Last Updated: Oct 20, 2016     Views: 467

I am not a glass scientist so I will defer to the glass science books to answer your question! :)  I will say that a scientist once explained color to me very simply as a chemical change that happens at the molecular level. Thus, once the glass is solidified and not molten, the glass is what it is! However, there are, of course, caveats to that...aren't there always.  For example some older glass formulas used manganese as a coloring agent.  Manganese reacts to UV rays and purples with exposure.  For that reason, manganese fell out of favor as an additive.   

I like this description of the manganese/solarization effect from's all about glass:

Protection Against UV Radiation

In certain situations, solar radiation can damage the color of objects exposed to it. This change in color is due to the gradual degradation of molecular links caused by high-energy photons. Such damage is caused by ultraviolet radiation and, to a lesser extent, shortwave visible light (in the violet and blue range). Solar radiation also causes the temperature to increase, thus accelerating this process.

Some glass products can combat this discoloration:

>Laminated glass with PVB interlayers absorbs more than 99% of UV radiation up to 380 nm

>Colored glass with a predominantly yellow-orange tint partially absorbs violet and blue light

>Glass with a low solar factor limits temperature increases

That said, no glass product can completely eliminate discoloration. In fact, in some cases, interior artificial lighting can also cause discoloration.


There is also a highly readable description of the underlying science behind coloring glass on the ceramicsweb page:

"The Nature of Glass         

- by Karl Platt

A few issues back we started a discussion on the elemental nature   of glass. Our interest then was to define how glass comes about   and to look at those substances which readily lend to it's existence.   Here we want to review what we've covered before, and take-up   how these basic glasses are manipulated to conform to studio   needs.

  Silica (SiO2) and Boric oxide (B2O3) form glasses readily all by   themselves and are the predominant glassformers used in Studio   glasses and glazes. SiO2 by itself will form an extremely viscous   melt, and then only at very high temperatures (+3000F). B2O3   melts at much lower temperatures, has higher fluidity, but as a   glass it reacts with atmospheric water to deteriorate.   The properties of these glasses is strongly tied strength   of the metal-oxygen bonds (Si-O or B-O) making them up.

  Pure SiO2 glass is relatively hard and has an exceedingly low thermal   expansion (5.5x10-7). Although technically useful, SiO2 glass is   difficult to make and hardly practical in a day to day sense. B2O3   glass, by contrast, melts low, is fluid and makes a highly   expansive, easily corroded glass. In general, this indicates on the   atomic level that Si-O bonds are notably stronger than B-O bonds.

  There are also structural differences between SiO2 and B2O3   glasses, but we'll take that up further later.

  SiO2 and B2O3 glasses have extreme properties. If more broadly   useful glasses are to be had, pure oxide glasses need to be modified   by adding other elements, called, naturally, Network Modifiers.

  The strong Si-O bonds, which account for the exceptional   properties of SiO2 glass, are perturbed by Network Modifiers.   Generally speaking, Network Modifiers affect the properties of a   pure SiO2 glass as follows:

  A. Thermal Expansion Increases
  B. Hardness Lowers
  C. Chemical Durability Lowers
  D. Density Increases
  E. Surface Tension
  F. Refractive Index Increases

  It is exceptional to use B2O3 as a lone glassformer. Usually, B2O3   and SiO2 are combined to form what is known as Borosilicate   glasses. Borosilicates are a diverse and widely used class of   glasses. We will also take them up further along. So keep them in   mind. For now though we will focus on the effects network   modifiers have on SiO2 based glasses.

  Since ancient times just about every element on the planet, and   some which weren't, have been added to silica host glasses. In   modern Studio setting, the range of compositions used has practical   limits. Of the entire palette presented by the Periodic Table of the   Elements, we use about 10% often and another 30% to a more   limited degree-- either as colorants or for special effects. The   remainder are either rare, dangerous or both.

  Table 1 below lists the elements other than oxygen and fluorine   most frequently used in most of the glasses you'll meet. This list   is arranged according to the field strength each atom holds as an   ion. Field Strength? Ions? Glad you asked.

  An ion is an atom which has either a net positive or net negative   charge. As such, an ion's interaction chemically is described taking   the point of view that each ion behaves like a little magnetic   sphere. A negatively charged ion, known as an Anion, such as   oxygen will draw to it positively charged ions, known as cations.   Conversely, either anions or cations alone will repel each other.

  Again, Cations (+) and Anions (-) have an affinity for each other   (+)i(-), but will repel themselves (+) < > (+).

  Ions become ions when electrons in the outermost region of an   atom are gained or lost to a neighboring atom. Electrons have a net   (-) charge and orbit about an atom's nucleus, which is composed   of protons (+) and neutrons (no charge).

  The amount of charge borne by an electron or proton is the same, thus for each proton there is an electron to balance it in orbit about   the nucleus. Well, most of the time sort of.   Anyway, electrons orbit about the nucleus

  Electrons take up orbital positions in a very orderly way. We'll   leave the gory details to chemistry books, but those farthest from   the nucleus are held less firmly than those closer in. Outer   electrons, as a result, may be shed to, traded with or shared   with any neighboring atom. That is so long as the other atom   is not selfish with its outer electrons. What blend of these   circumstances occurs when two atoms meet depending on the elements   and their environment. Which atom sheds or yanks how many electrons   to or from whom and how is the substance of chemistry.

  The outer electrons are known as Valence electrons. When one is missing, the   ion will be said to have a Valence of +1. When there's an extra available,   the ion is said to have a Valence of -1. Oxygen always has a Valence of   -2 and this dictates many of the reactions that make up the ceramic world;   like we'll see in an example with iron (Fe) below.

  The process of absolutely losing an electron is known out in the world   as oxidation. When iron (Fe) in a glass or glaze is subjected to an   oxidizing fire electrons are stripped from the Fe, which   gives it a net (+) charge. As we noted before   any oxygen (-2) out loose would rather not be and will take advantage of   the electron deficiency to combine with the Fe in the ratio of 2:3;   which we know as Fe2O3. The Fe when stripped of an electron takes on a net   positive charge and assumes extra oxygen, which are negatively charged to   so that the system comes into balance create a balance.

 Reduction is the reverse of this process. In the case of Fe, when recuced to   FeO it behaves as a network modifier and shares some rather low melting   compositions with SiO2. This is why reduced Fe glazes of tend to be   runny.

  Only a few elements prefer to be anions (-) and chief among them,   for our purposes anyway, is Oxygen. Oxygen happily reacts with   anything presenting enough positive charge. Oxygen is Big. With   an ionic radius of 1.4 it is larger than any of the cations in   Table 1. Accordingly, our glasses, on the basis of volume, are an   oxygen bulk knitted together by cationic metals. Cationic, not   catatonic; wake up, this is important.

  Alkaline metal oxides (alkalies) are foremost in altering the   properties of pure silica glass. In order of increasing atomic   weight, the alkaline metals Lithium (Li), Sodium (Na) and   Potassium (K) reside in the leftmost column of the Periodic Table.

  The alkalies have a valence of 1 and combine with oxygen in the   ratio 2:1 as Li2O,Na2O and K2O respectively. In an SiO2 melt   the alkalies perturb two Si-O bonds for each Li,Na,K-O bond   added to the melt. Each of the alkalies has a distinct impact on the   properties of the glass according to it's nature as an ion in the   glass. We should look at this more closely.

  Lithium is a very small, with an ionic radius of 0.78è, and is   highly charged for its size. As such, it really perturbs Si-O bonds,   and being so small it can freely diffuse through the network of Si-   O. Li2O glasses have broad light absorption properties and thus   colors formed within them often seems lifeless.

  Colors you see are the product of having the coloring ion (usually   transition elements) "squished" between the other components in the   glass. Li doesn't squish especially hard and the colors are subsequently   weakly defined.

  Sodium is the most widely used alkali. Abundant in nature, though   not always in a directly useful forms, sodium has played a huge   role in glasses since antiquity. Of the alkalies, it is of middling size   and has a moderate charge. It's ability to squishthe transition earths is   is middling and so are the colors produced.

  Potassium is larger and more weakly charged than either Sodium   or Lithium. As such, the extent to which it lowers viscosity is   lower. The potassium ion, fattish and weakly charged, has   interesting effects on colorants, by "squishing" hard which   lends to sharper spectral absorption and purer percieved colors.

  Almost immediately, alkali additions make an SiO2 melt more   fluid. By the time 10 mol% is added, viscosity will reach a   minimum for a given temperature. Figure 3 shows the effect the   alkalies have on viscosity. Note the steep slope of the curve   between 0 and 10 mol% and the relative position of Li to Na to K.

  The effect of the alkalies is to interpose themselves within the   glassy Si-O network where they represent a weak spot through the   imposition of weakly held oxygens. Subsequently, all Alkali-Silica   glasses will react readily with atmospheric water, which happily   assumes a stronger affinity for the weakly held oxygen. The extent   to which the alkali-silica glass is soluble increases with the amount   of alkali.

  Allied to low chemical durability and viscosity reductions is high   thermal expansion.

  Sodium-Silica glasses (sodium-silicate) is widely exploited for its   solubility. When dissolved in water the Na+1 and Si+4 operate   to suspend dirt which leads to their wide use in laundry detergents.   For the same effect, sodium-silicates are used in ceramic slips as   a suspending agent. Li and K silicate glasses produce similar   effects, but neither as well or as economically as sodium silicates.

  Solubility and excessive thermal expansions are, of course, highly   undesirable in studio applications. Fortunately there are many   substances which act to occupy the weakened oxygens present in   alkali-silica glasses.

  These are known as the alkaline earths, which are characterized by   a valence of +2. They combine with oxygen in the ratio of 1:1   and in alkali-silica glass they operate to tie up weakly held oxygen   internally, so that atmospheric water is thwarted from latching on   to them.

  The behavior of the alkaline earths is a bit more complicated than   with the alkalies. Each alkaline earth has its own sort of   personality. Magnesium (Mg) and Barium (Pb) represent the   extremes in terms of size. Lead is of course the heaviest.
  We will address the alkaline earths in the next paper."


I am attaching some scans from glass technology books that will explain much more precisely (and knowledgeably) this concept.  I agree with you that glass is such an interesting substance---its properties are unique!

Let me know if you need additional information and I will be happy to contact a glass scientist on your behalf if you need more explanations.