Due to the harmfull effects of lead and environment it is now tried to avoid the use of lead as much as possible. This is also the case in the glass industry and people are trying to find substitutes for lead-containingglasses. However, it is not only the base glass one has to consider: the metal oxides used for glass colouring usually give totally different colour tones if lead is not present in the base glass. 

Also some colouring materials are very harmful if one considers the occupational health and safety of the worker. Cadmium, selenium and nickel are all potential carcinogens and thus their use for colouring should be avoided, if possible. Materials which can be used for colouring and which have no severe health effects are iron(which oxides in glass to FeO and Fe203), cobalt (CoO), chrome oxide (supplied as Cr203 or another Cr3+ based compound, the Cr6+ compounds are considered to be carcinogens and thus they should not be used), copper (Cu20, CuO, Cu), manganese (MnO, Mn203), vanadium (V20s, V203, VO), titanium (TiO2) and lanthana (mainly cerium, neodymium and praseodymium). As seen, these elements can exist in glass in various oxidation states. The colour of the glass is strongly dependent on the oxidation state of the colouring oxide and thus it can be affected by e.g. other ions in the glass and the melting history of the glass (oxidation or reduction heating). 

Using iron compounds is the most inexpensive way of colouring glass. Iron colours the glass green; depending on the base glass and on melting the green colour may have blue or yellow tone. The amount of iron oxide used is usually between 0,5 - 5%. 

Cobalt colours glass blue. The colouring effect is very strong, even 0,02% of cobalt oxide in glass gives a clear blue colour. As cobalt can have only one oxidation state in glass (CoO), the colour is not sensitive to changes in the glass composition or melting process. 

Chrome oxide, Cr203 colours glass green. In the glass melt there is always also the other oxidation state, CrO3, present. This gives a yellow tone to green. Usually amounts of 0,1% - 0,5% of CrlO3 are used. 

Copper gives glass a blue colour, but when compared to cobalt the colour is more turquoise. Usually Cu2+ compounds are used for colouring. Typical amounts are 1-3 % of CuO. If copper oxide is reduced with a suitable heat treatment, copper-ruby glasses are formed. Manganese colours glass from red-brown to violet depending on the ratio of the two colouring forms. 0,1 % of MnO produces a light colour when 5% gives a dark one. 

Vanadium exists in glass in four different oxidation states and thus the colour it gives may change from green-brown to bluegrey. The usual colour is, however, green.
Comparing chrome and vanadium one can find out that chrome gives about five times more intense colour than vanadium with the same amount of colouring oxide.
Titanium itself does not have a colouring effect. It is used together with some other colouring agents to intensify or change the colour. Titanium changes the colour given by Mn2+ ions from colourless to yellow and the colour given by copper ions from blue to green. 

The light transmission (orabsorption) of glass is usually measured by using a spectrophotometer. In this device the spectral transmission i.e. the amount of light passing through the glass in different wavelengths is measured. Typically the measurement is done at 5 nm intervals. In addition to visible light (about 400 - 700 nm) the measurement range usually covers some uv-area (300 400 nm) and IR-area (700-1 100 nm). However, if thinking of the visual appearance, only the range of the visible light is important. The measurement result is shown as a spectral transmission curve, Fig. 1.

It is usually diffficult to see from a transmission curve what is the colour of the glass i.e. is it green, red, blue or some othet colour. Thus procedures have been developed to calculate from the spectral transmission curves the tristimulus values (X,Y,Z) which define the psychophysical colour for the specified observer, illuminant and geometry of illumination and view. The problem with the X,Y,Z values is that the colour space is not uniform. If a colour space more nearly uniform is wanted the CIELAB colour system is used. In this system values L*, a* and b* are calculated from X,Y,Z values. ClELAB system correlates closely with colour change as perceived by the human eye. The L* co-ordinate represents the grey scale so that L*= 100 represents pure white and L*= O represents black. The a* coordinate represents red (positive values) - green (negative values) axis of the colour space and b* represents the yellow (positive) - blue (negative) axis. Thus e.g. the colour L*= 50, a* = 30, b* = -1 5 would be given to mediumgrey, red-blue colour. 

Dhe aim of this study is to develop coloured glasses for lead-free base glass. Two different base glasses were chosen. The composition of these glasses is shown in Table 1. 

The colouring agents were supplied as oxides and the following oxides were used: Fe203, Cr203, Cu20, CuO, MnO, Mn203, V205, CeO2, Sb203, TiO2, SnO2, Nd203 and Co304.
The laboratory experiments were done in the Åbo Akademi University at the Department of Inorganic Chemistry where the suitable equipment for laboratory scale glass melting and spectraltransmissionanalysiswasavailable.Themeltingfurnace was an electrically heated pot furnace with SiC elements. The melting atmosphere was normal room air. 

Hi-alumina crucibles were manufactured by slip casting and fired to 1300šC. 300 g of the base glass was melted in the crucible at 1450šC and the glass was then poured onto a steel plate and the formed sample annealed. The sample was then analysed by spectrophotometer to ensure that no impurities (especially no iron) was dissolved into the glass from the crucible. 

For manufacturing the colour samples the Glass I glass, supplied as pellets, was crushed and the colouring metal oxides were mixed into the crush. The Glass 2 glass samples were manufactured using 2/3 of cullet and 1/3 of crystalline raw materials. The colouring oxides were mixed with the crystalline materials. The crucibles were first heated up to 900šC and then moved into the melting furnace ( 1450šC). After the crucible had reached this temperature. 150 g of glass raw materials was put into it and after it had melted another 150 g. Then the glass was left to homogenize for about 4 hours after which it was poured onto a steel plate and annealed. 

Dhe annealed samples were ground and polished so that the measurement thickness of the sample was about 5 mm. Then the spectral transmission was measured using the PerkinElmer spectrophotometer and the CIELAB values calculated from the measurement curves. The values given in this report are calculated for D65 standard illuminant and observer with 5 nm intervals over the wavelength range 380 to 780 nm. 

Figure 2 shows the spectral absorption curves for the two base glasses when I % of manganese oxide was used for colouring. If two or more metal oxides are supplied into the glass melt, one has to take in to account their relative oxidation/reduction reactions. 

If the metal oxides occur in glass only in one valence state, their combined effect can be calculated by merely adding the absorption curves. However, the metal oxides usually occur in several oxidation states and the resulting colour cannot be predicted by simple calculation. 

This can be seen e.g. in the combination of Ti and Ce ions, they both give very low a* and b* values, and by simple calculation one would accept the combination to be almost non colouring. However, in reality the combination has a fine yellow hue. It is also interesting to note from table 2, which shows the CIELAB values for different combinations of colouring oxides, how strong effect the different amounts of oxides of the system MnO-Cr203 has on the resulting colour. 

CONCLUSIONS AND FUTURE CHALLENGES

The laboratory scale experiments give the basic absorption curves for different colouring agents in case of electric melting. In production scale the furnace and the melting pot may significantly affect the colour.Additional experimentswill have to be made to study this. 

Compared to the absorption or transmission curves, the calculated CIELAB values (when compared to colour standard plates) give a visual observation of the colour. These values can be transferred into computers, such as into CAD systems. By this way a designer would always have the knowledge of the colours possible in the production. 

The measured values show that changing the base glass also greatly affects the obtained colour. In the case of combining several colouring oxides, the resulting colour is hard to predict. Small changes in the amount of colouring oxides may have a dramatic effect on the colour. 

ASTM Standards on Color and Appearance Measureme Philadelphia, 1987 

The authors would like to thank PhD Leena Hupa and Professor Kaj Karlsson from the Åbo Akademi University for the irextensive work and advice in the project. 

Info:

Raija Siikamäki,@
Research Scientist

University of Art and Design Helsinki UIAH
Department of Ceramics and Glass
Hämeentie 135 C 
FIN-00560 Helsinki, Finland 
phone: +358 9 75630395, fax: +358 9 75630275