Medieval glass making-from raw material to finished window, a very brief history

       Those of us living in the 21^st century know more than expected about the making of glass during the Middle Ages thanks to a 12^th century Benedictine monk writing under the pen name of Theophilus Presbyter.  Nine hundred or so years ago he assembled and wrote a three volume book De diversis artibus (On Diverse Arts). (The current standard translation is by Charles R. Dodwell, ed. & tr. Theophilus: 'The Various Arts' De Diversis Artibus,(1986) Oxford: Clarendon Press.)  He covered the art of painting, glassmaking, and metalworking.  From the work, it is deduced that the monk Theophilus was not only an artist but also an accomplished metalworker.  The book apparently circulated widely in Europe as there are English, German, Italian, and French manuscripts.  As a result, almost all of the text and the prologues to the books still exist, though with differences since these were hand-copied manuscripts, and several chapters on the addition of metal oxides as colorants are missing.  Three other later sources for information on glass making include Anthony of Pisa, Cennino Cennini, and Gregorius Agricola. (See Richard Marks, Stained Glass in England during the Middle Ages. (1993) Toronto: University of Toronto Press.  See especially the section on organizing workshops.)

A page from the British Library's manuscript of Theophilus' On Diverse Arts. The I in the left margin begins a chapter in Book I.

British Library Harley 3915 f. 19.  Made between 1150 and 1225 in northwest France or Germany

http://molcat1.bl.uk/IllImages/Kslides/big/K037/K037671.jpg

Materials for making glass

     It is the second volume of Theophilus’ On Divers Art, the book on glassmaking, which is of particular interest.  After the fall of the Roman Empire, glass making divided into a northern sphere and southern sphere over a period of centuries. The glass from north of the Alps differed in chemical composition from glass made south of the Alps.  Glass is melted sand or quartz, SiO₂.  Melting silica or sand takes temperatures in the range of 1623 C or over 3000 F, far above that temperature one can get from a wood fired furnace.  In ancient times, it was discovered that adding an alkaline compound to the sand allowed a much lower melting point.  This was called flux because it allowed glass to flow and thus it could be formed in molds or blown or both.

     Roman glass was made from sand, about 2.5% alumina (or aluminum oxide Al₂0₃), and lime (CaCO₃).  To this was added sodium carbonate (Na₂CO₃) from a natural salt called natron that was found in Egypt.  In Rome and around the Mediterranean, the flux was sodium carbonate (Na₂CO₃).  But north of the Alps, sodium carbonate became unavailable because of disruptions in trade.  So the glassmakers had to use a different material, potash or potassium oxide (K₂O) derived from the burning of wood.  Hence, one name for medieval European glass is forest glass.  The chemical compositions of glass tells us about when and where it was made:

Table of the chemical composition of glass sample by percentage (weight)

	Modern soda-lime-silica glass	Roman glass	Canterbury Cathedral 12th cen.	York Minster 12th cen.	Blunden’s Wood Glasshouse 14th cen.
Silica  SiO2	73.6%	67.0%	54.3%	61%	58.4%
Soda  Na2O	16.0%	18.0%	1.4%	2.5%	2.5%
Lime  CaO	5.2%	8.0%	16.9%	15.8%	13.1%
Potash  K2O	0.6%	1.0%	13.7%	5.3%	11.3%
Magnesia  MgO	3.6%	1.0%	4.2%	7.7%	6.7%
Alumina  Al2O3	1.0%	2.5%	1.1%	2.2%	0.9%
Iron Oxide  Fe2O3	--	0.5%	0.4%	1.3%	0.7%
Manganese Oxide  MnO2	--	--	0.83%	7.7%	1.1%
Phosphorus pentoxide P2O5	--	--	4.8%	4.1%	2.0%

Table adapted from: Corning Museum of Glass, Chemistry of Glass. (http://www.cmog.org/article/chemistry-glass) .  Andrew S. Meek. The chemical and isotopic analysis of English forest glass . Ph.D. dissertation University of Nottingham, 2011(http://eprints.nottingham.ac.uk/11965/1/MEEK_AS_PhD_Thesis_2011.pdf) Ian J. Merchant. English Medieval Glass-Making Technology: Scientic Analysis of the Evidence. Ph.D. dissertation, University of Sheffield. (<http://etheses.whiterose.ac.uk/3464/1/287681.pdf>)

     As can be seen in the table above, analysis of the potassium and calcium content of medieval glass demonstrates a distinctly different content than modern glass and ancient Mediterranean glass.  The potash-lime-silica glass is more easily damaged by prolonged exposure to condensation and rain.  It is the chemical composition that explains in part the problems with corrosion of medieval stained glass.  This is a separate topic and there will be more about this later.

     The ashes from different woods as well as the same wood grown in different soils and harvested at different times of the year all yielded some variety in the chemical content of the glass.  Theophilus wrote that he preferred the wood from beech trees.  In England, oak was more available than beech and so oak wood was also used as the source for the potash used in glass production.  The advantage of beechwood is that is contains trace amounts of manganese in various concentrations that can give glass a yellow to brown or purple color without adding anything further.  Because wood was needed both as a fuel for the glass makers’ furnaces and also as ash from burnt wood that was an ingredient in the glass, access to large quantities of wood meant that glass was made near forests but the source for sand could not be too far away.  Wood was not the only source for ash.  All sorts of other woody materials were burnt for ash-bean stalks, reeds, millet stalks, rushes, brambles, bracken, holly among others. 

Coloring glass

     The various wood ashes and minor contaminants in the sand gave glass some color without even adding further colorants.  Trace amounts of iron gave glass a blue green color.  Theophilus noted that molten glass changed from saffron yellow to a reddish yellow on further heating, probably based on its inherent manganese content.  So some colors were time and temperature dependent.

     Other colors were made with intentional additives.  Multivalent metal oxides were used.  Iron, copper, manganese and cobalt were especially useful.  As noted, manganese, that was a trace element in the wood ash, would give a range of yellow to brown colors.  When manganese was added to glass containing trace amounts of iron, the manganese removed the blue-green color and made the glass clearer.  Manganese added in higher quantities gives the purple colors known for thousands of years.  Since the individual elements had not yet been identified and the chemical composition of mineral ores not known, the chemistry of making colored glass seems to have been by trial and error.

Metal	Color in presence of oxygen (oxidized)	Color when low oxygen (reduced)
Cobalt oxide or cobalt carbonate	Blue	Blue
Cobalt in the presence of Cl- and lower temperature	Pink
Copper precipitated in the range of 10-8m & 10-9m		Red
Copper	Yellow-brown	Blue or green
Iron (II)oxide (FeO)	Blue to blue green
Iron (II, III) oxide, Fe2O3		Yellow to brown
Manganese	Colorless (in added in small amounts)	Weak orange
Manganese	Purple
Lead-tin oxide-silicate	Yellow
Tin oxide (old Byzantine glass) and antimony (old Roman glass)	White and opaque

            Eventually other additives were used to give a wider range of colors:

Additive(s)	Color
Chromium	Dark green to black
Chromium, tin oxide, arsenic	Bright green
Cadmium salts	Yellows
Gold	Ruby red
Nickel salts	Purple
Selenium	Red, also pink to orange
Titanium	Yellow to brown

            Finding a source for the glass colorants was difficult.  Glassmakers discovered that they could use old Roman mosaic glass to color their potash-lime-silica glass.  The blue glass in St. Denis in a suburb of Paris and York Minster in York was made from reused blue Roman mosaic glass.  It was not until the 1200s that cobalt was discovered and mined in Saxony. (Ian Freestone. New light on medieval stained glass by scientific analysis. Glass Circle News.(2014) Issue 134 Vol. 37(2):17-19.) 

     The reuse of waste glass, cullet, and old glass stripped from Roman buildings seems to have been an important source for color and opacifiers in early medieval glass. Antimony found in old Roman glass and tin from old Byzantine glass was being used to make glass in Anglo-Saxon England.( Ian C. Freestone, Michael J. Hughes and Colleen P. Stapleton. Catalogue of Anglo-Saxon glass vessels in the British Museum. London, The British Museum, (2008) 29-46.)

     Red glass was particularly difficult to make since the red produced by the addition of copper as metal filing or scrap copper was often quite opaque.  The additions of small amounts of metallic gold to the There is no evidence that gold was used to make red glass during the period 1100-1400.  Red glass from the period of 1100-1400 seems to have been made by glass flashing.

Furnaces

       Theophilus goes on to describe in Book II how to make a two chambered work furnace with a flat hearth and firebox of stones and clay and lined with clay.  The smaller furnace was for fritting, that is the initial heating of the sand and ash mixture and a larger one for actually melting the mixture and turning it into glass.  Various openings are described so that crucibles can be put in and taken out and gasses of combustion vented.  The furnaces thus described are really semi-permanent structures that  could be moved and rebuilt as needed.  Modern attempts to reproduce the furnace as described by Theophilus have failed.  But a furnace found at Blunden’s Glasshouse, Hambledon, Surrey, England dated to before 1330 revealed a central fire trench with two sides to the furnace.  The walls were built of sandstone.  It is not too dissimilar to the furnace described by Theophilus.  Next Theophilus gave instructions for how to build an annealing oven.  This oven is needed to slowly cool the glass down so it does not break on cooling, making the glass stronger.

Crown and Muff Glass, Flashed Glass

     Once the glass is melted, a glob of the viscous thick liquid is gathered onto a hallow rod and a bubble blown.  The glass is then transferred to another rod called a pontil.  The glass is reheated and spun to make a disc of glass with a thin edge and a bull’s eye in the middle where the pontil was attached.  This method of making window glass was called the “crown” method.  The glass had a shiny surface polish from the fire but the glass cut from the disc would have varying thickness.  The center section of undoubted used again to make more glass, but was not useable to make stained glass windows.  (Waste glass called cullet is used to make new glass.  Cullet has been found at all glassmaking sites excavated.)

     The second method of making glass for stained glass was called muff glass.  A flat piece of window glass is begun by blowing a bubble of glass. The bubble was swung to and fro on the blow‐pipe as it was being blown so that it becomes a long cylindrical bubble.  The ends were cut off the cylinder.  The cylinder was split along the middle and allowed to uncurl on a flat surface in an oven to produce a flat sheet of glass.  Annealing was accomplished by standing the glass on edge in an oven maintained at about 900F.  The advantage of the muff glass was production of a sizeable piece of flat more even thickness glass that had much less waste.  There was more useable glass for the stained glass window maker.  The disadvantage was that the glass lost it shiny surface polish during the manipulation to flatten the glass.

     Both these techniques are given the name of pot-metal glass since the coloring agents were added to the molten glass and the whole thickness of the glass is colored, not painted on later.

     Flashed glass is a technique used when a layer of colored glass is applied to a layer of clear or weakly colored glass, referred to as “white glass”.  The red glass made from copper was especially problematic because it would be very dark, almost black and it was also quite opaque.  To make the glass more translucent and redder, the red glass was applied in a thin layer to clear or weakly colored glass.  Alternatively, the clear or lightly colored glass could be dipped into the red glass with some partial mixing of the glass layers.  Flashed glass appears in stained glass window from the 12^th century at Canterbury Cathedral and 13^th century from the Cathedral of Saint-Gatien at Tours and the prophet windows at the Victoria and Albert Museum in London that probably came from the Cathedral of Saint Peter and Saint Paul at Troyes.  So flashed red glass made from copper was usual way of making red glass during most of the Middle Ages.

     In one study of medieval red glass that included glass from York Minster, red glass appeared as multiple layers of copper rich red glass over a thick layer of white glass.  This pattern differed from two others.  One showed a sandwich of red glass between a thin and thick layer of white glass.  The third showed one layer of red glass over white glass.  The multiple layers were not due to folding of the glass but rather the migration of copper ions.  Copper ions appear to move within the glass from an oxidized copper rich layer to the reduced copper poor layer.  The clear glass is gathered and then dipped into a copper rich glass before it is blown.  On reheating the red color in the glass appears.

The development of…red [was] critically dependent on the juxtaposition of an oxidised high-Cu glass, which is likely to have been pale blue or green when cool and a colourless reduced low-Cu glass.  The knowledge needed to produce the red glass is likely to have been restricted to a relatively limited group of craftsmen, and the process of transformation of the two weakly coloured glasses into red must have appeared a marvel to the medieval artisans. The symbolic signicance attached to the colour may have added to the impact of the transformation:- according to the future Pope Innocent III (1198-1216) red was the colour of blood, martyrdom and Christ's Passion (JJ Kunicki-Goldfinger, IC Freestone, I McDonald, JA Hobot, H Gilderdale-Scott, T Ayers. Technology, production and chronology of red window glass in the medieval period - rediscovery of a lost technology. Journal of Archaeological Science 41(1) 2014 89-105)

Three different types of flashed red or copper glass.  The top is the glass from York Minster,  The middle piece is the sandwich copper flashed glass and the bottom is modern flashed glass with one layer over a white base.  From the work of Ian Freestone.

     Given the amount of labor needed to get the raw materials and prepare everything for glass making, the profession required a large number of unskilled and skilled workers who could work together in a coordinated and synchronized fashion to produce the glass that the glass painter and window maker needed.

The Glazier: Glass Painting and Glass Staining, Cutting, Fitting and Assembling

     It seems as though almost all the colored glass was made on the mainland of Europe especially the Rhineland, Burgundy, Normandy, Flanders.  White (that is clear or weakly colored glass) and inherently colored glass was apparently made in England.  Once the glass was produced whether on the continent or in England, it went to the glazier, the window maker.  It appears as though the men making the windows were a distinctly different group from the glass producers, especially when large productions were involved.  This was similar to the methods of the Roman Empire where the production of raw glass was near to the sources for sand and the salt natron.  This was then sold to those who made the finished product of glass vessels or window glass. Theophilus wrote as though the glassmaker could also be the glazier.

     Designs for a window were drawn directly on to white washed tables.  Paper was not widely available, and it was expensive.  Tables were ready made and white wash was inexpensive. Pieces of colored glass were cut into rough shape with a dividing iron and then smoothed and fitted into the design using a grozing iron.  Then came the glass painting that could involve a number of steps so that facial expressions, folding of cloth, and other details could be applied to the glass.  The paint was a mixture of iron oxide, finely ground copper and powdered glass mixed with fixatives such as wine, urine and vinegar.  Subsequently the pieces of glass were fired one or more times to permanently fuse the paint to the glass.  The colors obtained ranged from brown to black.  Brushes for the paint were made from various animal hairs and could be exceedingly small for fine detail.  Once the glass was fired, the painted pieces were fit together with lead came, an H-shaped lead edging, and soldered to make up a panel for glazing the window.  After the panel was completed, it was waterproofed (cemented) with a mixture of chalk and linseed oil.  After chalk mixture had dried, the powdery residue was rubbed off.  The windows were usually affixed to the windows with iron rods to add strength.

The stain in stained glass: painting with silver salts

     The new advance in glass painting came about 1300 when it was discovered that silver salt solutions mixed with fixatives and clay could color the glass colors from bright yellow to intense yellow-orange after being fired in a kiln.  The most common compound was silver nitrate but silver chloride, silver sulfate, and silver oxide worked well when mixed with turpentine plus wine and urine as binding agents and covered with clay.  The glass was painted and kiln fired.  Once the clay is removed, the yellow appeared. The yellow actually migrated into the top layer of glass. The glaziers in the York area of England became masters of this glass staining technique and many examples survive.  Persons not only sported bright yellow hair but angels had golden wings and golden stars on gowns.  Bright yellow appear on armor and spears.  Skies had golden stars and halos became bright gold.

Nativity scene from Church of St. Peter and St. Paul, East Harling, Norfolk. 15th century. The silver stain was used to make Mary's hair blond, the straw gold. Joseph's staff yellow (Joseph has lost his head.), and the start of Bethlehem gold with golden rays.  Note the yellow in the canopy and edging.

http://www.norfolkstainedglass.co.uk/East_Harling/images/east_2c_big.jpg

Angel playing a bagpipe from west window, Coronation of the Virgin, St. Peter Church, Hungate. 15th century.  There is generous use of the yellow silver stain on feathers, clothing, hair and even bagpipe.

http://www.norfolkstainedglass.co.uk/angels/images/St_Peter_Hungate_wind.jpg

References (See also ones included in text)

John Blair and Nigel Ramsey, ed. English Medieval Industries: Craftsmen, Techniques, Products (1991) Cambridge: Cambridge University Press. J.R Charleston, “Vessel Glass” 237-264, Richard J. Marks, “Window Glass” 265-294

Sandra Davison. Conservation and Restoration of Glass, 2^nd Ed.(2003) Oxford: Elsevier Science

David Dungworth and Sarah Paynter.  Blunden’s Wood Glasshouse, Hambledon, Surrey: Scientific Examination of Glassworking Materials.  English Heritage Research Department Report Series 38-2010

Freestone I, Kunicki-Goldfinger J, Gilderdale-Scott H, Ayers T (2010) Multi-disciplinary Investigation of the Windows of John Thornton. In Shepherd M B, Pilosi L and Strobl S (eds) The Art of Collaboration. New York: Harvey Miller.

Kunicki-Goldfinger, J. J., Freestone, I. C., McDonald, I., Hobot, J. A., Gilderdale-Scott, H., & Ayers, T. (2014). Technology, Production and Chronology of Red Window Glass in the Medieval Period–Rediscovery of a Lost Technology. Journal of Archaeological Science 41, 89-105.