Steps at the detector change point in a UV/Vis/NIR instrument equipped with a specular reflectance accessory are due to four primary causes:
(1) Lack of total depolarization of the native light beam in the instrument.
(2) Change of either the size or shape of the beam image on the detector from the background correction to the sample.
(3) Difference in the noise profile between the UV/Vis and NIR regions.
(4) Lack of proper optical alignment in the beam path of the instrument.
Only two of the items above (1 and 4) are relevant to a serviceman performing an instillation or alignment of a reflectance accessory.
Item 2 is usually caused by the sample and the step can be more or less severe depending on the individual sample's optical characteristics.
Item 3 is controlled by the energy related parameters set in the method.
Items 2 and 3 are very important because they can be the source of detector steps in instruments where the reflectance accessory and instrument are properly aligned.
Steps at the detector/grating change due to noise are fairly easy to identify. These types of steps may be problematic for the customer but they do not indicate an optical alignment problem that needs to be fixed. This type of step can be minimized by adjusting the energy related parameters of the method (such as slit size, common beam mask size, and integration time. A noise step can be easily identified by performing multiple scans on the same sample under identical instrument parameters. A noise step will not be reproducible. It's direction and magnitude will be random from measurement to measurement. Here are some examples.
Figure 1 below shows three separate measurement runs on a NIST mirror at 8 degrees with a Universal Reflectance Accessory (URA). The appearance of a step at the detector/grating change point is due to the elevated noise of the NIR region in relation to the less noisy UV/Vis region. Note the random nature of both the size and direction of the step due to the random nature of the noise. The high %R of the mirror keeps the step to a minimal level.
 |
| Figure 1 |
Figure 2 shows two separate measurement runs of the blackened quartz window from a microcell. It was measured on a URA at 8 degrees. The %R of this type of sample is an order of magnitude less then the NIST mirror and is an excellent example of a low reflectance sample. Because of the lowered reflectance of this sample, the noise in the NIR region is larger and as a result the step appears more pronounced. But there is no doubt that this step is due to noise and not alignment.
 |
| Figure 2 |
Now let's take a look at another type of sample. Figure 3 is a spectrum of a low reflectance optical coating measured at 8 degrees on a URA. On this scale everything looks good, but what happens when we expand the scale?
 |
| Figure 3 |
This is the same low reflectance coating as above, but with the scale increased. Now it becomes apparent that we have a step at the detector/grating change that is not from noise. The spectrum in Figure 4 has a clear off-set for the %R value between the NIR and UV/Visible. To try and judge the size of the step by an empirical visual means will not be satisfactory. Visual inspection is too dependent on arbitrary scaling to be useful. Let's try another way...
 |
| Figure 4 |
In Figure 5 we have displayed the spectrum full-scale. There is no question at this scale that we have a step that is greater then the noise envelope on either side of the step.
To size the step we obtain the ordinate value on the high wavelength side of the step and the corresponding ordinate value for the low wavelength side of the step. Subtraction of the lower value from the higher yields the delta value of the step. This delta value is meaningless in itself and must be compared to the ordinate value (%R) to be meaningful. If we divide the delta value by the ordinate value at the step wavelength and multiply by 100, we obtain a relative percentage of the step to the ordinate signal.
For the spectrum below the step is about 1.6% of the ordinate value.
 |
| Figure 5 |
In Figure 6 is shown a spectrum of a NIST mirror measured on a URA. The URA is out of alignment and as a result has a step at the detector/grating change. The same procedure was used to obtain the relative step size. The value calculated for the step was about 0.2%R. Note here how looks can be deceiving. Although the relative step magnitudes are about the same for the spectrum below and the one above, the step in the spectrum here appears smaller than the step from the spectrum above. This is due to scaling differences. Note also that the relative here is 0.2% which is much lower than the 1.6% from the spectrum above; thus, the step below is small when compared to the level of reflectance signal.
 |
| Figure 6 |
Expressing the step as a percentage of the total reflectance signal is an excellent way to compare the relative magnitude of the detector change step anomalies.
