Wavelength accuracy may be adversely affected (or change) when an UV/VIS detector is serviced/repaired, moved, suffers a physical shock (bumped), large temperature charges occur, a lamp or other optical component is changed, a flow cell is changed, the optics become dirty or contaminated, or due to normal wear and age. The wavelength accuracy of any applicable detectors (e.g. UV, VIS, UV/VIS, DAD, PDA) should be measured on a regular basis as part of "Good Laboratory Practices" (GLP). Depending on the regulations or guidelines applied, most authorities require accuracy to be within 2 to 3 nm of a certified standard within the range used. In practice, we generally achieve accuracy of equal to or better than 0.5 nm across a range of UV / VIS wavelengths. Following good laboratory practice (GLP) requires that we establish the frequency and conditions which determine when they should be verified.plus have complete documentation of these wavelength checks.
We present a few suggestions in how to measure the detector wavelength accuracy of your HPLC UV / VIS module.
- Built-In Test Methods: Most instrument manufacturer's incorporate one or more wavelength accuracy checks directly built into their detectors. This allows quick and accurate measurement of the detector's wavelength accuracy for one or more wavelengths in an automated fashion. The instrument's utilize built-in filters (e.g. holmium oxide) which have been treated with chemicals to provide repeatable wavelength spectra which can be used to determine the accuracy of the detector (and adjust it to within specification in most cases, too). If your instrument has one or more of these built-in test filters, then follow the manufacturer's instructions for using them to measure the wavelength accuracy of your detector.
- Using a solution of high purity ANTHRACENE: Dissolved in an HPLC grade alcohol (i.e. Methanol ) or Acetonitrile (for low UV checks), anthracene has a lambda max of 251 nm. A solution concentration of ~ 1 ug / mL for HPLC use can be injected using a standardized method (SOP) and the area% evaluated, one-at-a-time, at several different wavelengths (for VWD or single wavelength detectors) as follows: 249, 250, 251, 252, 253 nm. Relative to the baseline, the areas should show a peak at 251 nm. If you have a scanning UV/VIS detector (aka: DAD or PDA), then you can scan all wavelengths around the 251 nm region and plot the results using just one run to obtain the same type of data.
- Using a solution of high purity CAFFEINE in HPLC grade water: Caffeine has two useful lambda maximums that we can use for wavelength accuracy checks in the ultraviolet region, 205 nm and 273 nm. We often prepare a range of solutions from 5 ug / mL to 500 ug / mL for linearity testing of UV/VIS detectors, but any of those same solutions could be used for wavelength accuracy checking (similar method as described above for anthracene).
- One of the most widely used methods requires a solution of HOLMIUM PERCHLORATE solution (NIST). Available for purchase from many chemical suppliers, this acidic solution provides excellent signals for calibration at well documented transmittance bands (i.e. 241.1, 287.1, 361.5 nm and many others out to ~ 640 nm, depending on the solution it is dissolved in). The detector's flow cell can be filled with the solution and measurements made. The solution is also available coated onto quartz slides and is in fact what is found and used in most detectors today for their built-in verification. However, you can still prepare your own test solution.
Notes: A reminder that the solution that you prepare the wavelength check standard(s) in will directly effect the results obtained. If you prepare it in a solution which has strong absorbance in or near the region you test, the results may be inaccurate (e.g. a test std dissolved in MeOH used to measure wavelength accuracy at 205 nm would not be an appropriate choice). Make sure your SOPs state exactly which solutions are used, how they are prepared and which flow cell are used to make the measurements! Flow cells with different dimensions (i.e. path lengths) will result in different signal outputs and different background solutions will also result in different results which can not be directly compared. For each test, you must use the same conditions to make all measurements.