Hameraj Singh

Noise, filters and detection limits

Abstract The “noise” of Chromatographic baselines has been investigated in regard to the detector, the nature and extent of filtering or smoothing, and the methodologies of qualitative and quantitative assessment: all in order to clarify the role such factors play in the determination and interconversion of some common types of detection limits. This study scrutinizes baselines from the flame photometric detector in single-channel continuous and ten-channel multiplexing versions; it also examines baselines from flame ionization and electron-capture detectors. It makes use of finite impulse response and non-weighted moving-average digital smoothing, as well as three-pole analog filtering. Baseline fluctuations are quantified by the standard deviation derived from the common root-mean-square (RMS) calculation, or from the less common least-squares Gaussian fit; peak-to-peak noise (Np-p) is estimated by procedures including or excluding presumed outliers. Individual results are expressed as the ratio of Np-p measurement and RMS calculation performed on the same data set. A wide variety of such ratios are then assembled from different detectors, filters, and smoothing conditions. They prove conclusively that -contrary to common belief—the conversion factor between the two types of measurements does vary: usually between 4 and 10, but occasionally even farther. Consequently, the conversion factor between the corresponding two types of detection limits varies as well. The Np-p/RMS ratio depends largely on the detector-of-origin, its condition, and the extent to which noise has been filtered. In contrast, the nature and sophistication of the filter hardly matters: either for the Np-p/RMS ratio or for the practical detection limit. This is because the slow undulations characteristic of heavily filtered baselines represent —at least in the detectors we used-dampened fast noise rather than aboriginally slow noise. Corresponding computer simulations, based on amplitudinally random noise smoothed by stationary boxcar or non-weighted moving-average filters, produce results strikingly similar to actual baselines. Simulated fast RMS noise correlates, as expected, with the square root (log-log slope = 1 2 ) of the filters time constant. The corresponding slopes for experimental noise are usually close to 1 2 as well. Most importantly, though, the simulated Np-p/RMS ratio varies strongly with the extent of smoothing -thus mimicking and thereby explaining the behavior of the experimental ratio.

Fundamental noise in three chromatographic detectors

Abstract Counting statistics are used to estimate the minimum theoretical noise of three Chromatographic detectors, by assuming that the standard deviation of their baselines equals the square root of their primary chemical events. These primary events are taken to be the observed generation of photons in the flame photometric detector, the emission of β rays in the electron-capture detector, and the formation of ion pairs in the flame ionization detector. The theoretically estimated and the experimentally observed noise agree in every case. This suggests that baseline noise in the three particular detectors is due, predominantly if not exclusively, to random processes involving the atomic structure of matter: therefore, it cannot be further reduced.

Chemiluminescent photon yields measured in the flame photometric detector on chromatographic peaks containing sulfur, phosphorus, manganese, ruthenium, iron or selenium

Abstract Photon yields — the number of photons generated per analyte atom — are of obvious analytical and mechanistic importance in flame chemiluminescence. However, such numbers are unavailable for spectral detectors in gas chromatography (as well as for most conventional spectroscopic systems). In this study, photon yields have been determined for the chemiluminescence of several elements in the flame photometric detector (FPD). The number of photons generated per atom of FPD-active element was 2×10 −3 for sulfur (emitter S 2 *, test compound thianaphthene), 3×10 −3 for phosphorus [HPO*, tris(pentafluorophenyl)phosphine], 8×10 −3 for manganese (Mn*, methylcyclopentadienyl manganese tricarbonyl), 3×10 −3 for ruthenium (emitter unknown, ruthenocene), 4×10 −5 for iron (Fe*, ferrocene) and 2×10 −4 for selenium (Se 2 *, dimethylbenzselenazole). Total flows, maximum thermocouple temperatures, and visible flame volumes have also been estimated for each element under signal/noise-optimized conditions in order to provide a database for kinetic calculations.

Dual-channel flame photometric detector for sensitive spectrum acquisition and variable-wavelength operation

Abstract A dual-channel flame photometric detector has been constructed with the first channel offering high light-throughput to a variable-interference-filter wheel. The wheel can be used stationary or spinning. The former selects a wavelength for conventional FPD operation, the latter provides 100-point spectra at less than 20 nm optical bandpass. The second FPD channel is of the conventional type. The detection limit ( S / N p−p =2) for phosphorus is 0.2 pg/s for either the stationary wheel or the conventional channel, and 0.1 ng of phosphorus produces a clearly recognizable HPO spectrum. An illustrative experiment shows the change-over from CH to CC luminescence in a methane-doped flame.

Accurate spectral scans of single chromatographic peaks

Abstract A detector equipped with two channels can fetch accurate luminescence spectra either from migrating chromatographic peaks or from other inputs of sharply varying concentration. To achieve this, one (the scanning) channel is continually referenced to another (the constant-wavelength) channel: the ratio of the two hence portrays correctly the intensity distribution of the spectrum. As a model system for testing this ratio approach on single gas chromatographic peaks, a flame photometric detector was connected to a 1 8 - m scanning spectrophotometer. No significant differences were found in relative spectral amplitudes between scanning the ascent versus the descent of a peak. Similar spectral constancy was obtained when scanning non-chromatographic variable inputs. With a view to practical use, three experiments were carried out that demonstrated the sensitivity and/or the spectral resolution obtainable from reference scans of single peaks. To wit, 5 pg of phosphorus produced a clearly recognizable HPO spectrum: the vibrational levels of S2 resolved well; and the presence/absence of atomic iron lines confirmed an earlier postulated energy limit for the chemiluminescent excitation reaction.

Polarization, relaxation and unrestrictedly linear response in a bipolar, constant-frequency electron-capture detector

Abstract To investigate bipolar constant-frequency regimes in the electron-capture detector (ECD), a “tripulser” was built. The tripulser was able to generate unit sequences of up to three pulses, individually defined as to width amplitude and relative position, with 600 ns to 1 s and 0 to 250 V definition ranges. On a commercial 63Ni two-chamber ECD (Tracor), the high-frequency region of bipolar pulsing (ca. 10 to 100 kHz) was explored. The detector showed clear polarization-relaxation (PR) effects within time spans (on the order of 10−5 s) that were commensurate with the theoretical mobility of electrons. Speculative evidence was found to suggest that PR kinetics, as driven by particular bipolar pulse sequences, resulted in changes to the (heterogeneous) charged-particle distribution and effectively allowed higher than usual concentrations of electrons (and cations) to exist in the ECD. Based on this evidence, a bipolar, constant-frequency drive was developed that, when tested on the Tracor ECD, showed good analytical performance. Most important (and in contrast to the behavior of any other unipolar constant-frequency mode) the bipolar (Tracor) ECD yielded strictly linear calibration curves—starting from the detection limit (5 · 10−18 mol/s of α-1,2,3,4,5,6-hexachlorocyclohexane at S/Np-p=2), over three orders of magnitude, all the way to an amount of analyte that totally exhausted the baseline current.

Phase-shift-free subtraction chromatograms from a dual-channel detector

Compared to single-channel analysis, dual-channel correlational chromatography can dramatically improve selectivity. Yet the extent of improvement is often limited by phase shifts due to the slow conventional electrometer/filter instrumentation used prior to signal deduction. Since fast amplification/correlation circuitry can avoid these phase shifts, a suitable electronic module with a millisecond time constant was constructed. In a demonstration experiment, it improved the linear selectivity of a chromium vs. a ruthenium compound to theoretical limits, i.e. by well over three orders of magnitude.

Correlation chromatograms from a triple-detector system

An electron-capture detector, two channels of a reactive-flow detector, and a flame-ionization detector mode were physically combined to investigate whether correlation (subtraction and conditional-access) chromatograms could be obtained from detectors that operated on entirely different principles and used separate, sequentially arranged monitoring chambers. The investigation was carried out with gasoline as a typical complex test sample. It showed that correlation chromatography, performed on different combinations of the three detectors, worked just as well as it had earlier on a solitary flame photometric detector. Relevant characteristics of correlational multi-detector systems are discussed in that context.

Time-integrated spectra from a flame photometric detector

A rotating, variable-wavelength interference filter has been used to acquire spectra from peak or baseline of a dual-channel, high-sensitivity flame photometric detector. On its spectral channel, the device monitors 100 data points from 400 to 700 nm ten times a second, and ensemble-averages them over seconds to (if needed) hours of acquisition time. The spectra can be smoothed, adjusted for background, corrected for filter transmission and photocathode profiles, and plotted with energy or photon ordinates. The extent of spectral fluctuation suggests the emergence of flicker noise at high analyte concentrations.

Occasional sub-ambient temperature programming performed in two “isothermal” gas chromatographs☆

Abstract By temporarily inserting large thermal masses into the column baths of existing “isothermal” gas chromatographs, linear temperature ramps have been obtained from constant energy inputs. Typically, the rise of temperature could be held linear (within ±3°C) over a more than 100°C range. Since the metal inserts can be cooled before insertion, this approach extends thermal programming to subambient temperatures (e.g. down to −130°C in one case of this study). While such a temporary arrangement cannot compete in speed and convenience with gas chromatographs specifically designed for electronic control and liquid nitrogen cooling, it offers the occasional user an alternative that is inexpensive to assemble and simple to use.