Image dissectors

The main detectors used in AES today are photomultiplier tubes (PMTs), photodiode arrays (PDAs), charge-coupled devices (CCDs), and vidicons, image dissectors, and charge-injection detectors (CIDs). An innovative CCD detector for AES has been described [147]. New developments are the array detector AES. With modem multichannel echelle spectral analysers it is possible to analyse any luminous event (flash, spark, laser-induced plasma, discharge) instantly. Considering the complexity of emission spectra, the importance of spectral resolution cannot be overemphasised. Table 8.25 shows some typical spectral emission lines of some common elements. Atomic plasma emission sources can act as chromatographic detectors, e.g. GC-AED (see Chapter 4). 

Image Converter and Image Dissector Cameras. See Vol 2, p C14-R under CAMERAS... 

These tubes have been employed with both one- and two-dimensional dispersive systems. For example, Harber and Sonnek (43) described an electronic scanning spectrometer based on an image-dissector photomultiplier in conjunction with a onedimensional dispersive system. Their system used a 12.7 cm Czerny-Turner mount with a reciprocal linear dispersion of... 

Figure 2. Image-dissector photomultiplier. (1) Sweep-coil electronics (2) Photomultiplier power supply (3) Focus-coil electronics (4) Display (5) Signal amplifier. (69).

In contrast to the image dissector, which measures the photon flux, the vidicon is an integrating device, where the target serves as a memory buffer, storing information until the scanning electron beam reads and erases it. 

Image dissector tube. Figure 2 shows a schematic representation of an image dissector tube. The active surface of the image dissector tube is an S-20 photocathode (2 ) similar to that... 

Figure 3. Three-dimensional representations of the spectrum from a mercury pen lamp recorded with an image dissector and a silicon target vidicon (30)...

After the appropriate dwell time has elapsed, the sample command is issued and initiates ADC conversion of the elemental intensity. The basic camera system is supplied with an 8-bit ADC, however, to take full advantage of the dynamic range available with the image dissector, an auxiliary signal amplifier/12-bit ADC module was incorporated into the system. 

Response characteristics of the silicon vidicon and image dissector for a variety of scan formats were evaluated and have been presented elsewhere (2, 29 30). Only those data most pertinent to analytical applications with the selected operating conditions are included here. 

These data show that for the image dissector there is excellent agreement between experimental and theoretical resolution and that the resolution of the image dissector system is about twice as good as that of the vidicon system. The deviation of the resolution of the image dissector from theoretical at 7024.05 A may be attributed to curvature of field in the reduced image of the spectral focal plane because the line is near the edge of the photocathode. 

It should be noted that the 200 pm slit width used here does not represent the smallest practical slit width that can be used with the image dissector. The photoelectrons from the photocathode of the image dissector are focused so that they produce a 1 1 image of the photocathode surface on a 38 pm diameter circular aperture. Only those electrons that are focused onto the aperture are passed to the dynode chain where they are amplified. However, the 200 pm slit width used here produces a slit image width of 51 pm on the photocathode, and thus, the ultimate resolution should be about 25% better than that shown in Table I,... 

Orders Wavelength (A) Expected Vidicon Image dissector... 

Wavelength accuracy. In order to evaluate the ability of each system to locate spectral lines, a preliminary wavelength calibration was carred out with the emission spectrum of a mercury pen lamp and then the peak maxima of several atomic lines from an iron hollow cathode lamp were located. The root mean square (RMS) prediction error, which is the difference between the predicted and the observed location of a line, for the vidicon detector system was 1.4 DAC steps. Because it is known from system calibration data that one DAC increment corresponds to 0.0125 mm, the absolute error in position prediction is 0.018 mm. For the image dissector, the RMS prediction error was 7.6 DAC steps, and because one DAC step for this system corresponds to 0.0055 mm, the absolute error in the predicted coordinate is 0.042 mm. The data in Table II represent a comparison of the wavelength position prediction errors for the two detectors. 

These values were calculated by multiplying the absolute errors of the predicted positions by the RLD at each wavelength. These data show that wavelength positions can be predicted somewhat more accurately with the vidicon than with the image dissector. 

Comparison of the luminous sensitivities of the image dissector and the silicon target vidicon Image Dissector Silicon Target Vidicon... 

To a first approximation, the noise level for the silicon vidicon is independent of the signal level. For repeated measurements of currents between 4 and 4,400 nA with the image dissector, a log-log plot of imprecision vs. signal is linear with a slope of 0.51 0.07 (30) confirming the expected shot-noise behavior. 

Performance data for the image dissector used to determine Cr, Cu, Fe, Mn, Ni, and Co in synthetic samples by atomic absorption. (With permission, Clin. Chem., 24, 602 (1978).)... 

For the image dissector, a log-log plot of standard deviation vs. current has a zero intercept of -1.48 0.18, suggesting a standard deviation of about 0.033 /i jp. This corresponds to a... 

This shows that the detection limit ratios should equal the square root of the image dissector current divided by the silicon vidi-con current, assuming the equal elemental sensitivities observed experimentally. Equation 3e, current and sensitivity data in Tables IV and V are used to compute detection limit ratios included in the last column of Table V. These detection limit ratios show that the image dissector has an advantage by a factor of 4 to 27 over the vidicon for every element and every wavelength examined. While detection limits are useful, it is desirable to compare the performance of the detectors at other concentration levels. 

Relative photometric errors. The fixed current error for the vidicon and the variable error for the image dissector have been discussed under the headings of independent and square-root errors (32J. Using eq 8c from the photometric errors paper (32), and noting that the relative absorbance error, RS, is equal to the relative concentration error, RSq, it can be shown that the relative concentration errors for the vidicon, RSq image dissector, RSp jp, are given by ... 

---