What is the difference between a .ch and a .uv file?

A .ch holds one number per time point: the signal from a single channel. A .uv holds a whole spectrum per time point, so instead of a line you have a surface, with time on one axis and wavelength on the other. That is why a .uv is much larger and much harder to display.

Why is DAD data harder to open than a normal chromatogram?

Two reasons. The file is a two dimensional array rather than a list, so a reader has to know the wavelength axis as well as the time axis, and it is far bigger, so naive reading is slow. Free tools have concentrated on the simpler .ch files first, and several describe their diode array support as experimental.

Which .uv versions can be read?

Versions 31 and 131. Version 31 is the legacy layout with a short header and single byte strings. Version 131 is the later one with a longer UTF-16 header. Both are read by open tools today.

Can I get a single wavelength trace out of a .uv file?

Yes, and it is usually what you want. Once the file is decoded you can take any wavelength that was recorded and read it as an ordinary chromatogram. That is the same thing ChemStation does when it shows you one trace from diode array data.

PeakWright / Guides / Reading .uv DAD data

Reading `.uv` DAD data without a licence

This is the hard one. A .uv file holds a full spectrum at every time point, and it is the Agilent file that free tools handle worst. If your diode array question has gone unanswered on a forum for years, this is why.

What a .uv file is

A normal chromatogram is a list: one signal value per time point. A diode array detector does not measure one wavelength, it measures a range of them at once. So a .uv stores, for every time point, the absorbance across every wavelength that was recorded.

That makes it a surface rather than a line. Time on one axis, wavelength on the other, absorbance as the height. It is the reason a .uv is often several megabytes where the equivalent .ch is a few hundred kilobytes.

This is also why diode array data is worth keeping. A single wavelength trace answers one question. The full array lets you ask later whether two peaks that co-eluted are actually the same compound, which you cannot do once someone has exported you a single channel.

Why it is harder than a .ch

Two reasons, and both are about the shape of the data rather than any protection on it.

It is two dimensional. A reader has to work out the wavelength axis as well as the time axis: where it starts, where it ends, and the step between points. Get that wrong and every spectrum is labelled with the wrong wavelengths.
It is big. An 8 MB file holds roughly a million values. Anything that reads it carelessly is slow, and anything that holds several of them at once runs out of memory.

Because of that, most free tooling started with .ch and got to .uv later. Several projects still describe their diode array support as experimental, which is an honest label rather than a criticism.

Which versions exist

Version	Header	From
31	Short header, single byte strings	Legacy ChemStation.
131	Longer header, UTF-16 strings	Later ChemStation and OpenLab.

Both are read by open tools today. The layout of each, including the offsets that were measured rather than published, is written down in the format reference.

What works today

A browser viewer, for a look

The viewer on this site reads .uv versions 31 and 131 and draws the trace, in your own tab, with the file never leaving your machine. That last part matters more with diode array data than with anything else, because these files are large and people are reluctant to upload them.

Be honest about what a viewer gives you: a picture and the numbers behind it. It is not a replacement for a data system.

rainbow, for Python

rainbow reads Agilent .uv data along with .ch and .ms, and installs with pip install rainbow-api. If you want the full array as a matrix to work on, this is the most direct route and nothing here improves on it.

chromConverter, for R

chromConverter handles ChemStation and OpenLab .uv files in versions 131 and 31. Some of its formats go through external tools you install separately, so check which route yours takes.

OpenChrom, for a full application

OpenChrom imports Agilent diode array data through a converter plugin. Its own documentation has described DAD and UV support as experimental, so treat it as promising rather than settled, and check your result against something else before you rely on it.

The thing you probably want

Most of the time the question is not "give me the whole array", it is "show me the trace at 254 nanometres". Once a .uv is decoded, any recorded wavelength can be pulled out and read as an ordinary chromatogram. That is exactly what ChemStation does when it shows you one trace from diode array data.

If that is your situation, check the .D folder first. Many methods write a .ch for the wavelengths of interest alongside the .uv, and reading the .ch is simpler and faster. See what is inside a .D folder.

One caution about size. A multi hour run at a fine wavelength step produces a very large array, and any tool that holds all of it in memory will feel it. If a .uv is slow to open, that is the shape of the file rather than something being wrong.

Getting the numbers out

Exporting a two dimensional array to CSV needs a decision about layout: one row per time point with a column per wavelength, or a long format with three columns. Which one you want depends entirely on what reads it next. There is more on this in getting ChemStation data into CSV.

Try it on your .uv

Drop the file in and see whether it reads. It runs in your browser, so nothing is uploaded, and if the file is a version we do not handle you get a clear message rather than a wrong chromatogram.

Open the viewer The .uv format reference

PeakWright is not affiliated with or endorsed by Agilent Technologies. "ChemStation" and "OpenLab" are their trademarks.

Reading .uv DAD data without a licence