sword" episode of The Curse of Oak Island aired in Canada last night, and Dr. Christa Brosseau is now free to discuss her analysis and results. In the period of time between now and when those of us in the U.S. got to see the episode, I've seen many incorrect (and, frankly, insulting) statements about Brosseau's work coming from those who have staked all their credibility on the sword being an ancient Roman artifact. We owe her a "thank you" for participating in this endeavor.
I am very happy that Brosseau has agreed to let me publish this statement here. I'm sure that many people who will read this are not experts in metals analysis. Brosseau is. Even if you don't understand every detail of this summary, take note of the manner in which she describes the rationale for the tests she undertook, her methods, the equipment that she used, and her results. That's what professionals do.
Brosseau tells me that she has been following this blog with interest, and will try to answer questions posed to her here. Have a question about the analysis after reading what she has to say? Fire away. I especially welcome serious questions from people who still "believe" in the sword. Insulting nonsense comments will be deleted.
Here is what Brosseau sent me:
I am very happy that Brosseau has agreed to let me publish this statement here. I'm sure that many people who will read this are not experts in metals analysis. Brosseau is. Even if you don't understand every detail of this summary, take note of the manner in which she describes the rationale for the tests she undertook, her methods, the equipment that she used, and her results. That's what professionals do.
Brosseau tells me that she has been following this blog with interest, and will try to answer questions posed to her here. Have a question about the analysis after reading what she has to say? Fire away. I especially welcome serious questions from people who still "believe" in the sword. Insulting nonsense comments will be deleted.
Here is what Brosseau sent me:
Executive summary:
A metal sword was brought in for chemical testing during the summer of 2015. Objectives included determining if the sword was bronze or brass, and whether or not the chemical constituents could be ascertained, both qualitatively and quantitatively, in an effort to date the artifact. Both elemental and molecular testing was conducted. The chosen tests included scanning electron microscopy (SEM), energy dispersive x-ray spectroscopy (EDS), confocal Raman microscopy and surface enhanced Raman spectroscopy (SERS). Multiple samples of both the patina, as well as the base metal were taken from the tip of the sword. A sample of the base metal was also taken from the hilt to test the alloy composition. Two samples of an “added layer”, present on the front of the hilt was also taken for analysis. After chemical testing and analysis, the results conclude that this object represents a modern leaded brass with an artificial patina with a likely earliest manufacture date of ~1880 AD. The added layer is a lead oxide putty, applied for an unknown reason.
Introduction:
The determination of the chemical composition of copper alloys can be important in the dating of art historical objects formed from such alloys. Brass (Cu/Zn alloy) and bronze (Cu/Sn alloy) constitute the vast majority of copper alloys of interest since antiquity. In this analysis, two questions were put forward: (1) Is this sword a bronze or a brass and (2) Can the elemental composition of the base metal and/or the chemical constitution of the patina offer any clues as to the age of the object. The answer to the first question is readily provided in a standard elemental analysis. A zinc content above trace levels, and the absence of a significant amount of tin would suggest a brass. For brasses, the zinc content is a powerful indicator of chronology.1 The first functional brasses were produced by the Romans, and were prepared using a process wherein a zinc ore was reacted with copper metal at high temperatures. This smelting process occurs at temperatures of at least 1100°C, which is above the boiling point of metallic zinc. Thus, most of the zinc is lost to evaporation, putting a thermodynamic limit on how much zinc can be incorporated into a brass using this method. This upper limit is widely agreed upon to be 28%, based on the analysis of a multitude of early brasses, as well as more recent experiments aimed at recreating this early method.1 Therefore a brass that is being put forward as ancient which contains a zinc content greater that 28% is suspicious. This is especially true if the alloy has greater than trace amounts of lead or tin, which inhibit absorption of zinc in the alloy. In post-Medieval Europe, a more sophisticated cementation method was developed, which allowed the incorporation of more zinc into the alloy, up to 33%.1 Hence, dating from about the sixteenth century, some brasses have been found containing up to 33% zinc. Zinc content above this level is reflective of a more modern brass, produced using a more modern direct method (speltering), the patent for which was taken out in 1738 in Britain.2 The older cementation method was largely abandoned by the mid-1800s.
While the zinc content is the most important element used to date brasses, other factors will be important as well, including a determination of the relative purity of the alloy (which might suggest a modern refinement) and the presence of modern pigments and/or resins will also be key indicators of modern manufacture.
Methodology:
The following chemical analysis tools were utilized in this study:
1. Scanning Electron Microscope (SEM) – TESCAN MIRA 3 LMU Variable Pressure Schottky Field Emission Scanning Electron Microscope
This electron microscope provides micro and nanoscale resolution of a sample. Features such as slag in the metal and grain boundaries in the patina can be elucidated.
2. Energy Dispersive X-ray Spectroscopy (EDS) – INCA X-max 80mm 2 EDS equipped with silicon drift detection (SDD)
The EDS unit is coupled to the SEM, and allows one to determine the elemental make- up of a particular region in the sample down to ppm levels. In this method, the sample is bombarded with a high energy electron beam, which excites core electrons in the sample. Upon decaying back to the ground state, energy is released in the form of x-ray radiation, which is monitored. The energy of these emitted x-rays is very specific to an individual element, and thus an elemental map of the sample can be obtained. This technique is akin to x-ray fluorescence (XRF), with a few notable differences. XRF can be used in a portable fashion, but this only allows for investigation of the surface of an object. In addition, portable XRF typically lacks the sensitivity and accuracy of EDS. When destructive analysis is possible, EDS is preferred. This will be used to determine the elemental composition of both the base metal and the patina.
3. Confocal Raman Spectroscopy – This instrument combines a confocal microscope with a Raman spectrometer, and provides a vibrational signature of the sample, such that the molecular species can be identified. This technique is based on light scattering. This will be used to evaluate chemical components of the patina that are Raman active, most notably any additives, resins or dyes/pigments.
4. Surface Enhanced Raman Spectroscopy (SERS) – This technique uses a thin layer of noble metal nanoparticles (Ag) over the surface of the sample, such that a much greater Raman signal can be obtained. This technique is particularly useful for samples which exhibit intense fluorescence, which can quench the Raman signal, or where the concentration of analyte is very low. This will be used to evaluate chemical components of the patina.
Sampling and preparation of samples for analysis:
For this analysis, several samples of the sword were collected for analysis. In each case the sample size was ~1-10 mg.
Tip of sword: base metal filings, patina flakes
Hilt of sword: base metal filings, patina flakes
Hilt of sword: added layer of unknown origin (this was applied to a few areas of the sword, for an unknown reason)
The patina flakes and the added layer were removed directly using tweezers. For the base metal, a small part of the sword was polished free of patina to expose the un-corroded metal, cleaned with solvent, and then several filings were taken using a steel file. Samples of the base metal of the hilt were taken at a later date in order to confirm the sword was one continuous metal.
For the SEM-EDS the samples were prepared by the SEM technician on carbon tape studs. For Raman spectroscopy, the samples were analyzed directly on glass coverslips. For SERS, the samples on microscope slides were treated with a layer of silver nanoparticles prepared in house according to an established protocol.3
Results and Discussion
Metallurgical analysis of the base metal:
SEM-EDS analysis was conducted on the base metal filings. It was noted that the zinc content in this alloy is very high, at 35% ± 1% at 95% confidence, and thus is representative of a modern brass. This places the earliest manufacture date for the piece at 1738. In addition, the lead content is very high (2.6% ±0.7). Closer examination of the metal showed bright regions, referred to as slag. In this region the heavy metals (Pb, As, Sb, etc.) are typically co-localized. EDS analysis of the slag however showed only lead, suggesting that the lead is of high purity, and thus likely refined. Also, the copper itself is quite pure, with only very trace amounts of contaminants (As (one sample only), Sn, etc.) detected. Thus the copper itself is also likely refined. Electrowinning of metals was first demonstrated in 1847, and the practice was first patented in 1865.4 Commercial plants for electrowinning of copper ore existed in 1870 in Wales and in the early 1880s in the USA.5 Thus, the likely actual earliest manufacture date for this brass is 1870-1880. This sword constitutes a modern leaded brass. The lead would have been added to facilitate casting and to improve the strength of the alloy, and also to reduce corrosion.
Metallurgical analysis of the added layer:
In several places on the sword, an added layer was noted. This layer had a metallic lustre and was putty-like in nature. Results confirm this as a lead oxide, likely a lead white putty. The reason for this added layer is unclear; it may have been added to simulate a lead overlay or to hide an area where a defect in casting was present.
Morphological and Elemental analysis of the patina:
The patina was examined using SEM-EDS. This allows one to monitor the morphology of the sample, and obtain the chemical composition for a desired area of interest. While the samples were not prepared in cross-section, it is still evident that the patina is uniform in structure and morphology, which points to an artificial rather than natural patination process. Also, the patina was very easily removed, which also suggests an artificial patination. The patina in general, is fairly complex, however the chemical composition is dominated by copper, oxygen and chloride. This suggests the patina is mostly a copper hydroxychloride. The source of the chloride would be from either salt water corrosion, or from chemical patination using chemical reagents, for example chloride salts such as CuCl2 or NH4Cl. Since the extent of corrosion is limited, the latter explanation is the most likely. In addition, the patina contains iron and sulfur, which may also indicate chemical patination, through the use of CuSO4, Fe(NO3)3 or Na2S2O3, all commonly used reagents for brass patinas. In order to further evaluate the patina further at the molecular level, Raman spectroscopy and SERS was performed.
Molecular Analysis of the Patina:
The patina sample was evaluated using Raman spectroscopy, to ascertain the chemical make-up of the patina. Raman spectroscopy is particularly sensitive to substances which have the ability to scatter light well, through the presence of heavy atoms, or through the presence of electrons which are delocalized. Thus, it is an excellent method for the analysis of both inorganic and organic pigments. In this case, we were interested in determining whether any pigments might have been used to produce the patina. In several places, the recorded Raman signal was an exact match for Prussian Blue, a modern blue pigment. Prussian blue was first available for use in Europe in the early to mid-1700’s, and was in extensive use for the next several hundred years. The use of Prussian blue for the restoration of historical bronzes has been noted in several conservation publications, including the restoration of bronzes from the Fonderia Chiurazzi in Naples.5 The identification of Prussian blue in the patina strongly suggests that the patina is of modern origin, post-1734 AD, and was produced artificially.
To conduct further analysis, SERS was performed. In this method, a thin layer of silver nanoparticles is applied to the surface of the sample, such that the Raman signal can be increased significantly. The SERS data identified multiple pigments in the same sample. Prussian blue was again identified, along with lead white and what appears to be a good match for Cerulean blue. Cerulean blue was first available for use in Europe in 1821. In addition, strong peaks in the region of 1500 cm-1 suggest the presence of a synthetic organic pigment, likely a yellow azo dye, which would be of 19th or 20th century manufacture. Peaks at 2800-3000 cm-1 due to the C-H stretching vibrations of organic materials are also suggestive of this. The likely presence of Cerulean blue and an unknown synthetic organic pigment suggest the sword was manufactured post-1820 AD.
Conclusions:
The Oak Island sword represents an out-of-place artifact with no established provenance. In this case, scientific analysis of the object can aid in answering some fundamental questions regarding the piece, such as the nature and composition of the alloy.
In this case, it was determined that the alleged Roman sword is in fact a modern leaded brass with an artificial patina. Based on the materials analysis, this sword had to have been manufactured post-1738, and more likely post-1880, based on the materials and process of manufacture used.
References:
1. P. Craddock. “Scientific Analysis of Copies, Fakes and Forgeries”. Elsevier, Oxford, UK, 2009.
2. J. Day. “Copper, Zinc and Brass Production” in J. Day and R. F. Tylecote, The Industrial Revolution in Metals, Institute of Metals, London, 1991.
3. “Surface-Enhanced Raman Spectroscopy: A Direct Method to Identify Colorants in Various Artist Media” C.L. Brosseau, K. Rayner, F. Casadio, C.M Grzywacz, R.P Van Duyne. Analytical Chemistry, 2009, 81(17) p.7443.
4. “Copper Leaching, Solvent Extraction and Electrowinning Technology” Ed. J.V. Jergensen II, 1999.
5. “The Restoration of Ancient Bronzes: Naples and Beyond” Ed. E. Risser and D. Saunders. J. Paul Getty Museum, Getty Publications, 2013.
Additional Notes:
A metal sword was brought in for chemical testing during the summer of 2015. Objectives included determining if the sword was bronze or brass, and whether or not the chemical constituents could be ascertained, both qualitatively and quantitatively, in an effort to date the artifact. Both elemental and molecular testing was conducted. The chosen tests included scanning electron microscopy (SEM), energy dispersive x-ray spectroscopy (EDS), confocal Raman microscopy and surface enhanced Raman spectroscopy (SERS). Multiple samples of both the patina, as well as the base metal were taken from the tip of the sword. A sample of the base metal was also taken from the hilt to test the alloy composition. Two samples of an “added layer”, present on the front of the hilt was also taken for analysis. After chemical testing and analysis, the results conclude that this object represents a modern leaded brass with an artificial patina with a likely earliest manufacture date of ~1880 AD. The added layer is a lead oxide putty, applied for an unknown reason.
Introduction:
The determination of the chemical composition of copper alloys can be important in the dating of art historical objects formed from such alloys. Brass (Cu/Zn alloy) and bronze (Cu/Sn alloy) constitute the vast majority of copper alloys of interest since antiquity. In this analysis, two questions were put forward: (1) Is this sword a bronze or a brass and (2) Can the elemental composition of the base metal and/or the chemical constitution of the patina offer any clues as to the age of the object. The answer to the first question is readily provided in a standard elemental analysis. A zinc content above trace levels, and the absence of a significant amount of tin would suggest a brass. For brasses, the zinc content is a powerful indicator of chronology.1 The first functional brasses were produced by the Romans, and were prepared using a process wherein a zinc ore was reacted with copper metal at high temperatures. This smelting process occurs at temperatures of at least 1100°C, which is above the boiling point of metallic zinc. Thus, most of the zinc is lost to evaporation, putting a thermodynamic limit on how much zinc can be incorporated into a brass using this method. This upper limit is widely agreed upon to be 28%, based on the analysis of a multitude of early brasses, as well as more recent experiments aimed at recreating this early method.1 Therefore a brass that is being put forward as ancient which contains a zinc content greater that 28% is suspicious. This is especially true if the alloy has greater than trace amounts of lead or tin, which inhibit absorption of zinc in the alloy. In post-Medieval Europe, a more sophisticated cementation method was developed, which allowed the incorporation of more zinc into the alloy, up to 33%.1 Hence, dating from about the sixteenth century, some brasses have been found containing up to 33% zinc. Zinc content above this level is reflective of a more modern brass, produced using a more modern direct method (speltering), the patent for which was taken out in 1738 in Britain.2 The older cementation method was largely abandoned by the mid-1800s.
While the zinc content is the most important element used to date brasses, other factors will be important as well, including a determination of the relative purity of the alloy (which might suggest a modern refinement) and the presence of modern pigments and/or resins will also be key indicators of modern manufacture.
Methodology:
The following chemical analysis tools were utilized in this study:
1. Scanning Electron Microscope (SEM) – TESCAN MIRA 3 LMU Variable Pressure Schottky Field Emission Scanning Electron Microscope
This electron microscope provides micro and nanoscale resolution of a sample. Features such as slag in the metal and grain boundaries in the patina can be elucidated.
2. Energy Dispersive X-ray Spectroscopy (EDS) – INCA X-max 80mm 2 EDS equipped with silicon drift detection (SDD)
The EDS unit is coupled to the SEM, and allows one to determine the elemental make- up of a particular region in the sample down to ppm levels. In this method, the sample is bombarded with a high energy electron beam, which excites core electrons in the sample. Upon decaying back to the ground state, energy is released in the form of x-ray radiation, which is monitored. The energy of these emitted x-rays is very specific to an individual element, and thus an elemental map of the sample can be obtained. This technique is akin to x-ray fluorescence (XRF), with a few notable differences. XRF can be used in a portable fashion, but this only allows for investigation of the surface of an object. In addition, portable XRF typically lacks the sensitivity and accuracy of EDS. When destructive analysis is possible, EDS is preferred. This will be used to determine the elemental composition of both the base metal and the patina.
3. Confocal Raman Spectroscopy – This instrument combines a confocal microscope with a Raman spectrometer, and provides a vibrational signature of the sample, such that the molecular species can be identified. This technique is based on light scattering. This will be used to evaluate chemical components of the patina that are Raman active, most notably any additives, resins or dyes/pigments.
4. Surface Enhanced Raman Spectroscopy (SERS) – This technique uses a thin layer of noble metal nanoparticles (Ag) over the surface of the sample, such that a much greater Raman signal can be obtained. This technique is particularly useful for samples which exhibit intense fluorescence, which can quench the Raman signal, or where the concentration of analyte is very low. This will be used to evaluate chemical components of the patina.
Sampling and preparation of samples for analysis:
For this analysis, several samples of the sword were collected for analysis. In each case the sample size was ~1-10 mg.
Tip of sword: base metal filings, patina flakes
Hilt of sword: base metal filings, patina flakes
Hilt of sword: added layer of unknown origin (this was applied to a few areas of the sword, for an unknown reason)
The patina flakes and the added layer were removed directly using tweezers. For the base metal, a small part of the sword was polished free of patina to expose the un-corroded metal, cleaned with solvent, and then several filings were taken using a steel file. Samples of the base metal of the hilt were taken at a later date in order to confirm the sword was one continuous metal.
For the SEM-EDS the samples were prepared by the SEM technician on carbon tape studs. For Raman spectroscopy, the samples were analyzed directly on glass coverslips. For SERS, the samples on microscope slides were treated with a layer of silver nanoparticles prepared in house according to an established protocol.3
Results and Discussion
Metallurgical analysis of the base metal:
SEM-EDS analysis was conducted on the base metal filings. It was noted that the zinc content in this alloy is very high, at 35% ± 1% at 95% confidence, and thus is representative of a modern brass. This places the earliest manufacture date for the piece at 1738. In addition, the lead content is very high (2.6% ±0.7). Closer examination of the metal showed bright regions, referred to as slag. In this region the heavy metals (Pb, As, Sb, etc.) are typically co-localized. EDS analysis of the slag however showed only lead, suggesting that the lead is of high purity, and thus likely refined. Also, the copper itself is quite pure, with only very trace amounts of contaminants (As (one sample only), Sn, etc.) detected. Thus the copper itself is also likely refined. Electrowinning of metals was first demonstrated in 1847, and the practice was first patented in 1865.4 Commercial plants for electrowinning of copper ore existed in 1870 in Wales and in the early 1880s in the USA.5 Thus, the likely actual earliest manufacture date for this brass is 1870-1880. This sword constitutes a modern leaded brass. The lead would have been added to facilitate casting and to improve the strength of the alloy, and also to reduce corrosion.
Metallurgical analysis of the added layer:
In several places on the sword, an added layer was noted. This layer had a metallic lustre and was putty-like in nature. Results confirm this as a lead oxide, likely a lead white putty. The reason for this added layer is unclear; it may have been added to simulate a lead overlay or to hide an area where a defect in casting was present.
Morphological and Elemental analysis of the patina:
The patina was examined using SEM-EDS. This allows one to monitor the morphology of the sample, and obtain the chemical composition for a desired area of interest. While the samples were not prepared in cross-section, it is still evident that the patina is uniform in structure and morphology, which points to an artificial rather than natural patination process. Also, the patina was very easily removed, which also suggests an artificial patination. The patina in general, is fairly complex, however the chemical composition is dominated by copper, oxygen and chloride. This suggests the patina is mostly a copper hydroxychloride. The source of the chloride would be from either salt water corrosion, or from chemical patination using chemical reagents, for example chloride salts such as CuCl2 or NH4Cl. Since the extent of corrosion is limited, the latter explanation is the most likely. In addition, the patina contains iron and sulfur, which may also indicate chemical patination, through the use of CuSO4, Fe(NO3)3 or Na2S2O3, all commonly used reagents for brass patinas. In order to further evaluate the patina further at the molecular level, Raman spectroscopy and SERS was performed.
Molecular Analysis of the Patina:
The patina sample was evaluated using Raman spectroscopy, to ascertain the chemical make-up of the patina. Raman spectroscopy is particularly sensitive to substances which have the ability to scatter light well, through the presence of heavy atoms, or through the presence of electrons which are delocalized. Thus, it is an excellent method for the analysis of both inorganic and organic pigments. In this case, we were interested in determining whether any pigments might have been used to produce the patina. In several places, the recorded Raman signal was an exact match for Prussian Blue, a modern blue pigment. Prussian blue was first available for use in Europe in the early to mid-1700’s, and was in extensive use for the next several hundred years. The use of Prussian blue for the restoration of historical bronzes has been noted in several conservation publications, including the restoration of bronzes from the Fonderia Chiurazzi in Naples.5 The identification of Prussian blue in the patina strongly suggests that the patina is of modern origin, post-1734 AD, and was produced artificially.
To conduct further analysis, SERS was performed. In this method, a thin layer of silver nanoparticles is applied to the surface of the sample, such that the Raman signal can be increased significantly. The SERS data identified multiple pigments in the same sample. Prussian blue was again identified, along with lead white and what appears to be a good match for Cerulean blue. Cerulean blue was first available for use in Europe in 1821. In addition, strong peaks in the region of 1500 cm-1 suggest the presence of a synthetic organic pigment, likely a yellow azo dye, which would be of 19th or 20th century manufacture. Peaks at 2800-3000 cm-1 due to the C-H stretching vibrations of organic materials are also suggestive of this. The likely presence of Cerulean blue and an unknown synthetic organic pigment suggest the sword was manufactured post-1820 AD.
Conclusions:
The Oak Island sword represents an out-of-place artifact with no established provenance. In this case, scientific analysis of the object can aid in answering some fundamental questions regarding the piece, such as the nature and composition of the alloy.
In this case, it was determined that the alleged Roman sword is in fact a modern leaded brass with an artificial patina. Based on the materials analysis, this sword had to have been manufactured post-1738, and more likely post-1880, based on the materials and process of manufacture used.
References:
1. P. Craddock. “Scientific Analysis of Copies, Fakes and Forgeries”. Elsevier, Oxford, UK, 2009.
2. J. Day. “Copper, Zinc and Brass Production” in J. Day and R. F. Tylecote, The Industrial Revolution in Metals, Institute of Metals, London, 1991.
3. “Surface-Enhanced Raman Spectroscopy: A Direct Method to Identify Colorants in Various Artist Media” C.L. Brosseau, K. Rayner, F. Casadio, C.M Grzywacz, R.P Van Duyne. Analytical Chemistry, 2009, 81(17) p.7443.
4. “Copper Leaching, Solvent Extraction and Electrowinning Technology” Ed. J.V. Jergensen II, 1999.
5. “The Restoration of Ancient Bronzes: Naples and Beyond” Ed. E. Risser and D. Saunders. J. Paul Getty Museum, Getty Publications, 2013.
Additional Notes:
- The composition of the base alloy is most closely in line with a free machining brass. (360 brass = ~61% Cu, 35% Zn, 0.45% Fe, 3% Pb). No Mg or Si was present in the base metal.
- It is likely that this reproduction originated in Naples, or somewhere nearby, given the presence of Prussian Blue in the patina. Someone familiar with restoration of Italian bronzes may have been involved in its production.
- The hilt was sampled after the blade. This was because if the object had been authentic, taking a sample for analysis from the much more detailed hilt would be undesirable. Once the sword was found to be of modern origin however, then sampling from the hilt was completed. As mentioned above, both the blade and the hilt were found to have an identical composition.
- All tests conducted on this piece are state-of-the art analysis methods, not “out-dated” methods, as suggested by a certain individual.
- During the filming of this segment, it was the end of July, and the lab AC was broken. I may have appeared red during filming for the simple reason that I was about to faint from the heat! Also, I may have appeared nervous because I am a scientist, not an actor. Being on television isn’t really my gig.
- All persons in the lab must wear appropriate PPE in the lab (glasses, coats, etc.). That is standard practice in chemistry labs, and hence why I asked them to wear the gear during the shoot. The PPE was certainly not television props.
- Anyone interested in knowing anything further can contact me directly and I will be very happy to discuss the work with them. I am currently considering preparing this work for peer-reviewed publication. Stay tuned!
RSS Feed
