top of page

Light sources

On this page, we discuss lamps for wavelength calibration, and broad-band lamps for absorption spectroscopy

We are developing lamps for UV, visible, and UV-vis spectroscopy.  Tungsten halogen lamps are commonly used to provide broad band visible light in spectroscopy applications.  The most common source of UV light for spectroscopic measurements is a deuterium lamp.

  In addition, we provide guidance below for building a lamp for calibrating your spectrometer based on the emission spectrum of commonly available compact fluorescent lights (CFL).

A fluorescent lamp for wavelength calibration

The cheapest and easiest way to calibrate the wavelength axis for your spectrometer will be with light from a fluorescent lamp.  The spectra to the right (and seen all over this web site) is from a compact fluorescent lamp CFL.


  The light from a florescent lamp is emitted at discrete, known wavelengths.  The wavelengths of light from such a lamp are reported here.  The figure below the spectrum shows a home built CFL light source with a fiber optic connection.  The fiber connection is not necessary to calibrate your spectrometer but helps with keeping things aligned when something is moved about.  Aside from the fiber connection, all parts of this lamp were purchased at the local big-box hardware store.  They include: an electrical box, light switch, ceiling box, a light fixture, some wire and CFL

  Once you can acquire the CFL spectrum, you are ready to use the calibration user interface to set the wavelength calibration parameters and load them into the non-volatile memory of your spectrometer.  Once the calibration parameters are stored in EEPROM, the values will be loaded into the normal user interface any time you use the instrument.  An embedded video demonstrating how the instrument is calibrated can be found on this page.


Tungsten Halogen Lamps for visible spectroscopy

​For a visible lamp, we use 12 volt 10 watt bulbs powered by a 1 amp "wall wart" DC power block.  Output from this lamp is shown in the figures below.  As you can see, light intensity from this type of lamp drops off substantially at wavelengths shorter than 400 nm.


We use the 10 watt bulbs because they are cheap and easy to power and do not generate tons of heat.  If you use higher power bulbs you will need a more expensive power source and if you build an enclosure out of plastic or wood, you run the risk of setting your stuff on fire.

We built a safe and reliable broadband lamp for visible absorbance spectroscopy in a metal electrical box we got at the local big box hardware store and some small pieces of aluminum.  The figure on the upper right shows the sockets for the bulb and the power jack on their aluminum mounts.  The mounts bolt to the electrical box, which is then bolted to a piece of flat aluminum.  On the lower right is the finished lamp.  Notice that the 3D printed parts are mounted remote from the main housing.  This was done to limit heat exposure.  

In this lamp, we use a two pin bulb and a G4 socket.  You can order these from us or find them at your local lighting store or on Amazon.  One thing I have noticed about using commercial lighting bulbs for spectroscopy is the presence of fairly sharp lines in the spectrum.  I suspect the lines are from halogen absorptions.  I suspect that if you spend the big bucks at a spectroscopy store, you will get a bulb with less halogen and the lines will not be as pronounced.  The figure on the lower right shows a blown up section of the figure on the lower left.  These figures both show 15 overlaid spectra.  As you can see, the lines don't move or change in intensity, so they should not affect absorbance spectra.


Deuterium lamps for UV and UV-vis Spectroscopies

Spectroscopy in the ultraviolet is more expensive than in the visible.  Deuterium lamps have power requirements that are more stringent than visible lamps.  They require a high voltage to begin an arc discharge then a lower voltage to maintain the arc.  The lamp itself has to have quartz windows to be transparent to UV and an appreciable amount of deuterium in its atmosphere.

That being said, there is some question in my mind whether the prices charged for UV and UV vis lamps are appropriate.


We have successfully built a UV-vis lamp using components largely purchased from vendors we found on  The only major issue with the lamp is heat dissipation, which would not be a problem if the housing we purchased (figures below) were made entirely of metal, rather than having plastic end plates. 


We had aluminum blocks milled to mount deuterium lamps we purchased.  The figure above right shows a block and bulb.  The forward face of the block has mounting holes for M5 machine screws and a through hole for the light path.  The light path hole passes all of the way through the block to allow mounting a tungsten halogen bulb behind the deuterium lamp (seen in the figure below left).  As shown in the figure below left, the UV-vis lamp requires two AC / DC converters, one for the deuterium lamp (inside the aluminum block) and the other for the tungsten halogen lamp (mounted inside a copper pipe, attached to the aluminum block).  In this configuration, light from the tungsten halogen bulb passes through the aluminum block and a hole in the electrodes of the deuterium lamp and exits co-incident with the UV.

bottom of page