B&W Optical Spectrometer
An optical spectrometer is a device used to measure the relatively intensity of different wavelengths in a light source. Essentially, you put light in and it tells you what wavelengths are present in that light.
You could say our eyes are crude spectrometers, they tell us how much of the input light sits with ~100nm of 572nm, 528nm and 430nm. These are really poor specifications for an optical spectrometer. The fact that we have such poor spectrometers attached to our heads has probably contributed significantly to humanities confusion about the nature of light.
Because of the limitations of the human visual system, children are educated in the primary colors of light. This distracts from a scientifically more useful understanding of the nature of visible light as a continuous phenomena.
A spectrometer helps us reveal the continuous nature of light.
Optical spectrometers come is all shapes and sizes. Originally, using prisms (famously by Newton) but more recently using diffraction gratings.
Decent spectrometers are relatively expensive, typically Ocean Optics USB spectrometers cost in the $1000 range. But a few years ago some really interesting “untested” 473nm Raman spectrometer units started popping up on eBay. At that time they were in the $15 range, and for this price you’d get a spectrometer, laser and a few other interesting optical components. Various users had written software to interface with the spectrometer module.
What were these originally designed for? Well, turns out they were sold as the NuSkin Pharmanex, a skin based antioxidant detection platform, which found its way on to Dr Oz.
I find the idea of accurate detection from skin Raman spectrometry somewhat suspect and NuSkin seems to have had a somewhat troubled history in this regard. But these early Pharmanex units seems to be well engineered. It seems like they’ve mostly taken OEM B&W Tek parts and shoved them in a box. Back when they were $15 I picked up a few to use for various projects. Here’s the module itself:
With the cover removed, you can see the optical setup:
The optical block uses the Czerny-Turner configuration, also used in Monochromators, a diffraction grating is used to split out the light and a concave mirror is used to direct the spectrum toward a sensor:
In this case, the sensor is a ILX511. This is a very common Sony Linear CCD sensor, which I’ve played around with a bit before.
Essentially it’s a 1 pixel long camera:
The spectrum will be registered on the sensor. There are a few issues here, firstly the sensor isn’t uniformly sensitive across the spectrum. So, it’s difficult to make quantitive power measurements with this device. For the most part though, this isn’t want I use this instrument for.
The second issue is that the spectrometer requires wavelength calibration. The wavelength as projected on the sensor is not totally linear. As such the software needs to correct the spectrum based on known peaks.
For this you need a reference spectrum. Turns out the easiest way to get a nice spectrum with lots of clean peaks is an old fashioned CFL lightbulb. The spectra from these bulbs is surprisingly clean, and I’ve found them useful in all sorts of projects.
With the spectrometer in good calibration. I find this spectrometer a handy tool for all sorts of projects. Most commonly I use this for checking the wavelength/stability of lasers and identifying specifications of unknown filters.
While there are other tools better suited for quantitive power measurements (which we’ll be looking at later (perhaps consider subscribing?)) for $15 this has been a really handy tool to have around the lab!
Follow on Twitter. All content CC-by-SA.