Manganese crystalline glazes (high alkali, silica, and alumina) are usually created by saturating a feldspathic glaze with between 15-60% manganese dioxide. During the cooling cycle, manganese precipitates out of the molten glaze and crystallizes on the surface, producing lustrous, satiny surfaces.
UNDERSTATEMENT: Manganese Dioxide is extremely hazardous to your health!!!
Breathing in Manganese dust when mixing these glazes or breathing the off-gassing vapor when firing WILL GIVE YOU PARKINSONS-LIKE SYMPTOMS BEFORE ULTIMATELY KILLING YOU, PAINFULLY. HEAVY GLOVES, DUST MASKS, AND VENTILLATION ARE CRITICAL.
There is a long history of lustrous, metallic glazes. The first examples are thought to be from the early ninth century in an around what is modern day Iraq. Archeological evidence suggests that early examples originated from Mesopotamia in Fustat, which was then the capitol or Egypt. The oldest surviving examples were often multi-colored stains and iridescent sheens derived from copper and silver compounds. These compounds were usually manufactured by dissolving coins into acids and then mixing the resulting solution with earthenware clay. This mixture was then calcined and then finely ground. The resulting pigment was then mixed with a carrier (usually lavender oil) and applied to lead or tin glazed pots and re-fired to dull red heat. The pots were then held in an extremely smoky reduction environment at various temperatures and lengths of time, which resulted in surfaces ranging from olive-green, brown, amber, orange, yellow, crimson, and a very dark red which was sometimes so dark as to look almost black.
Although the history and development of reduced-pigment lusters is long and storied, it was a more or less consistent sequence. It isn’t until the 19th century that one starts to find examples of resinate lusters. This resulted in the development of materials almost identical to modern ‘liquid gold’ and ‘platinum’ lusters. In Europe in the 1870s a revival in the technology and development of luster glazes saw a further refinement of reduced glaze lusters, most notably in the studios of William De Morgan, Massier, Kähler, and Zsolnay. This notable shift was the result of the use of higher firing clays, which French ceramicist Louis Franchet believed could offer the complete range of earlier pigment-lusters, but without a lot of the trouble. Aside from the obvious temperature differences, the main difference between pigment and reduced glazes is that glaze lusters are generally less subtle, less mellow, and offers a wider, more brilliant range of color.
I began research on this project in an attempt to find a brilliant, iridescent glaze similar to Zsolnay’s famous Eosin glaze, which has a very obvious bright reflective rainbow iridescent quality. Initial research suggested that Zsolnay’s effects were the result of the thin application of copper, silver or bismuth to a pre-fired glaze – firing to fusion point, and then reducing the kiln atmosphere during the cooling cycle. This method is documented extensively in Greg Daly’s book Lustre. Having had some glimmers of success with iron saturate glazes in reduction cooling environments, I proposed a solution that did not; 1.) involve expensive silver or bismuth oxides, or caustic salts such as stannous chloride or copper sulfate, and 2.) involve a postfiring or overly exotic and difficult to repeat firing schedule. In the end, a satisfactory solution was some combination of feldspathic glazes with 30-60% Manganese Dioxide, following closely in the steps of David Shaner, Lucie Rie, Hans Coper, John Tilton, and historical Rockingham ware.
Reduced-pigment luster. Nearly all historical luster made before 1800 fits in this category. The result of calcining copper, silver, and bismuth oxides with earthenware or laterite clays, and applying the resulting mixture to a maturely fired lead or tin glaze surface. The piece is then refired and held in heavy reduction at dull red heat allowing for a thin layer or metallic oxide to fuse with the surface of the glaze. After the firing, the earthenware is wiped away, revealing a nano-thick layer of iridescent metal.
Resinate luster. Usually made with dissolved gold, platinum, or other noble metals and suspended in an organic binder. Generally fired to a low temperature, with the organic compounds burning out and fluxing a thin, even layer of metallic oxides with the surface of the work. Developed around 1800, very common in industry, very toxic.
Reduced Glaze Luster. Generally higher porcelain and stoneware temperature. Usually cover the entire surface of a form. Relies on metallic saturated glazes precipitating out thin layers of reduced metallic oxides which deposit in a thin layer on the top of the glaze. Generally more brilliant and operate across a wider spectrum of interrupted light.
Materials: I found that nearly all of my iridescent surfaces contained some percentage of manganese. The exception is a traditional Tenmoku glaze fired in standard reduction, and then ‘struck’ at 1840F for 1:20-2:00 hours. Strike firing, or striking the kiln is a glass term which refers to increasing the fuel supply and thus creating a reducing atmosphere around 1800F. Initial tests suggested that manganese saturated glazes promoted richer iridescent surfaces regardless of a strike firing. Additions of other oxides were often counterproductive to glossy surfaces and generally resulted in unpleasant black, rough surfaces. Copper, Iron, Chrome, Nickle, and Cobalt were all tested alone and in conjunction from .1 -> 20%. The character of the underlying glass matrix of was usually beer bottle brown, so I tested extensively to change the color of the glass without effecting the iridescent surface – to date I still don’t have a simple solution to this problem. Granular Manganese seemed to produce brighter colors as well as promoting streaking ‘hares-fur’ effects in faster cooling, and acting as ‘seeds’ to crystal formation on slower cooling cycles. My ideal concentration of granular manganese was 2% and fine manganese dioxide at about 27%.
Most recipes called for 50-70% feldspar, and after testing all of the available feldspars, I found that Nepheline syenite promoted a much smoother, regular iridescence. Custer feldspar promoted iridescence across a wider spectrum, but promoted intense crystallization as to appear almost pixellated. Kona f4 promoted a more matte, golden green/purple sheen. Other feldspars promoted a lustrous brown glass with varying degrees of light to moderate iridescence.
The addition of silica promoted a lightening of the glass matrix, as well as a sugary, semi- shiny sparkling satin luster. Silica beyond 15% eliminated iridescence. Alumina additions to the glaze produced a semi-matt honey colored glaze.
I found that the clay body had a huge impact on the color and quality of the iridescence. The most successful clay bodies were grolleg based porcelains, with only the highest percentages of manganese based glaze recipes showing even the slightest luster on stoneware recipes.
Finally, glaze thickness was perhaps the most critical aspect of obtaining iridescence at high temperature. This was complicated as these glazes are extremely runny. Even slight overfiring resulted in glazes running off the pot. There was a need to find a balance between adding clay and silica to the feldspar and manganese without diluting the concentration of available metal oxides and feldspar. It was also extremely difficult to apply these glazes consistently, and fire them in such a way as to reach maturity without overfiring.
Firing: All tests were fired in high temperature gas kilns. I usually fired to 1260C, or Orton cone 10. A majority of my testing was in standard cone 10 reduction firing, with a 1 hour body reduction at cone 012-> cone 08, and a 6-10 hour firing from cone 08-> cone 10. Recipes with 15% copper produced a striking gold color in oxidation environments, and glazes in oxidation firings bubbled and boiled up between cone 7-9, which suggests a similar thermal reduction similar to oil spot glazes.
Cooling: Most of my firings were in small soft brick or fiber kilns, so the possibility of extended cooling cycles was limited. I found that crash cooling seemed to promote smoother, less brilliant surfaces, and a moderately fast cool was ideal in creating a balance between bright color and reasonably smooth surface. Longer cooling promoted larger crystals to a point, and excessively long cooling cycles promoted a matte surface. Reduction cooling remains an exciting possibility which mostly extended beyond the scope of my research. A very interested mottled crystal growth was observed on bottle forms cooled with a 3 hour reduction hold at 1840F.
 Caiger-Smith, Alan. Lustre Pottery: Technique, Tradition, and Innovation in Islam and the Western World. London: Faber and Faber, 1985. Print. Pg. 21
 Caiger-Smith, 1985, Pg. 177
 “Iridescence in Lepidoptera”. Photonics in Nature (originally in Physics Review). University of Exeter. September 1998. Retrieved April 27, 2012.
Britt, John. The Complete Guide to High-fire Glazes: Glazing & Firing at Cone 10. New York: Lark, 2004. Print.
Caiger-Smith, Alan. Lustre Pottery: Technique, Tradition, and Innovation in Islam and the Western World. London: Faber and Faber, 1985. Print.Pg 149
Conrad, John W. Black Pearl and Other Saturated Metallic Glazes. Santa Ana, CA: Falcon Division of Aardvark Clay, 2010. Print.
Currie, Ian. Revealing Glazes Using the Grid Method. Australia: Bootstrap, 2000. Print.
Daly, Greg. Lustre. London: A. & C. Black, 2012. Print.pg. 131
Hamer, Frank, and Janet Hamer. The Potter’s Dictionary of Materials and Techniques. London: & C Black, 1991. Print.
My name is Matt and I'm a potter living in Southeast Alaska. I've been an artist/teacher/potter for the past decade, and I got my start in ceramics in high school some 18 years ago. These days I make my living selling wheel thrown pottery that sits at the intersection of ceramics/science/mineralogy/and geology.
View all posts by mattfiske