Technical Appendix: Compositions of Enamels on the George Watch

Technical Appendix: Compositions of Enamels on the
George Watch
MARK T. WYPYSKI
AssociateResearchScientist, ShermanFairchild CenterforObjectsConservation, The MetropolitanMuseum of Art
Chemical analyses were done of six different enamels
from the George watch in the collection of The
Metropolitan Museum of Art (acc. no. 17.190.1475)
for comparison with other reported Renaissanceperiod enamel compositions. Analyses using an energy
dispersive X-ray spectrometer (EDS) and scanning
electron microscope (SEM) were done of four different translucent enamels and of two opaque enamels
from the watch's back cover and dial.
Enamels are made of glass, either produced
specifically for the purpose or reused from other
objects, which is fused in place onto a metal substrate.
Studies of European late medieval and Renaissance
enamel compositions have revealed that there was a
change in the general compositions of the glass used
for enameling beginning about the early fourteenth
century, apparently in connection with the rise in the
use of translucent enamels on gold and silver substrates.1 Most known enamels dating from the thirteenth and early fourteenth centuries have been found
to have soda-glass compositions (a glass mixture with
soda [sodium oxide] as the dominant alkali) containing relatively large amounts of potassium, magnesium,
and calcium oxides.2 However, beginning about the
early fourteenth century and continuing into the
fifteenth, most of the translucent enamels analyzed
can be characterized as having mixed-alkali compositions, that is, compositions with relatively high levels of
both sodium and potassium. These mixed-alkali glasses are usually associated with rather low levels of aluminum, magnesium, and calcium. Magnesium and
calcium serve as the main stabilizing agents for glass
compositions, and the extremely low levels of these elements in many mixed-alkali enamel compositions can
account for the poor condition of many enamels from
this period. Little or no lead is generally found in these
enamels, except for the opaque enamels, where it is
associated with the addition of crystalline tin oxide as
a white opacifying agent, or in opaque yellow and
green enamels, where it is associated with a yellow colorant/opacifier such as lead-tin yellow. Most mixedThe notes for this article begin on page 15 2.
alkali enamels from the fourteenth and fifteenth centuries were found to have a sodium-to-potassium ratio
of approximately one-to-one, or else to contain an
excess of potassium over sodium.
By the end of the fifteenth century, however, some
enamels were beginning to be used that contained
much more sodium than potassium. These enamels
may be more accurately described as having soda-glass
compositions with relatively large amounts of potassium, rather than being mixed-alkali. Like most of the
enamels from the fourteenth and fifteenth centuries,
these soda-glass enamels continued to have relatively
low levels of aluminum, magnesium, and calcium.
Some differences are seen between the red enamels
and most other translucent enamels. In red enamels,
the use of mixed-alkali compositions with relatively
more potassium than sodium apparently persisted well
into the sixteenth century or possibly even the seventeenth. And red enamels, unlike most translucent
enamels from this period, also often contain relatively
high levels of magnesium and calcium. (Unpublished
analyses performed at The Metropolitan Museum of
Art of fifteenth-sixteenth-century northern Italian
enameled copper vessels and Limoges sixteenthcentury painted enamel plaques show that at least
some translucent enamels from this period also have
relatively high levels of magnesium and calcium,
although mostly on copper substrates rather than gold
or silver.) These types of enamels appear to have been
in use throughout the Renaissance period and may
have continued in use until as late as the early nineteenth century. Enamels dated to the second half of
the nineteenth century and later, including some
attributed to the Vasters and Castellani workshops,
have been found to have decidedly different compositions, usually lead-potash or lead-alkali compositions,
with some different colorants and opacifiers than
those found in earlier compositions.3
Four translucent enamels from the George watch
were examined. Quantative analyses, reported in Table
1, were performed on samples of the green, blue, and
red enamels from the back cover.4Non-destructive surface analyses were done on the yellow as well as on the
150
The Metropolitan Museum of Art
is collaborating with JSTOR to digitize, preserve, and extend access to
Metropolitan Museum Journal
®
www.jstor.org
TABLE 1. QUANTITATIVE
ANALYSES
(WEIGHT
Translucent Translucent Translucent
red
blue
green
-
Sodium-Na2o
Magnesium-MgO
Aluminum-A1203
Silicon-SiO2
Sulfur-SO3
Chlorine-C1
Potassium-K20
Calcium-CaO
Titanium-TiO2
Manganese-MnO
Iron-FeO03
Cobalt-CoO
Nickel-NiO
15.9
o.3
o.6
71.5
0.5
1.3
3.1
0.5
nd
nd
0.5
o.6
0.2
Copper-CuO
Zinc-ZnO
3.6
nd
Arsenic-As203
Tin-SnO2
Lead-PbO
0.5
nd
nd
Bismuth-Bi2O3
1.1
OF ENAMELS
ANALYSES
PERCENTAGES)
WEIGHT
Element-Oxide
12.2
0.2
2.4
o.7
1.1
66.7
0.3
o.8
60.9
2.7
nd
15.0
5.6
nd
o.3
0.1
1.1
4.2
nd
nd
8.6
nd
nd
nd
nd
nd
0.2
0.8
0.5
nd
nd
0.8
nd
nd
0.5
nd
nd
(APPROXIMATE
PERCENTAGES)
Element-Oxide
-
14.9
EDS SURFACE
TABLE 2. SEMI-QUANTITATIVE
EDS ENAMEL
Sodium-Na2o
Translucent
yellow
I
16
Opaque
white
Opaque
green
10
10
<1
<1
Magnesium-MgO
Aluminum-Al,03
Silicon-SiO2
<1
Sulfur--S03
Chlorine-C1
Potassium-KO0
Calcium-CaO
Titanium-TiO2
<1
40
nd
45
nd
<1
<1
<1
Manganese-MnO
Iron-Fe2O3
Cobalt-CoO
Nickel-NiO
Copper-CuO
Zinc-ZnO
Arsenic-As2O3
Tin-SnO2
Lead-PbO
Bismuth-Bi2O3
1
6o
1
3
1
nd
1
2
2
nd
<1
2
<1
nd
nd
12
<1
<1
nd
nd
nd
nd
1
nd
nd
nd
nd
3
<1
nd
nd
nd
nd
nd
22
20
nd
2
<1
nd
6
30
nd
nd = not detected
Other oxides, such as P205, Cr203, Sb203, and BaO, were also looked for but not detected.
opaque white and green from the watch dial (only
enamels with some previous loss or damage were sampled to avoid compromising undamaged surfaces).
Because of various problems inherent in surface analysis, this type of analysis can only provide approximate,
semi-quantitativeresults, which are reported in Table 2.
The translucent green, blue, and yellow enamels were
found to have soda-glass compositions with relatively
large amounts of potassium and low levels (approximately one percent or less) of magnesium, calcium,
and aluminum oxides. The blue enamel was noted as
exhibiting a greater degree of decomposition than the
other colors, especially the red. This is not surprising in
light of the enamel compositions. The total amount of
the stabilizing elements magnesium and calcium was
found to be less than one percent by weight in the blue,
whereas in an average stable glass it is usually about five
to ten percent. While the green and yellow enamels also
have low percentages of these elements, they contain
large amounts of iron oxide, which can help to improve
the chemical resistance of glass.
The green enamel was found to be colored with
large amounts of both copper and iron oxides, while
the yellow contained a very large amount of iron. The
inclusion of large amounts of metallic colorants has
been documented for many other translucent enamel
compositions. These large amounts of colorants were
apparently required to achieve the desired hue of the
thin translucent layers over the metal. The dark blue
translucent enamel was found to contain a relatively
large amount of cobalt oxide. Cobalt is a rather strong
colorant, and cobalt blue glass is generally found to
contain no more than about two tenths of one percent
of cobalt oxide, while the enamel tested was found to
contain three times this amount. Small amounts of
nickel, arsenic, and bismuth were also found in this
enamel. These elements, especially bismuth, are relatively rare in glass compositions and appear to be associated with the origin of the cobalt ore used to make
this enamel. Cobalt-containing glass and enamels from
the thirteenth, fourteenth, and fifteenth centuries are
usually found to contain small amounts of zinc, apparently from the use of a Syrian cobalt ore source rich in
zinc.5 The fifteenth century saw the widespread
reliance on European cobalt ore sources, such as those
from Saxony which yielded nickel-, arsenic-, and
bismuth-rich ores, for the production of glass, enamel,
and the pigment smalt.
151
The red enamel, unlike the other translucent colors,
was found to have a mixed-alkali composition with
somewhat more potassium than sodium and relatively
high levels of magnesium and calcium. The colorant in
this enamel is a reduced form of copper oxide. A trace
amount of tin oxide was also noted in this enamel.
Traces of tin and lead are often associated with red glass
and enamel, as they apparently act as reducing agents
and help to raise the solubility of the copper oxide.
Surface analyses were also done of two opaque
enamels, a white and a green. An obvious difference
between these enamels and the translucent ones is that
these were found to contain large amounts of lead
oxide. The white enamel also contains a large amount
of crystalline tin oxide, a white colorant and opacifier.
Lead oxide is almost alwaysassociated with tin oxide in
glasses, as it was apparently added to help the conversion of metallic tin to tin oxide. White enamels
opacified with tin oxide were used at least as early as
the end of the twelfth century, although generally with
much smaller amounts of tin and an excess of lead to
tin. White enamels from the fifteenth century and
later, however, have been found to contain much higher percentages of tin, many more than twenty percent
by weight, usually with an approximately one-to-one
ratio of lead oxide to tin. The opaque green enamel
was also found to contain tin. Analyses of some of the
opacifying crystals in the enamel revealed that most if
not all of the tin is present in the form of lead-tin yellow, rather than white tin oxide. Other Renaissance
opaque yellow and green enamels have been found
that contain either lead-tin yellow alone or a mixture
of lead-tin yellow and lead antimonate yellow crystals.
The green color in this enamel was achieved by the
addition of some copper oxide, which by itself produces a blue or turquoise color, to yellow enamel.
Although the information currently available on
Renaissance enamel compositions is somewhat sparse,
all of the enamels examined from the George watch
were found to be entirely consistent in composition
with what is known about enamels dating from the late
fifteenth to the seventeenth centuries. Unfortunately,
based on current research, there appear to be few, if
any, compositional criteria for distinguishing between
Renaissance period enamels and enamels dating from
the eighteenth or early nineteenth century. Ongoing
research at the Metropolitan Museum and elsewhere
will help to shed more light on the different enamel
compositions of these periods.
NOTES
i. M. T. Wypyski and R. W. Richter, "Preliminary Compositional
Study of 14th and 15th c. European Enamels," Techne6 (1997),
pp. 48-57; M. Bimson and I. C. Freestone, "Rouge Clair and
Other Late 14th Century Enamels on the Royal Gold Cup of the
Kings of France and England,"Annalesdu ge CongresdelAssociation
Internationalepour l'Histoiredu Verre,Nancy 1983 (Liege, 1985),
pp. 209-22; M. Schreiner, I. Schmitz, W. Baatz, and B. Campos,
"The Degradation of Enamel on Medieval Silver Objects of the
Kunsthistorisches Museum in Vienna/Austria," The Ceramics
CulturalHeritage,CIMTEC(Florence, 1995), pp. 603-12.
2.I. Biron, P. Dandridge, and M. T. Wypyski, "Techniques and
Materialsin Limoges Enamels," in EnamelsofLimoges:1100-1350,
exh. cat., MMA (New York, 1996), pp. 48-62, 445-50; S. G. E.
Bowman and I. C. Freestone, "Early 14th c. Enamelwork: A
Technical Examination of the Hanap Cover of All Souls College,
Oxford," Techne6 (1997), pp. 41-47.
3. M. T. Wypyski, "A Survey of 16th and 17th Century European
Enamel Compositions" (in preparation).
4. The enamels were analyzed using energy dispersive X-ray spectrometry. The three enamel samples were taken by flaking off very
small pieces, on the order of a cubic millimeter in size or less, with
the use of a steel scalpel. The samples were prepared for analysis
by embedding them in epoxy or polyester resin and grinding with
silicon carbide paper to expose the sample interiors. The cross
sections were then polished with cerium oxide and given a highvacuum carbon coating for conductivity before analysis. Weight
percentages of the elements detected were calculated against wellcharacterized reference glasses. For EDS analysis of glasses, the
relative variation in the calculated percentages for the major element oxides has been determined to be less than two percent for
silicon, about five percent for sodium, potassium, and calcium,
and about ten percent for magnesium, aluminum, and the metals
such as copper, manganese, and iron. The minimum detection
limits for the elements titanium to zinc were found to be under
one tenth of one percent. The minimum detection limits for phosphorus, lead, barium, arsenic, antimony, and tin oxides, however,
were found to be much higher, about one half of a percent by
weight, mainly due to peak overlap problems. For details, see
M. Verita, R. Basso, M. T. Wypyski, and R. J. Koestler, "X-Ray
Microanalysisof Ancient GlassyMaterials:A Comparative Study of
Wavelength Dispersive and Energy Dispersive Techniques,"
Archaeometry
36, no. 2 (1994), pp. 241-51.
5.I. C. Freestone, "Looking into Glass," in Science and the Past
(London, 1991), pp. 37-56; I. Soulier, B. Gratuze, andJ. N.
Barrandon, "The Origin of Cobalt Blue Pigments in French Glass
from the Bronze Age to the Eighteenth Century,"in Archaeometry
94: Procedingsof the 29th InternationalSymposiumon Archaeometry
(Ankara, 1994), pp. 133-40; M. Verita, "AnalyticalInvestigation
of European Enameled Beakers of the 13th and 14th Centuries,"
Journalof GlassStudies37 (1995), pp. 83-98; E. Ciliberto, I. Fagala,
G. Pennisi, and G. Spoto, "Bulk and Surface Characterization of
Early Pigments: Case Study of Renaissance Smalt," Scienceand
TechnologyforCulturalHeritage3 (1994), pp. 163-68.