Guides

Wine Cellar Lighting: The Reason Good Bottles Are Dark Green

Archlior
Wine Cellar Lighting: The Reason Good Bottles Are Dark Green

Wine Cellar Lighting: The Reason Good Bottles Are Dark Green

A wine cellar is the only room you will design where the lighting specification has a chemical consequence for the contents.

Not a comfort consequence. Not an atmosphere one. A measurable chemical reaction, with a documented mechanism, that happens in a narrow band of wavelengths some light sources emit heavily and others barely touch.

The winemakers already know this, which is why the evidence is sitting in every wine shop. Good champagne comes in dark green glass, almost black. Serious white burgundy is green too. Cheap prosecco comes in clear glass, and so does most rose, because clear glass sells and those wines are drunk young.

That glass is not a styling decision. It is a filter.

So the cellar you are drawing has two jobs that pull against each other. It has to be a room somebody wants to walk into, with material and depth and atmosphere. And it has to not slowly damage what it was built to hold.

This guide covers both: the chemistry that constrains the specification, then how to design a good room inside that constraint.


What Light Actually Does to Wine

The fault has a name. Winemakers call it light-struck taste, or goût de lumière, and the same phenomenon is known in beer and in milk.

The chemistry is well established across a substantial body of peer-reviewed research.

Wine contains riboflavin, vitamin B2, mostly produced by yeast during fermentation. Concentrations vary considerably by yeast strain and can exceed 150 micrograms per litre.

Riboflavin is a photosensitiser. When it absorbs light, particularly between 370 and 450 nanometres, it jumps to an excited state and becomes chemically reactive.

In that state it attacks methionine, a sulfur-containing amino acid present in wine at an average of three to four milligrams per litre.

The products are sulfur compounds, principally methanethiol and dimethyl disulfide, and they have extremely low detection thresholds. The published descriptors are consistent and unappealing: cooked cabbage, rotten eggs, onion, garlic, wet wool, damp cardboard.

And the fruit disappears too. The research describes the fault as arising from two directions at once, the loss of floral and fruity notes alongside the development of the off-flavours.

There is a threshold, which is useful to know. Research found that methionine degradation does not occur when riboflavin sits below 50 micrograms per litre. Above that, the more riboflavin present, the more sulfur compounds form and the stronger the fault.

And it is faster than most people expect. In controlled laboratory exposure, riboflavin was completely degraded within two hours.


Which Wines Are Vulnerable

Not all equally, and the pattern is worth knowing because it changes what you are protecting.

Sparkling wine is the most exposed. Wines made by the traditional method undergo secondary fermentation in the bottle and spend extended time on lees, which is exactly what raises riboflavin content.

White wine is next, and clear glass is the worst case. Research notes that light-struck taste occurs most frequently in white wine bottled in clear glass, and that it causes meaningful economic losses to producers.

Red wine is comparatively protected, because its polyphenols absorb in the relevant range and act as a natural shield. Comparatively protected, not immune.

Which produces a useful design principle. In a mixed cellar, the champagne and the whites are what your lighting decisions are really about.


The Part That Changes How You Specify

Here is where the chemistry becomes a lighting decision, and it is unusually specific.

The damage happens between 370 and 450 nanometres. So the question for any light source is simple: how much does it emit in that band?

Fluorescent lighting is a poor choice, for a precise reason. Fluorescent and compact fluorescent lamps produce characteristic mercury emission spikes at 405 and 436 nanometres. Both of those sit squarely inside riboflavin's absorption range. A fluorescent tube is, in effect, aiming at exactly the wavelengths that cause the fault.

Warm white LED is measurably better. Research measuring the emission spectra of different sources found that warm white LED lamps have comparatively low output in the critical range, and notes that newer LEDs minimise or eliminate emission between roughly 370 and 442 nanometres.

And daylight is the worst of all, being broad and strong across the entire damage window as well as being uncontrolled.

So in a wine store, lamp choice is not a comfort decision. It is a chemistry decision, and it is one of the very few interiors where that is literally true.


What That Means for Source Selection

Four decisions follow directly from the chemistry. Everything after this section is design.

Specify warm white LED, 2700K or lower. Warmer sources carry less energy at the short-wavelength end by definition, which is exactly the end that matters. It also looks better on timber and stone, so there is no trade-off.

Avoid fluorescent entirely, including legacy tubes in an existing cellar. This is the clearest single recommendation in this article.

Confirm UV output with the manufacturer. Quality white LED emits essentially no ultraviolet, a genuine advantage over the halogen and fluorescent sources it replaced, but verify rather than assume.

Control duration, not just intensity. Damage accumulates as intensity multiplied by time, the same principle that governs artwork and textiles. A cellar is occupied for minutes at a time, so occupancy sensing with a short auto-off delay reduces total annual exposure by an order of magnitude compared with lights left running. It is the highest-value control decision in the room, and it costs almost nothing.

And keep daylight out. A cellar with a window has an uncontrolled, broad-spectrum source varying through the day. If glazing exists it needs blinds, and no bottle should ever sit in direct sun.


Designing the Room, Not Just Protecting the Bottles

A note on what follows. Everything above rests on published photochemistry research. Everything below is design practice: dimensions, positions and levels drawn from how these details work in comparable applications, not from any standard, because none exists for wine storage. It is offered as a considered starting point rather than as a requirement.

That said, this is the part about making a cellar worth walking into.

The tension you have to resolve

A cellar has two jobs that pull against each other.

Storage wants darkness. That is the whole argument of this article.

Assessment wants light. Judging a wine means looking at its colour, its clarity, the depth at the rim, the bead in a sparkling wine. That needs a decent level and accurate colour rendering, and it cannot be done in the gloom that suits storage.

Resolve it by separating them in space and in circuit. The racking sits in low light. The tasting or pouring surface gets its own, brighter, separately controlled zone that is only on when someone is standing there. Two circuits, two levels, one room.

The tasting and pouring zone

If there is a table, a counter or a decanting surface, treat it as the one place in the cellar where colour accuracy matters.

Higher level, locally. Somewhere around 200 to 300 lux on the surface is a reasonable working target, which is bright enough to read a label and judge a colour without lifting the whole room.

CRI 90 minimum, and ask for R9. Wine colour is almost entirely in the red and amber part of the spectrum, which is exactly what the headline CRI figure fails to capture. A source with poor deep red rendering will make a young red look brown and an aged one look dead. This is the same specification point as a butcher's counter, for the same optical reason.

Keep the source out of the glass. A fixture reflected in a glass or in a decanter is distracting and it interferes with judging clarity. Light the surface from above and slightly forward, and check it by standing where someone would actually pour.

A neutral surface helps. A dark timber counter under warm light makes every wine look darker than it is. A pale stone or a neutral tray gives the eye a reference.

On its own switch. The tasting light should never be part of the general cellar circuit, because it defeats the exposure control that protects everything else in the room.

The entrance and the transition

Most cellars are entered from a considerably brighter space, and that transition is usually ignored.

Walking from a lit hallway into a dark cellar means arriving unable to see. The eye needs time, and in a room with a step, a threshold or a stone floor that is a practical problem as well as an unpleasant one.

Give the entry zone a slightly higher level than the racking, so the eye has somewhere to land on arrival before adapting further. It also makes the room feel considered rather than abrupt.

And light the threshold itself. Cellars very often have a step down. A low level marker or a lit nosing solves it, and the same argument applies here as on any stair.

Lighting the architecture

This is what turns a store into a room.

Graze the stone. A cellar is usually built from materials worth revealing: stone, brick, board-formed concrete, timber. Raking light along those surfaces gives the room depth and character while putting almost no light on the bottles. It is the ideal technique here, because it lights the architecture and leaves the collection in shadow.

Light the racking structure, not its contents. A concealed strip washing the face of a timber rack reveals the rhythm of the joinery and the geometry of the bottle ends without illuminating the bottles themselves. This is the single most useful compositional move in a cellar.

Use the vault or ceiling. Many cellars have an arched or vaulted ceiling, and washing it turns the shape into the main feature. It also lifts perceived brightness without adding light at bottle level.

Leave real darkness. A cellar should have shadow in it. The contrast between a few lit surfaces and a dark room is what gives the space its atmosphere, and it happens to be exactly what the wine wants.

Display, and the glazed wine wall

Most clients want to show something, and increasingly the wine is not in a cellar at all but behind glass in a living space.

That changes the risk considerably. A glazed wine wall sits in an occupied room lit for hours rather than a store visited for minutes. It often receives daylight. And its lighting is usually specified for drama, with sources aimed straight at the bottles.

Four moves make it defensible.

Light the racking, not the bottles. Grazing the timber or metal structure gives the dramatic effect while keeping direct light off the glass.

Pick one niche or shelf as the display element rather than lighting the whole wall, and keep the rest of the collection in shadow.

Fill it with the robust bottles. Reds, preferably ones being drunk within a few years rather than cellared for a decade.

Put the display circuit on its own timer, separate from any general room lighting, so it is not glowing quietly at three in the morning.

That gives the client the feature they want, concentrates the exposure on the bottles least vulnerable to it, and leaves the champagne in the dark. If a collection contains anything genuinely valuable, particularly champagne or aged white burgundy, keep it in a dark store and put the robust reds on show. It costs nothing and it is worth raising at design stage.


Fixtures, Positions and Buildability

The decisions above are strategic. These are the ones that end up on a drawing.

Where the light physically goes

Under the front edge of each rack shelf. A recessed linear strip in an aluminium profile, set into the underside of the shelf above, washing down the face of the bottles below. This is the workhorse detail in a cellar, and the two things that determine whether it looks good are the same as anywhere else: a diffuser so you see a line rather than dots, and a deep enough reveal that the strip itself is never visible from standing height.

In the vertical divider between bays. A strip recessed into the side of a timber upright, throwing light across the bottle ends rather than down onto them. This produces the rhythm of a cellar rather than a flat wash, and it puts noticeably less light on the bottles than a shelf-mounted strip.

Grazing the stone wall. A recessed linear or a run of small aperture fixtures set 150 to 300mm from the face, raking down the material. Same setback rule as any grazing application, and the same warning: it reveals the finish, so agree the standard with the contractor.

Washing the vault or ceiling. A concealed cove at the springing point of an arch, or at the wall and ceiling junction, throwing light upward. This lifts the whole room and puts nothing at bottle level.

Recessed floor uplights are worth avoiding. They sit where bottles are carried, they collect dust and spilled wine, and they throw light upward directly into the racking, which is the opposite of what the collection wants.

The reveal detail that decides everything

A strip mounted flush under a shelf will be visible. Someone standing in the aisle looks slightly upward at the shelves above eye level and sees the source directly.

Set it back behind a lip. The strip should sit far enough inside the joinery that the emitting surface is hidden from any standing sightline, and the light emerges as a glow from the underside rather than a visible line. Roughly 20 to 30mm of return is usually enough, but the honest test is to check it against the actual shelf heights and a standing eye level on the drawing.

This has to be resolved with the joinery, not after it. A shelf detail that does not accommodate a profile and a return cannot be fixed once the racking is built, which is why the lighting detail belongs in the joinery drawings rather than in a separate lighting package.

Environment, and why it constrains the fixture

A cellar is a cold, often damp, sometimes below-ground room, and that changes the specification.

Humidity is the main issue. Cellars are typically held at relatively high humidity to protect corks, which is the opposite of most interiors. IP44 as a working minimum for fittings within the cooled space, and higher where condensation is likely or where the room is below ground and subject to any water ingress. Confirm the actual conditions with whoever is specifying the cooling, because a wine room and a cool store are not the same environment.

Cold is generally fine for LED, and in fact LEDs perform slightly better at low temperature, which is one of the few places that is true. Drivers are the more sensitive component.

Put drivers outside the cooled space wherever possible. They generate heat, which the cooling system then has to remove, and they are the component most likely to fail and need replacing. A driver in an accessible cupboard outside the cellar solves both problems at once.

Every watt inside the room is a watt the cooling system pays for twice, once to run the fitting and once to remove the heat. It is a small load in a domestic cellar and a real one in a large commercial store.

Maintenance, which is where cellars go wrong

Racking is fixed joinery, often built in, often full.

Ask the question at design stage: when this strip fails, how does anyone reach it? If the answer involves emptying a rack of several hundred bottles, the detail is wrong.

Favour continuous runs that can be pulled from one end, profiles with removable diffusers, and connections that are accessible without dismantling joinery.

Specify a long-life product and a driver you can actually buy in ten years. A proprietary integrated fitting that is discontinued is a genuine problem in a room that will not be renovated for decades.


Wine Storage Lighting Checklist

Source selection

  • Warm white LED, 2700K or lower
  • No fluorescent or compact fluorescent anywhere, including legacy fittings
  • UV output confirmed with the manufacturer
  • Emission in the 370 to 450nm range checked where the collection is valuable

Exposure control

  • Occupancy sensor with short auto-off delay
  • No daylight reaching bottles, blinds fitted to any glazing
  • Display circuit separate from general lighting and timed
  • Light aimed at racking, circulation and structure rather than at bottles

Fixtures and detailing

  • Strip set behind a lip, hidden from standing sightlines at every shelf height
  • Reveal depth checked against actual shelf heights on the drawing
  • Lighting detail included in the joinery drawings, not issued separately
  • Aluminium profile with diffuser specified, no bare strip
  • Grazing setback of 150 to 300mm where stone is lit
  • No recessed floor uplights in circulation

Environment

  • IP44 minimum inside the cooled space, higher where condensation is likely
  • Actual humidity confirmed with whoever specifies the cooling
  • Drivers located outside the cooled space where possible
  • Heat load assessed against the cooling system

Maintenance

  • Access route to every strip identified before joinery is built
  • Continuous runs pullable from one end
  • Removable diffusers specified
  • Product and driver likely to remain available long term

Design

  • Tasting or pouring surface on its own circuit, around 200 to 300 lux
  • CRI 90 minimum with R9 requested at the tasting zone
  • No fixture reflected in a glass or decanter at the pouring position
  • Entry zone at a slightly higher level than the racking
  • Threshold or step lit
  • Stone, brick or vault grazed to light the architecture
  • Racking structure lit rather than its contents
  • Genuine darkness retained in parts of the room

Collection

  • Sparkling wine and clear-glass whites identified as the most vulnerable
  • Valuable bottles stored in the dark rather than on display

Seven Mistakes

1. Fluorescent lighting in a wine store. Its mercury emission spikes at 405 and 436 nanometres sit inside the exact band that causes light-struck taste.

2. Aiming light at the bottles. Light the racking and the room. The bottles do not need to be lit to be seen.

3. Leaving cellar lighting on. Damage accumulates as intensity multiplied by duration, and duration is the variable you control for nothing.

4. Daylight in a wine room. Broad spectrum, strong across the whole damage window, and uncontrolled.

5. Treating a cellar as a store rather than a room. Grazed stone, lit racking and a properly considered entrance turn a cold store into architecture, and none of it requires putting light on the bottles.

6. Detailing the strip after the joinery is drawn. A shelf that cannot accommodate a profile and a return cannot be fixed once built, and the reveal is what decides whether the source is visible.

7. Putting the champagne on display. Sparkling wine carries the most riboflavin because of extended lees contact, which makes it the most vulnerable thing in the collection.


Frequently Asked Questions

Does light really damage wine? Yes, through a well documented chemical reaction. Wine contains riboflavin, vitamin B2, produced mainly by yeast during fermentation, at concentrations that can exceed 150 micrograms per litre. When riboflavin absorbs light between roughly 370 and 450 nanometres it becomes chemically reactive and oxidises methionine, a sulfur-containing amino acid present at around three to four milligrams per litre. The reaction produces sulfur compounds including methanethiol and dimethyl disulfide, which have very low detection thresholds and are described in the research as cooked cabbage, rotten eggs, onion, garlic and wet wool. The fault is called light-struck taste, and the same phenomenon is known in beer and milk.

Why are wine bottles green or brown? The glass is a filter. Dark glass absorbs a substantial part of the 370 to 450 nanometre band that triggers the reaction, which is why serious champagne and white burgundy come in dark green bottles while cheaper wines sold on shelf appeal come in clear glass. Research confirms that light-struck taste occurs most frequently in white wine bottled in clear glass. It is a protective choice rather than an aesthetic one.

What lighting should a wine cellar have? Warm white LED at 2700K or lower, and no fluorescent. This is not a matter of preference. Fluorescent and compact fluorescent lamps produce mercury emission spikes at 405 and 436 nanometres, both of which fall inside riboflavin's absorption range, so a fluorescent tube emits strongly at exactly the wavelengths that cause the fault. Research measuring source spectra found warm white LED has comparatively low output in the critical range. Confirm UV output with the manufacturer, keep daylight out entirely, and put the room on an occupancy sensor.

Which wines are most at risk from light? Sparkling wine first, because traditional method wines undergo secondary fermentation in the bottle and spend extended time on lees, which raises riboflavin content. White wine next, particularly in clear glass. Red wine is comparatively protected because its polyphenols absorb in the relevant range and act as a natural shield, though protected is not the same as immune. In a mixed collection, the champagne and the whites are what your lighting decisions are really protecting.

How quickly does light damage happen? Faster than most people assume. Under controlled laboratory exposure, riboflavin in model wine was completely degraded within two hours. Real-world exposure in a cellar is far less intense than a test rig, but the principle that matters is that damage is cumulative: it accumulates as intensity multiplied by duration. That is why an occupancy sensor with a short auto-off delay is the single most valuable control decision in a wine store, since it reduces annual exposure by an order of magnitude compared with lights left running.

How do you light a wine cellar that also has a tasting area? Separate them, in space and in circuit. The racking sits in low light because that is what protects the wine, while the tasting or pouring surface gets its own brighter zone that is only on when someone is standing there. A working target of around 200 to 300 lux on the surface is enough to read a label and judge a colour without lifting the level of the whole room. Specify CRI 90 minimum and ask for the R9 value, because wine colour lives almost entirely in the red and amber part of the spectrum, which is exactly what the headline CRI figure fails to capture. Keep the source out of the glass, and put the tasting light on its own switch so it does not undermine the exposure control protecting everything else.

How do you make a cellar feel like a room rather than a store? Light the architecture and leave the collection in shadow. Grazing stone, brick or board-formed concrete reveals the material and gives the room depth while putting almost no light on the bottles, which makes it close to the ideal technique here. Wash the face of the racking to reveal the rhythm of the joinery rather than illuminating its contents. If there is a vault or arched ceiling, wash it, since that lifts perceived brightness without adding light at bottle level. And retain genuine darkness in parts of the room, because the contrast between a few lit surfaces and a dark space is what gives a cellar its atmosphere, and it happens to be exactly what the wine wants.

What about the entrance to a cellar? It is usually ignored and it matters. Most cellars are entered from a considerably brighter space, so walking in means arriving momentarily unable to see, which is a practical problem in a room that often has a step or a threshold. Give the entry zone a slightly higher level than the racking so the eye has somewhere to land before adapting further, and light the threshold itself with a low level marker or a lit nosing. It makes the room feel considered rather than abrupt, and it removes a genuine trip risk.

Where should the light strips actually go in wine racking? Two positions do most of the work. Under the front edge of each shelf, in a recessed aluminium profile with a diffuser, washing down the face of the bottles below. And in the vertical dividers between bays, throwing light across the bottle ends rather than down onto them, which produces the rhythm of a cellar and puts less light on the bottles. The detail that decides whether either looks good is the reveal: the strip must sit behind a lip, far enough inside the joinery that the emitting surface is hidden from a standing sightline, so light emerges as a glow rather than a visible line. Around 20 to 30mm of return is usually enough, but check it against actual shelf heights and standing eye level on the drawing.

What IP rating do cellar light fittings need? IP44 as a working minimum inside the cooled space, and higher where condensation is likely or the room sits below ground with any risk of water ingress. The reason is that wine cellars are typically held at relatively high humidity to protect corks, which is the opposite of most interiors, so the environment is more demanding than a dry room of the same temperature. Confirm the actual design conditions with whoever specifies the cooling, since a wine room and a cool store are not the same environment.

Where should the drivers go? Outside the cooled space wherever possible, in an accessible cupboard or plant area. Two reasons. Drivers generate heat, which the cooling system then has to remove, so every watt inside the room is paid for twice. And a driver is the component most likely to fail and need replacing, which is a considerably easier job outside a room full of racking. LEDs themselves cope well with cold, better than at normal room temperature in fact, so the fittings are rarely the problem.

Is a glass-fronted display cellar a bad idea? Not necessarily, but it needs different handling. A display cellar sits in an occupied room lit for hours rather than a store visited for minutes, it frequently receives daylight, and its lighting is usually specified for drama with sources aimed at the bottles. Three adjustments make it defensible: light the racking rather than the bottles, keep the display circuit separate and timed rather than running all evening, and store genuinely valuable bottles in the dark while putting robust reds on display. That last decision costs nothing and is worth raising with the client at design stage.


Notice

This article is published by Archlior for general information and educational purposes. It is written for architects, interior designers and specifiers as an introduction to lighting wine storage.

It is not professional lighting design advice, oenological advice, engineering advice, or a statement of regulatory compliance for any project.

The photochemistry described here is drawn from published peer-reviewed research as cited in the text, including studies on riboflavin-mediated photo-oxidation of methionine in wine. Findings are summarised from published abstracts and papers rather than reproduced in full, and readers should consult the primary sources directly where accuracy matters. Susceptibility to light-struck taste varies with wine composition, yeast strain, bottling, closure and glass colour, and the figures quoted are representative values from published research rather than predictions for any specific wine.

Source selection guidance reflects the published spectral characteristics of different lamp types as measured in the research cited. The design sections covering fixture positions, reveal depths, illuminance levels, IP ratings and detailing are design practice rather than sourced standards, drawn from how comparable details perform in other applications. There is no regulated lighting requirement for wine storage and no published standard governing it, so these figures are considered starting points to be developed with the specific joinery, cooling design and site conditions of a project. Where a collection is of significant value, obtain specialist storage advice.

Wine storage also depends on temperature, humidity, vibration and orientation, which are outside the scope of this article. All electrical work must be carried out by a suitably qualified electrician working to the regulations in force locally.

Archlior accepts no liability for any loss, cost or damage arising from reliance on the information in this article. Readers act on it at their own discretion.

Last reviewed: August 2026.