anchor

color in glass

Pure glass is transparent because nothing in it can absorb a visible photon. Color is what happens when glassmakers deliberately insert something that can — and there are four fundamentally different things to insert.

Pure silica glass is transparent across the visible spectrum for a precise reason: it has nothing that can absorb a visible photon. The electronic band gap of silica is about 9 electron volts — large enough that the energy threshold for electronic absorption falls deep in the ultraviolet, well above the visible range. At the other end of the spectrum, the Si–O bond vibrations absorb in the infrared. The visible band sits in the quiet window between those two edges. On top of that, glass is amorphous — structurally homogeneous at the scale of visible wavelengths — so there is nothing to scatter light either.

What you see when you look through a pane of window glass is a material that has been refined to keep that window clean. To color glass is to deliberately insert something into that window — something that can absorb or scatter photons whose wavelengths range from about 400 nm (violet) to 700 nm (red). What you perceive as the glass’s color is the transmitted spectrum: white light enters, the chromophore subtracts certain wavelengths, and the remainder reaches your eye. The four mechanisms differ in what does the subtracting.

The four mechanisms

Dissolved ions

The oldest and most common mechanism is metal ions dissolved directly in the glass melt — functioning like a dye dissolved in water. Glassmakers have exploited this for thousands of years: cobalt for deep blue, manganese for amethyst, iron for bottle green, copper for turquoise.

The physics is called d–d electronic transitions for transition metals: the outer electrons of ions like Co²⁺, Cr³⁺, or Fe²⁺ are split into energy levels by the surrounding oxygen atoms, and they absorb photons whose energy matches the gap between those levels. Because the energy levels depend on the ion’s oxidation state and the exact geometry of the surrounding oxygen coordination, the same element can give different colors in different glasses — or in the same glass under different furnace conditions.

Rare-earth ions work by a related but distinct mechanism: their absorbing electrons are shielded inside the ion’s electron shell, so their transitions are sharper and less sensitive to the glass host. Neodymium gives a characteristic violet-pink; erbium gives pink; erbium-doped fiber amplifiers exploit the same ion’s sharp 4f transitions at a different wavelength (~1.5 µm in the near-infrared) to amplify telecommunications signals.

→ Transition metal colorants · Rare-earth colorants

Metal colloids and nanoparticles

The second mechanism requires no ions: it uses metallic particles — gold, copper, or silver — precipitated inside the glass at sizes of tens of nanometers. At that scale, the conduction electrons of the particle oscillate collectively in resonance with the incoming light wave, absorbing strongly at a specific wavelength. This is called surface plasmon resonance.

Gold nanoparticles absorb green and blue light and transmit red — producing the deep wine-red of gold ruby glass. Copper nanoparticles give a similar but warmer red. The striking thing about both is that the color does not exist in the melt: gold ruby glass is colorless when it comes out of the furnace. The particles only form during a controlled reheat called striking, and the glassmaker’s skill lies in reading the color as it blooms.

Silver behaves differently in this family: rather than deep-body ruby color, it produces a yellow surface stain when silver ions migrate into the glass and reduce to nanoparticles near the surface — the basis of silver stain in the medieval stained glass tradition.

→ Colloidal color · Silver stain

Semiconductor band-gap crystals

The third mechanism produces the most saturated colors glass can make. Cadmium sulfoselenide (a mixture of CdS and CdSe) can be precipitated as tiny crystals inside the glass. These crystals are semiconductors, and like all semiconductors they have a band gap: photons with energy above the gap are absorbed; photons below it pass through. Because the gap sits in the visible range, this creates a razor-sharp absorption edge — everything above the cut-off frequency absorbed, everything below transmitted.

By adjusting the sulfur-to-selenium ratio, glassmakers can tune the cut-off across the yellows, oranges, and reds, producing the saturated signal colors used in traffic lights and railway semaphores. At very small crystal sizes — below about 10 nm — quantum confinement effects shift the edge further, and the color becomes size-dependent rather than composition-dependent. Glass was, in fact, the medium in which quantum dots were first experimentally demonstrated: Ekimov’s 1981 work used copper chloride (CuCl) nanocrystals in glass, with cadmium-based systems following shortly after.

→ Selenium ruby glass

Structure and scattering: color without a chromophore

The fourth mechanism produces color — or white opacity — without any chromophore at all. The source of the optical effect is physical structure rather than chemistry: particles or layers that redirect light rather than absorbing it.

→ Continue reading below in Structure and scattering

Structure and scattering

When glass contains a fine dispersion of particles or phase-separated droplets whose size is comparable to the wavelength of light, it scatters rather than transmits. The result can be milky opacity, selective coloring, or the dramatic two-color effect called dichroism — depending on particle size, composition, and whether you are looking at transmitted or reflected light.

Opal glass

Opal glass achieves its milky, opalescent appearance through light scattering from a dispersed second phase embedded in the base glass. The second phase can be fluoride microcrystals (calcium or sodium fluoride, precipitated by adding fluoride compounds to the batch) or phase-separated droplets formed when an initially homogeneous melt separates into two glassy regions on cooling — a process called spinodal decomposition or nucleation-and-growth depending on the conditions.

The particles scatter short wavelengths (blue, violet) more strongly than long ones, following approximately Rayleigh-like behavior for very small particles. This gives opal glass its characteristic blue cast in reflected light and a warm, yellowish-orange transmission — the same physics that makes the sky blue and sunsets orange. At larger particle sizes the scattering becomes more uniform across wavelengths, producing the neutral milky white of dense opaline tableware.

The effect is purely structural: remove the dispersed particles and the glass is clear. The color is in the microstructure, not the composition.

Dichroic glass

Modern dichroic glass — used in jewelry, art glass, and architectural panels — achieves its shifting, iridescent color through thin-film interference rather than scattering or absorption. Extremely thin metal oxide layers (titanium dioxide, silicon dioxide, and others) are deposited on the glass surface in a vacuum chamber, building up a precise optical stack. When light reflects off and transmits through these layers, the waves interfere with each other: certain wavelengths reinforce in reflection while their complements pass through in transmission. The result is a glass that appears one color in reflected light and a complementary color in transmitted light — and shifts across the spectrum as the viewing angle changes.

This is a different mechanism from the dichroism of the Lycurgus Cup, the famous Roman cage cup that appears green in reflected light and red when lit from behind. The Lycurgus Cup’s effect comes from gold-silver nanoparticles (colloidal mechanism, above) that both scatter and absorb. The thin-film interference of modern dichroic coatings and the nanoparticle scattering of the Lycurgus Cup produce superficially similar two-color effects through entirely different physics.

Redox: the cross-cutting control

All four mechanisms depend — directly or indirectly — on the oxidation state of the elements involved. Iron in glass can be Fe²⁺ (blue-green absorber) or Fe³⁺ (yellow-brown absorber), or a mixture of both, depending on the oxygen level in the furnace atmosphere and the composition of the batch. Copper can exist as Cu²⁺ ions (turquoise) or Cu⁰ nanoparticles (ruby colloid). The same element, different oxidation state, entirely different color — or no color at all.

Redox control is therefore not one mechanism among four but the variable that governs which mechanism is operating and how efficiently. It is also the tool for decolorizing glass: when the iron in cheap raw sand would otherwise tint the melt green, glassmakers add oxidizing or reducing agents — or small amounts of complementary colorants like selenium or manganese — to cancel the tint and restore transparency.

→ Redox in the furnace

← back to the map