Snowflake glass is a unique type of 19th-century Chinese glass that, as the name implies, invokes images of snow. The effect is created with varying amounts of white inclusions (small particles inside the glass) and small air bubbles in a colorless or translucent white base glass. In Chinese, this type of glass goes by multiple names which can be translated as follows: lotus root powder ground (藕粉地), snowstorm (霏雪地), and saliva ground (唾沫地). Some of the names refer to its common use as a background glass with a contrasting overlay color, often red, in cameo-carved objects.
The best-known example of this type of glass in our collection is the Snowflake Warrior Vase, a large cased and cameo-carved vessel that uses snowflake glass as the background layer with a thick red overlay (read more about the story told through its intricate carvings here). While the Warrior Vase is the largest and most impressive object made with snowflake glass in our collection, there are many smaller objects, such as snuff bottles, rings, bracelets, and buckles, which also use snowflake glass.
Interest in the Warrior Vase from our visiting Carpenter Foundation Fellow, Dr. Shelly Xue, prompted a joint research project with curators, conservators, scientists, and glassblowers into snowflake glass which until now, had not been extensively studied.
The team chose 42 snowflake glass objects, including the Snowflake Warrior Vase, to explore the variety of snowflake glass and what the observed differences might mean about where, when, and how the objects were produced.
Research projects like this typically start with simpler research techniques, usually visual examination, and progress to more complicated and time-consuming ones like chemical analysis. This process helps us group objects that are similar so that we can limit the number of objects we need to do chemical analysis on. It also allows us to refine the questions we hope to answer with the analysis (usually the question is: why do these groups look different?).
All of the objects were examined visually and microscopically where we primarily looked at the size, shape, and density of the bubbles and inclusions as well as the opacity of the base glass and sorted the objects into rough groups based on the differences in these characteristics.
We also looked at all of the objects under ultraviolet irradiation (UV). Under UV, certain materials fluoresce, or glow in distinctive ways. Glass itself does not always fluoresce under UV, but if it does it may tell us something about its composition. An icy blue fluorescence, for example, suggests a high lead content. While UV cannot reliably identify glass compositions, it can be useful to compare glasses and establish similarities. In this case, UV examination helped to further sort the objects into three primary groups (based on the fluorescence of the snowflake and colored overlay glasses).
We selected twenty of these objects for compositional analysis with X-ray fluorescence (XRF) based in part on these groupings. XRF is a non-destructive analytical technique used to determine which elements are present in a material. For glass, XRF analysis can only get semi-quantitative results, because of variations in how well the instrument can detect different elements. In fact, lighter elements such as sodium, which are often major components of glass, cannot be picked up at all.
The XRF analysis confirmed three distinct glass formulations, sodium-borosilicate glasses (based on the absence of other elements), lead-free potassium-lime glasses, and high-lead potassium-lime glasses. These groups mostly correlated with the three groups created based on the visual, microscopic, and UV examinations. The significant differences in the composition of these three groups suggest that they were made from different recipes, likely at separate production sites and/or time periods.
27 objects (26 with red overlay, including the Snowflake Warrior Vase; 1 with transparent blue overlay):
- Lots of variation in visual characteristics
- Opacity of the base glass ranges from clear to cloudy, skewed more towards the clear side.
- Inclusions range from few to many but are generally very small to medium in size and often form small clusters.
- Number of bubbles also varies but is generally in the intermediate to high range.
- UV examination:
- Snowflake glass: variety of orange-pink fluorescence under longwave UV.
- Red glass overlays: yellow-orange fluorescence under shortwave UV.
- XRF (14 objects):
- No detectable fluxing and stabilizing components which indicate a sodium-borosilicate composition by default (sodium and boron are not detectable by the XRF instrument).
- Red glass colored with copper
10 objects (8 with red overlay; 1 with green overlay; 1 with red base and snowflake overlay):
- Visually similar to the sodium-borosilicate glasses, but different under UV.
- UV examination:
- Snowflake glass: yellow fluorescence under longwave UV.
- Red glass overlays: no fluorescence
- XRF (4 objects):
- High potassium and calcium; no lead.
- Red glass colored with gold
5 objects (all with opaque blue overlay):
- Visual examination:
- Base glass is much less translucent
- Large variety in the size and shape of the inclusions creates a mottled effect very different from the other glasses in which the inclusions are more distinct.
- Opaque bright blue overlay glass found on all five objects points to a more contemporary, early 20th-century date.
- UV examination:
- Snowflake glass: icy blue fluorescence (typical of lead glass) under shortwave UV.
- Blue glass overlays: no fluorescence
- XRF (2 objects):
- High potassium, calcium, and lead.
- Blue glass colored with cobalt
In many ways, this research generated more questions than it provided answers, but it has greatly increased what is known about this unique and fascinating glass that is representative of 19th-century Chinese glass production. One of the things we still don’t know is how the snowflake effect was produced, so look for an upcoming blog on how we tried to figure that out through experimental recreations at the furnace.
I would like to thank everyone involved with this research project, in particular Dr. Shelly Xue, Dr. Gregory Merkel, Harry Seaman, Lianne Uesato, Dr. Jane Cook, and the CMoG collections staff.