Noise and the OLIS RSM 1000 Vs. Diode Array Systems
diodearraycartoon
rsmcartoon

Since the RSM spends 0.001 seconds to measure the entire spectrum, some noise is inevitable. However, this is a “good” noise, i.e., random ‘shot noise’ from the PMTs which has minimal effect on the quality of the kinetic and spectral fitting results.

The RSM spectrophotometer collects very large data sets with intensity vs time and intensity vs wavelength information. Thus, rather than just single wavelength measurements, which give only the rates, the RSM offers the advantage of wavelength information (scans). This advantage is very reasonable, yielding both more absolute rates and much higher confidence in drawn conclusions.

RSM scans are not “spectra.” RSM scans are “two dimensional data points with wavelength and amplitude information.” Multiple scans become the 3D data matrices which the mathematics use to calculate the constituent spectra and kinetics. Therefore, when looking at our data, do not think of a scan as a spectrum, no matter how much it might look like a “spectrum.” A scan from the RSM is no more a spectrum than a data point is the kinetic trace on the spectrum from a conventional instrument. A scan from the RSM is one of hundreds or thousands of scans (2D data points) which are used as an ensemble, just as hundreds of one dimensional data points are used to form a spectrum or kinetic trace from a conventional spectrophotometer.

Sensitivity to the fourth or fifth decimal place is achievable, in seconds, given sufficient signal. Very high photometric accuracy is assured with photomultiplier tubes as the detectors.

The object of time resolved spectroscopy is to determine the kinetic behavior of a chemical reaction and, if possible, to also obtain absorbance spectra of the reacting species. The ideal high speed spectrophotometer could collect an instantaneous spectrum with no photometric error and no noise. Practically, one collects the spectrum as rapidly as possible. Rapid scans require high photometric sensitivity, since even very bright light sources do not provide many photons (relatively speaking) in one millisecond. To obtain high sensitivity requires high gain settings on the PMTs which results in increased noise, but again, noise which is random, and therefore which has minimal deleterious effect on the analysis of the data.

Noise And Diode Array Systems

Diode array spectrophotometers are heralded as modern, neat, all-electronic, no-moving-parts, etc. However, they have a fundamental limitation: silicon photodiodes are very insensitive. This inherent insensitivity forces the designer of a diode array spectrometer to maximize the light throughput. He does this by using very bright light sources, single grating spectrographs (higher efficiency and higher stray light than double grating monochromators), large entrance slits (relatively poor optical resolution) and long integration times (as the photodiode array sums up the light incident on it over time) which lead to ‘smearing’ of the spectral information over time. Each of these factors contributes to less confidence in the photometric accuracy of the collected spectral information. Errors ‘tricks.’

Further, the optical arrangement of diode array systems forces the sample being examined to be very brightly illuminated with broad-band or white light and forces the detectors to be far from the sample (there is an intervening spectrograph or post-dispersive element) which means that potentially ruinous photolyzing effects occur and that turbid, particulate samples cannot be examined with the highest fidelity.

diodearraycartoon

This less than ideal optical layout causes “noise” (error) in the photometric data which can be very hard to remove; in fact, it is difficult to know for sure whether it is even an issue, since it is inherent in the instrumentation. Thus, while the diode array might appear to provide cleaner looking spectra (depending upon collection conditions), there is the real likelihood of photometric error.

Single scans from the RSM might appear to have more noise, but this is not photometric noise. That is, the RSM is more likely to result in the CORRECT PHOTOMETRIC ANSWER which is easily and ‘automatically’ removed when the data are fitted using the Olis Robust Global Fitting algorithms (our mature extension of factor analysis).

All these issues lead to the same conclusion: Diode array systems are not the spectrophotometers of choice when one wants the best possible time resolved spectra.

The Olis RSM is a dual beam, double monochromator spectrophotometer with all optical characteristics the same as other such spectrometers, with the addition of the highest possible data collection speed. All support hardware is chosen for its contribution to achieving the correct answer photometrically.

rsmcartoon