Rob Morris

When miniature spectrometers were introduced to the instrumentation market in the 1990s, they benefited from a sort of perfect storm of technological circumstances. The development of detectors for mass-volume markets lowered overall system costs dramatically, and the growth of fiber optics made it much easier to "guide" to the spectrometer the light that interacts with samples. Even the evolution of personal computers — which became faster, smaller, and more convenient — allowed spectrometers to process high-speed, high-resolution spectral data.

Yet advances in optics perhaps have played an even more significant role in the evolution of miniature spectrometers. Today, because of new and improved optical bench components and designs, miniature spectrometers are much more than clever curiosities. They are robust analytical instruments found around the world in laboratories, on process lines, and in the field.

Conventional Wisdom Gets in the Way

One of the advantages of being a novice is that you don't know enough to be limited by conventional thinking. That certainly was the case at my company and other early manufacturers of miniature spectrometers. We were smart and resourceful, but we really weren't experts in the "contributing-regularly-to-refereed-journals" sense, in the field of spectroscopy. So, when a customer had a particular need, the first thought wasn't why that requirement couldn't be met, it was more about wondering why it hadn't been done before.

Not surprisingly, the miniature spectrometer pioneers became very good at solving application problems. They already had proved they could make a miniature version of a Czerny–Turner optical bench, an innovation that some very smart people had dismissed, so why not enhance or change the optical components in that spectrometer bench as well?

Thus began a two-decades-long series of improvements, innovations, and insightful applications of new optical technologies that have yielded the highest-performing, most robust miniature spectrometers ever.

And optical components have been at the heart of it all.

Improving Detector Performance

Charge-coupled devices (CCDs) are the detector of choice for miniature spectrometers, offering the advantages of small size, modest pricing, and suitable sensitivity for a wide range of applications. Still, CCDs aren't perfect. Performance tradeoffs can be significant. However, imaginative use of optical coatings and components helps to mitigate those compromises.

Optical Coatings

Although CCDs have the advantage of low levels of readout noise, their polysilicon capacitors absorb UV light and limit response below 350 nm. One remedy to this problem is to coat the CCD array with a phosphor to absorb UV light and emit visible light. This results in a detector with good UV response for many applications. Further refinement of such coatings (both the coating formulations and how and where the coatings are applied to the detector) will yield additional performance gains.

Order-Sorting Optics

Spectrophotometers typically use a diffraction grating for wavelength discrimination. Diffraction gratings are similar to prisms in that white light is spread into a spectrum by redirecting light at angles that are wavelength-dependent. However, each wavelength of light striking a grating is diffracted into several angles. The angles are whole-number multiples, or orders that vary with wavelength. So, for example, first-order light at 200 nm will be diffracted at angle 1 and at angle 2, which is exactly twice as large as angle 1. Angle 2 is also the same as the first-order angle taken by light at exactly twice the wavelength, or 400 nm. If the application requires measurements at 400 nm, and if 200-nm light is also present, then the signal at the pixel in question will include contributions from both wavelengths. (If, on the other hand, there is no deep-UV light, then only 400 nm will be detected.)

The second-order light typically is removed by using filters. For folks using scanning monochromators, accounting for second- and third-order is as simple as popping in the appropriate bandpass blocking filter. But for full-spectrum devices such as miniature spectrometers, this was not an option. Instead, bandpass filters had to be moved into the optical bench light path during an experiment — a suitable solution, but perhaps not as elegant as some other approaches.

Dichroic filters provided another option to the order-sorting challenge. Dichroic filters selectively transmit or reflect light according to their wavelength. These filters can be patterned onto a single substrate, such as a linear variable filter (LVF), that is applied to the CCD's window. LVFs can be designed to match the dispersed spectra and provide the correct blocking at each pixel in an array (versions designed for 200–850 nm and 350–1000 nm are available today). The benefit to this technology is that the filter acts as a second-, third-, and higher-order filter, and because the filter is installed at the time of manufacture, there's never any need to adjust it. Transmission efficiency is affected only marginally.

The development of LVFs for miniature spectrometer optical benches inspired another application: a discrete version of the LVF for use elsewhere in the spectrometer setup. With the ability to combine high-pass and low-pass linear variable filters, manufacturers could use the dichroic technology to spectrally shape the excitation energy from broadband sources used for cuvette-based fluorescence. Now you could manually adjust the center and bandpass wavelengths in an LVF filter carrier (fixture) installed at the cuvette holder.