by Alexander Scheeline
Chemistry
Portable Visible Absorption Spectrometry: Two Dimensions Instead of One
hink of a visible absorption spectrometer or a visible fluorimeter. What do you see? A scanned diffraction grating and photomul- tiplier? A fixed diffraction grating and a linear charge coupled device (CCD) or linear diode array? These have been mainstays for quan- tification for decades. Are there any other approaches one might take? One alternative is to use a color camera (a rectangular imager) as a detector. “Hey, I can use my digital camera to do spectrometry!” says the eager, but naïve, potential user. For qualitative use, of course this is cor- rect. But for serious, quantitative work, there are problems.
T
The problems are associated with both the detector and the dispersion element. Cameras are optimized to give pictures that are appealing to people who prefer rich, saturated colors. Nonuniform illumination of the camera (due, for example, to lens vignetting) is usually compensated for in software or firmware. Some cameras are smart enough to brighten parts of images that are in shadow. The pictures look better than real life. But a spectrometer should not do such compensation. The imbalance in color is evidence of light absorption or fluorescence!
Spectrometers have to be robust with respect to all sorts of challenges: temperature changes, humidity, vibration, dirt, dust, and mechanical stress. Some of these are not much of a problem in the laboratory, but are severe for portable, field instruments. How can wavelength calibration be done in a 25 °C lab and still be accurate at the South Pole or in the Sahara Desert? Wouldn’t a smart approach calibrate the instrument at the time and point of use rather than in the factory?
As first reported at ScIX/FACSS 2013 and shown as a prototype (the AAH-200) at Pittcon® 2014, SpectroClick (Champaign, IL) is using novel grating, light source, and software technology to allow real-time cali- bration of absorption and fluorescence spectrometers. Instruments are handheld and targeted for on-site, at-sample use. The patented grating arrangement fills a camera with spectral information. Below we will look at an image, explain its unique properties, and contrast the image with that of conventional instruments.
Figure 1 shows a SpectroBurst™ (SpectroClick), a centrosymmetric image with hundreds of diffraction orders all obtained from a single collimated beam. Near the center, dispersion is low, as much as 15-nm pixel–1
. Since
the undispersed 50-µm entrance aperture images to a spot 9 pixels in diameter, the resolution is roughly 150 nm, only as good as the resolution would be from the red/green/blue (RGB) encoding used in BMP and JPG files. Toward the edge of the image, the range from 420 nm to 700 nm covers hundreds of pixels, and demonstrated resolution is approximately 9 nm. If a 25-µm entrance aperture had been used, resolution would improve to 5 nm. It is obvious in the picture that some orders are cleanly imaged, while others are drastically overexposed. Only orders where intensity is transduced approximately linearly are useful for spectrometry.
Figure 1 – A SpectroBurst showing hundreds of diffraction orders of vari- ous dispersions and throughputs. While dispersion depends exclusively on radius, throughput is a function of both radius and the grating geom- etry giving rise to the order.
What about the other orders? Suppose one is doing absorption spec- trometry. The light will disappear from the weak orders at a modest absorbance. The orders previously in saturation will come out of satura- tion. Thus, for absorption spectrometry, as concentration increases, one uses less dispersed orders to keep the dynamic range of the measurement within the dynamic range of the detector. Within limits set by dispersion and stray light, the dynamic range of the grating/detector combination is the product of the dynamic range of the camera times the variation in throughput from order to order of the grating. Simply by spreading the light out more, third order would have one-third the illumination level (per pixel) of first order. Increasing the dynamic range from there is caused by some of the orders being diffracted multiple times by the grating stack and by the vignetting of regions far from the image center. An 8-bit per color camera can provide a dynamic range of 5000, or between 14 and 15 bits. This is not as great as the dynamic range of a good diode array or large pixel CCD, but the camera is significantly less expensive, and the gratings are also low-cost plastic transmission gratings.
The ability to use wide dynamic range is one thing; delivering useful chemical analysis is something else. There are numerous steps between obtaining a SpectroBurst and having a useful spectrum. A significant part of SpectroClick’s development work has been to devise wavelength and intensity calibration algorithms that are fast and adequate for spectro- scopic needs. Simply plumbing the literature is inadequate. Let’s look at some of what is involved.
AMERICAN LABORATORY • 37 • SEPTEMBER 2014
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48 |
Page 49 |
Page 50 |
Page 51 |
Page 52