Argon fill, and how long it actually stays in the unit
Argon fill appears on almost every glazing quotation and almost nobody asks the obvious question, which is whether it is still in there in fifteen years. This page is about what it does, how long it lasts, and why the answer matters less than it sounds.
What argon is doing
The cavity between two panes transfers heat by conduction through the gas and by convection as the gas circulates from the warm pane to the cold one.
Argon is denser than air and less thermally conductive. Being denser it also convects less readily, so the circulation that carries heat across the cavity is slower.
Both effects work in the same direction, and the result is a measurable improvement in the U-value for a gas that is inert, colourless, non toxic and cheap.
Why argon and not something else
| Gas | Performance | Cost | Where used |
|---|---|---|---|
| Air | Baseline | Free | Older units, and nothing current |
| Argon | A useful improvement | Very low | Effectively standard |
| Krypton | Better again, especially in a narrow cavity | Considerably higher | Slim units, some triple glazing |
| Xenon | Best | Prohibitive | Effectively never, domestically |
Argon is the point on that curve where the improvement is real and the cost is close to nothing, which is why it became standard rather than optional.
Cavity width, and the interaction
Gas fill and cavity width work together, and the best width differs between gases.
Too narrow a cavity and the panes conduct through the gas too readily. Too wide and convection sets in, circulating heat across the gap, so performance falls again. Each gas has an optimum, and argon’s is wider than krypton’s.
Which is why krypton appears in slim units and in triple glazing, where two cavities have to fit within a limited overall thickness and each one is therefore narrow. In a normal double glazed cavity, argon is doing its best work and krypton would add cost for a small gain.
It leaks, slowly
No seal is perfect. Gas migrates out and air migrates in, slowly, through the seal materials themselves rather than through a defect.
Industry standards set a maximum permitted annual leakage rate for a compliant unit, expressed as a percentage of the fill per year, and it is a low single figure. A well made unit performs better than the limit.
At those rates a unit that started well filled retains a useful proportion of its argon for a long time. It is a gradual decline rather than a failure, and there is no moment at which it stops working.

What the decline is worth
This is the reassuring part and it is worth stating plainly.
The improvement argon gives over air is real and modest. So the loss as it slowly leaks is a fraction of a modest number, spread over decades. A unit that has lost a meaningful part of its fill after twenty years performs slightly worse than it did and considerably better than a 1990s unit with no coating and no gas.
Which means argon is worth specifying because it costs almost nothing, rather than because its retention is critical.
You cannot check it without equipment
There is no way to tell by looking. Argon is colourless, odourless and at the same pressure as the air it replaced.
Instruments exist that measure the concentration in an installed unit non destructively, and they are used in testing and research rather than on domestic call-outs. Nobody is going to come and check your rooflight.
The practical consequence is that the fill is a matter of specifying a reputable product from a supplier whose units are made to the standards, rather than something to verify afterwards.
The seal is what matters
Everything about gas retention comes back to the edge seal, which is also what keeps moisture out.
A sealed unit has a primary seal on the spacer, generally a butyl that is very good at holding gas, and a secondary seal outside it doing the structural work of bonding the panes. The primary seal is the gas barrier and the secondary seal protects it.
So a unit that is losing gas quickly is a unit whose seal is compromised, and that same seal failure will eventually admit moisture and produce visible misting. The gas goes first and the misting is what you see.
Overhead units work their seals harder
A point specific to rooflights and worth knowing.
In a wall, the glass weight goes down through the frame. Overhead, the secondary seal is holding the outer pane against gravity for the life of the installation, and the whole unit is sagging very slightly under its own weight continuously.
Add the daily thermal cycle, which on a south facing rooflight is larger than on any wall, and the edge seal of an overhead unit is working harder than the same unit in a window. It is one of the reasons overhead units should be built for the purpose rather than being a window unit laid flat.
Pressure, and the altitude question
A sealed unit is sealed at the pressure and temperature of the factory, and it flexes when those change.
That is why panes visibly bow slightly in hot weather and dish in cold, and why units transported over high ground are sometimes fitted with a capillary tube to equalise and then sealed. Locally this is not a factor, since the highest ground in our radius is a few hundred feet.
What does matter is that this flexing works the seal continuously, which is part of why edge seals eventually fail. A larger unit flexes more, which is why very large sealed units have shorter typical lives than small ones.

How units are filled
Worth knowing, because it explains why fill quality varies between manufacturers.
Argon is introduced as the unit is assembled, displacing the air in the cavity before the seal is completed. How completely it displaces the air depends on the process, and modern automated lines achieve a high concentration consistently where older or manual methods achieved less.
Which means two units both described as argon filled can start life at noticeably different concentrations. A unit filled to a high concentration on a good line retains a useful amount for far longer than one filled to a minimum, and the difference compounds over decades.
Larger units and the flexing
The size of a unit affects how hard its seal works, and it is worth knowing when specifying a large rooflight.
A larger pane flexes more for the same change in temperature or pressure, because it is less stiff over its span. That flexing works the edge seal continuously, and over decades it is one of the reasons very large sealed units have shorter typical lives than small ones.
On a rooflight, where the unit is also sagging under its own weight and taking a larger daily thermal cycle than any wall, the effect is more pronounced. It is not a reason to avoid large units. It is a reason to buy a good one and to expect a large overhead unit to be replaced sooner than a small vertical one.
What actually ends a unit’s life
Not gas loss. The desiccant.
Inside the spacer is a desiccant that absorbs residual moisture and any that gets in over time. Its capacity is finite, and when it saturates the cavity can no longer stay dry and the unit mists between the panes.
That is the normal end of life of a sealed unit, it is gradual, and it is what people actually see. Gas loss precedes it silently and costs a small amount of performance along the way.
Guarantee periods and what they cover
Manufacturer guarantees on sealed units generally cover the seal failing and producing visible misting within a stated period.
They rarely say anything about gas concentration, for the practical reason that nobody can demonstrate it after the fact without specialist equipment. So the guarantee is on the visible outcome rather than on the performance.
That is reasonable and it is worth understanding, because it means the claim you might make is about misting rather than about a U-value having drifted.

Whether to specify it
Yes, without thinking about it. It is close to free, it is standard on current products, and the only way to not have it is to buy an old or unusually cheap unit.
The question worth asking a supplier is not whether the unit is argon filled but what the initial fill concentration is, because a unit filled to a high concentration retains a useful amount for longer than one filled to the minimum.
Any reputable supplier will answer that. It is a reasonable indicator of how carefully the units are made.
Questions worth asking a supplier
- Is the unit argon filled, and to what initial concentration?
- What spacer is in it, and is it warm edge?
- Is the unit made for overhead use rather than being a window unit?
- What is the guarantee period against seal failure, and what does it require of the installation?
- Is the quoted U-value whole unit or centre pane, and at what size?
None of those is an awkward question and any supplier worth using answers all five without hesitation. A supplier who cannot say what spacer is in their units is telling you something.
Where it sits against everything else
Argon belongs in the same category as a warm edge spacer: small, cheap, worth specifying, and not where the large differences are.
The large differences on a rooflight are the age of the unit, whether the kerb or reveal is insulated, whether the vapour control layer is sealed to the frame, and whether the solar control suits the orientation. A unit with argon on an uninsulated kerb loses more at the perimeter than the gas saves across the middle.
Specify it, and do not choose between products on it.
Replacing an old unit
The useful comparison when somebody asks whether their fifteen year old rooflight is still performing.
A unit from the mid 2000s probably has a low emissivity coating and argon, and both are still substantially doing their job. A unit from the 1990s may have neither, and it is losing several times as much heat as a current one regardless of what has happened to any gas.
So the age of the unit tells you far more than any speculation about the fill, and it is what we look at first.
What we specify as standard on the glazing
We specify argon filled, warm edge units as standard and say so on the quotation rather than leaving it implied. We ask suppliers about initial fill concentration on larger orders. And when somebody asks whether the argon has gone out of their old rooflight, we tell them it will have gone down, that it matters less than they think, and we look at the age and the kerb detail instead.
Argon is a cheap improvement that fades slowly. The seal it lives behind is the part that decides the life of the unit.
Thinking about this job? We survey across Worthing, West Sussex and the South Downs and give you a fixed price before any work starts, with a 10-year workmanship guarantee.