Optimizing Spectrometer Resolution and Throughput in Absorbance and Fluorescence Measurements
by Miriam Mowat, Applications Scientist
Modular spectroscopy makes possible thousands of applications by giving users the flexibility to combine optical bench components and accessories in various configurations. The availability of interchangeable slits in the new Flame spectrometer adds even more modularity. Indeed, Flame can be quickly and easily optimized for a particular measurement or technique, without disrupting its internal optics and eliminating any need for recalibration.
Slit size is one of the determining factors in the optical resolution of the spectrometer, along with the dispersion of the diffraction grating and the number of detector pixels. The choice of slit size also involves design trade-offs: A larger slit increases throughput, but at the expense of optical resolution. A smaller slit yields higher optical resolution, but decreases throughput. The optimum choice of slit size requires weighing those two consequences, and depends heavily on the application.
The interchangeable slit design of Flame helps to eliminate design trade-offs, providing users with greater experiment flexibility in the same instrument. In this application note, we demonstrate the benefits of switching from a narrow slit for high resolution in sharp peak absorbance measurements to a wider slit for high throughput in fluorescence measurements
Absorbance of Holmium
For over 40 years, holmium oxide has been a standard for calibration of spectrometers, usually as holmium oxide glass, due to its consistent, characteristic absorbance in the UV and Visible ranges (1), (2). Here, we have taken a simple absorbance measurement of holmium, as its characteristic peaks are sharp and narrow and require good resolution to be properly resolved. We measured a NIST-traceable liquid sample of holmium oxide in a quartz cuvette (Starna Scientific).
Here is the equipment used in the absorbance setup:
The holmium oxide absorbance spectrum was measured with each of the six different slit widths ranging from 5 µm to 200 µm, and then with the absence of any slit. Figure 1 shows the spectrum measured for each slit. The effect of slit size on measured peak width at half maximum of the peak at 530 nm is demonstrated in Figure 2.
It is evident that with increased slit size, there is a broadening effect on the peaks. In cases where the peaks are very close together, it becomes difficult to distinguish them from each other. This is clear in a number of places in Figure 1. For example, in the range from 460-480 nm there are several small peaks that become increasingly smoothed (rounded) as the slit size increases until it is difficult to define them at all. Similarly, if there is a peak with a shoulder it is gradually smoothed until it is not visible with the larger slits, such as at 540 nm and at 640 nm.
This effect can be crucial in applications that require the accurate identification of narrow peaks. With Flame, a user can select a smaller slit to resolve closely aligned peaks for absorbance or emission and later switch to a larger slit for low light fluorescence measurements.
Fluorescence of Fluorescein in Water
For many applications, resolving sharp spectral peaks is not the focus. Instead, when an experiment is limited by the strength of signal or time restrictions on acquisitions – i.e., a short integration time is necessary — high throughput becomes the most important consideration.
This is the case with fluorescence measurements as the signal can be very low, especially at low concentrations of the fluorescent molecule. Often fluorescent markers are used in low concentrations to identify branded liquids without changing their color when viewed under ordinary conditions. This requires knowledge of the lower limit of concentrations that that can be detected. Additionally, certain applications may have very short integration times – measuring samples on a production line, for example – that will affect how much signal the spectrometer can acquire.
In this experiment, we will investigate the effect that slit size has on throughput when measuring the fluorescence of fluorescein in water.
Here is the equipment we used in the fluorescence setup:
The fluorescein samples were illuminated with an LED light source emitting at 470 nm, the peak absorbance wavelength for fluorescein. This excites the fluorescein molecules and causes them to de-excite, releasing photons at a wavelength of 520 nm. The fluorescence emission was measured using the Flame spectrometer at 90o to the illumination direction to minimize light from the excitation source reaching the Flame. To reduce bleed of the excitation light source further, a 495 nm filter was placed between the cuvette and the collection leg optics of the cuvette holder
Using the most fluorescent sample, the 10-4 molar solution, the integration time necessary to reach roughly 65,000 counts was found (Figure 3). This is a very strong signal, just below spectrometer saturation, and demonstrates the basic effect of throughput on signal strength.
Figure 4 shows the reduction in throughput as slit size is narrowed. With smaller slit size, the required integration time rapidly increases as the light that can pass through the slit is greatly reduced. This can have very tangible benefits when speed of measurement is of high importance.
Next, we measured seven different concentrations of fluorescein at the lowest integration time that fluorescein response could be detected. The limit of detection was defined as the spectrum at 520 nm reaching 400 counts above the baseline level. This is a somewhat arbitrary value that was based on an estimate of 3x the standard deviation of the noise. For each slit size and concentration combination, the integration time was reduced (or increased) until this limit was reached (Figure 5).
This data shows the relationship between slit size and throughput, with larger slits allowing the use of far shorter spectrometer integration times. This sort of measurement can give an indication of most the appropriate slit size for greatest efficiency in each acquisition.
It is worth noting that the 10-2 molar solution requires longer integration times than many of the solutions containing a lower concentration of fluorescein. This does not follow the trend of the others, where higher concentrations equate to more fluorescence and thus, lower integration times. This is simply because at this high concentration the optical density of the solution is heavily increased, and the light is not able to pass through as easily. This alludes to another situation where the high throughput of a large slit size can be beneficial, to maximize the signal from samples of high optical density.
It is clear that in this application, the high throughput of the larger slit is more useful than the improved resolution of the narrower slit.
As our experiments demonstrated, decreasing the slit size of the Flame spectrometer improves spectral resolution, but also reduces throughput. Good resolution is crucial in applications that require accurate measurement of sharp, narrow peaks, whereas high throughput is crucial where low signal can be a limitation. With the interchangeable slit design of the Flame, Ocean Optics has made it possible to change the spectrometer’s optical resolution and throughput performance simply by switching these easy to replace slits. Interchangeable slits give users more freedom in spectrometer design, with no need for spectrometer rework at the manufacturer’s facility or recalibration in the field.
1. Homium Oxide Glass Wavelength Standards. Allen, D. W. 6, 2007, Journal of Research of the National Institute of Standards and Technology, Vol. 112, pp. 303-306.
2. Starna. Starna, Reference materials, Holmium Oxide – UV and Visible wavelength. [Online] [Cited: 02 04, 2015.] http://www.starna.com/ukhome/d_ref/wave.html.