UV transmittance: the water property that decides whether a UV stage is sized right

Shine a beam of ultraviolet light through a fixed thickness of water and part of it does not come out the other side. The proportion that does is the transmittance. In water treatment the thickness is standardised at one centimetre, so a UVT figure is always shorthand for “at 1 cm path length” — and a figure quoted at a different path length is a different number for the same water.

The behaviour is the Beer–Lambert relationship. Attenuation compounds multiplicatively with the light's path, which is why the arithmetic runs on a logarithmic scale rather than a linear one:

Two consequences follow immediately, and both are useful when reading a specification or a laboratory report. The first is that UVT and absorbance are one measurement written two ways, so a report that gives either can be converted. The second is that transmittance figures do not add and subtract in the way a percentage usually invites — the useful intuition is to think in absorbance when comparing waters, and to read out in UVT only at the end.

The related index a laboratory is often asked for is UV254, the absorbance at 254 nm. UV254 is used as a surrogate for the dissolved organic content that causes most of the attenuation, and it is quoted as an absorbance rather than a percentage. A report that gives UV254 gives you the same information as one that gives UVT, in the other currency.

Transmittance is a function of wavelength, so a UVT figure is incomplete without the wavelength it was measured at. In water treatment the long-standing convention is 254 nm, and that convention is the reason a bare “95 %” usually means 254 nm — and also the reason a bare figure can be misread when the treatment device is not a 254 nm source.

The distinction matters because the two families of ultraviolet source do not emit at the same place in the spectrum. A low-pressure mercury lamp emits close to a single line at 254 nm, which is exactly where the convention sits. A light-emitting diode source emits over a band with its own peak, and that peak is not 254 nm. Water does not attenuate equally across that range, so the transmittance relevant to an LED source is the transmittance at the source's own band — which is a value a 254 nm report does not directly supply.

For a buyer the practical consequence is a single added clause in an otherwise ordinary question. Rather than asking “what UVT does your unit need?”, ask “at which wavelength was that figure measured, and is it the same wavelength as the source in the unit?” A supplier who can answer it is working from a measurement; a supplier who cannot may be repeating a number inherited from a 254 nm convention.

Our own comparison of UVC-LED and mercury-lamp sources sets out where the two families differ as sources. The point here is narrower and only about measurement: whichever family a unit uses, the transmittance that decides its sizing is the transmittance at that unit's own emission.

Both quantities describe light and water, both are reported as a small number, and both are routinely treated as if one stood in for the other. They are different measurements of different things.

Turbidity is measured by how much light is scattered out of a beam by suspended particles, and it is reported in nephelometric units. UV transmittance is measured by how much light is removed from a beam, dominated by dissolved material that absorbs it. A water can be low in turbidity and still have a modest UVT, because the material doing the absorbing is in solution and scatters very little. Equally, a turbidity reading can move without UVT moving, because the change may be in suspended matter that scatters rather than in dissolved matter that absorbs.

The practical form of this is one sentence worth keeping: a turbidity result within whatever limit applies to it is not evidence that UVT is adequate, and a UVT result is not evidence about turbidity. Where a specification depends on transmittance, the figure to obtain is a transmittance figure.

These are the questions that turn “will this unit work for me?” into something a supplier can answer with a measurement rather than a reassurance. They are written to be asked in order, and the first two are the ones that decide whether the rest are meaningful.

1. At what UVT was this unit rated, and at what flow? A rating is a pair. One number without the other is not a condition.
2. At which wavelength was that UVT measured, and does it match the unit's source? This is the clause that catches a 254 nm figure being carried over to a source that does not emit there.
3. What is the minimum transmittance the unit is intended to be applied at? A stated lower bound tells you whether your water is inside the intended application before any discussion of performance begins.
4. If my water measures lower than that, what changes — flow, chamber, or upstream? The three responses in the sizing section are not interchangeable, and the answer tells you which one the supplier expects you to make.
5. How should I have my water measured, and with what report? A laboratory transmittance or absorbance figure at a stated path length and wavelength, taken at the point the unit will occupy.
6. What upstream conditions should the water meet before the unit? Most ultraviolet units are qualified on pre-treated water; the qualifying conditions belong in the document set, not in the conversation after an installation.

The value of asking in this form is that it is answerable. Each question asks for a stated condition rather than an assurance about performance, and a supplier who works from measurements will have the answers. Where a unit's documentation already carries them, the same list works read the other way round — as a check that the documents you have been given are complete. Our own data sheet and document position is written to be read that way, including which parameters are issued per order rather than published.