Instrumentation Application of a Silicon-Target Vidicon Detector to Simultaneous Multielement Flame Spectrometry D. G. Mitchell, K. W. Jackson, and Κ. Μ. Aldous Division of Laboratories and Research, New York State Department of Health, New Scotland Ave., Albany, N.Y. 12201
A multichannel, atomic ab sorption spectrometer incor porating a vidicon detector is simple to construct and operate. It enables the rapid simultaneous determination of 10 or more elements in solution
Of the three analytical techniques, atomic emission (AE), atomic ab sorption (AA), and atomic fluores cence (AF) spectrometry, only in AE have multielement systems achieved a degree of popularity, with the longestablished photographic method of recording emission spectra and, more recently, the multiple slit-multiple detector (direct reader) type of in strument. Both of these systems leave much to be desired, however, because of the tedious measurement and in terpretation of the photographic spec tra and the alignment and geometri cal difficulties of the direct reader. An extensive range of sophisticated instrumentation is now available for AA, which is simpler and more selec tive than AE and has become better established than the more recently developed AF spectrometry. A few AA and AF systems have been de signed for the simultaneous analysis of more than two elements, but un fortunately, none is commercially available, and this is undoubtedly the most serious drawback of these two techniques. There are now a number of important potential applications for multielement analysis, such as the monitoring of elemental water pollu tants, wear metals in lubricating oils, and toxic metals in biological materi als. In principle, multichannel AF spec trometers are easier to design than AA instruments, since in the former case it is not necessary to direct all of the incident radiation along the same optical path, and so it is a simple matter to focus radiation from a number of hollow cathode lamps and/ or electrodeless discharge lamps into the atomization cell. There are several approaches to multielement analysis, and these have recently been reviewed by Busch and Morrison (1). One approach in volves the use of multichannel detec tors such as linear photodiode or pho
totransistor arrays (2) and television camera tubes. Busch and Morrison have suggested that the latter type is the most promising detector for the future. In such a device, dispersed ra diation leaving a monochromator strikes a photosensitive surface to re lease electrons which develop a charge pattern across a target. The charge density at each point on the target is a function of the incident ra diation intensity, and the position is a function of wavelength. An electron beam searches the target and neutral izes each point of charge to produce a video signal. The photosensitive surface may be incorporated as part of the target, as in the silicon vidicon; or it may be a photocathode from which photoelectrons are accelerated toward a separate target, as in the image orthicon, the secondary elec tron conduction (SEC) vidicon, and the silicon intensifier target (SIT) vid icon. The first application of television tubes to atomic spectrometry was probably by Benn et al. (3), who in 1949 used an image orthicon for AE to examine low-intensity lines during a gaseous combustion reaction. This detector has a glass target, and the electrons in the search beam in excess of those required to neutralize the positive charge at a particular point on the target are reflected into a fivestage secondary emission electron multiplier. The anode current of the multiplier then generates the output signal voltage. These workers ob served that for the short exposure times needed to follow chemical reac tions, the orthicon was about 50 times more sensitive than the photo graphic plate. The orthicon, however, exhibits poor signal-to-noise (S/N) characteristics {4) and has now been superseded by the vidicon, which has a simpler method of measuring the charge pattern, in that the video sig nal is the magnitude of the neutraliz-
ANALYTICAL CHEMISTRY, VOL. 45, NO. 14, DECEMBER 1973 · 1215 A
ing current. A vidicon tube with an antimony sulfide target has been successfully applied to molecular spectrometry (5), but this target does not show a linear response to photon flux, and a correction must be applied. The most useful vidicon tubes for the analytical chemist will probably be the silicon tube, the SIT tube, and the SEC tube, which has a KC1 target. The useful characteristics of these tubes are compared in Table I. Their relative merits have been described in detail elsewhere (6). Briefly, the SIT is useful for measuring low light intensities, since it provides a higher gain than the SEC vidicon. The SEC, however, has a much lower dark (leakage) current, permitting long integration by accumulating charge on the target without seriously affecting the S/N ratio. The SIT is more expensive than the SEC, but it is also far more robust and less likely to be damaged from overexposure. For AA, the silicon vidicon is probably adequate, and in this application the higher cost of the SIT and SEC vidicons would hardly be justified. However, for AE and AF, where low light levels must often be detected, an SIT or SEC tube with its higher gain would be preferable. Margoshes (7, 8) proposed the use of an SEC vidicon in the construction of a computerized AE spectrometer. This camera tube was to be used in conjunction with an échelle monochromator to simultaneously measure signal intensities at up to 2048 wavelengths. The tube can be programmed (9, 10) such that instead of the electron search beam regularly sweeping the target, a number of Χ Υ coordinates are stored in the comput er memory and these direct the elec tron beam to a number of predeter mined positions on the target surface. This random access system increases the dynamic range of the vidicon by a time integration technique. Areas of high illumination are read more fre quently than weakly illuminated areas, thus permitting more charge to accumulate in the latter. The SEC vidicon described by Margoshes has a UV-transmitting sapphire window and is responsive at wavelengths as low as 165 nm. Such a system would almost certainly be the most suitable for emission work. (A detector of this type with random access program ming is marketed by EMR Photoelec tric, Princeton, N.J.; but like most commercially available vidicons it does not respond in the UV and would consequently be of little use for atomic spectrometry.) A silicon-target vidicon with good response down to about 200 nm is commercially available from SSR In struments Co., Santa Monica, Calif. Although its S/N ratio is generally
Table 1. Comparison of Vid icon Tube Characteristics Silicon target
Dynamic range 103 Internal gain 1 Dark current, electrons sec - 1 cm"', 10 20°C 10 Quantum efficiency, 400 nm 50 Maximum integration time, s e c 3 Relative cost 2
SIT
SEC
103 2 X 103
102 10*
ΙΟ'»
103
15
12
3 10
10* 7
a Depends on temperature and area per resolution element.
somewhat poorer than that attainable with a photomultiplier tube, the in strument can simultaneously detect signals from at least eight elements. A spectrometer incorporating this de tector is described below. It is easily constructed, and the system is no more complex to set up than most single-element instruments. A large number of optical channels give it a great deal of flexibility, and it pos sesses the highly desirable feature that it can be used with atom cells producing transient signals. Application to Flame Spectrometry
During AE measurements, a highresolution detection system is re quired to isolate the spectral line from the often complex background emission spectrum. For the vidicon to be useful as a multielement detector, a spectral range across the target of at least 150 nm is desirable. A monochromator with a sufficiently low lin ear dispersion to give this range has a relatively poor resolution, and during emission work it might be necessary to use an échelle system, such as that of Margoshes, to obtain adequate resolution. The resolution required for AA and AF is much lower, and the vidicon should be better suited as a detector for these two methods. At the time of this writing, no results have been published for multielement AF spectrometry using a vidicon. An SIT, however, with its much improved sensitivity, should make the system feasible for the detection of fluorescent signals. The principal difficulty during AA work is the multiplexing of sufficient suitable light sources along the same optical path, and this problem is not encountered during AF measurements when it is only necessary to focus all the lamps at a point within the atomization cell. From this consideration AF is a highly attractive
1216 A · ANALYTICAL CHEMISTRY, VOL. 45, NO. 14, DECEMBER
1973
method. It has already been shown that the vidicon should be feasible for multielement AA (11), and this method possesses the advantage that an SIT is not required. Also, a variety of hollow cathode lamps are now available containing up to six elements; and if, say, eight elements are to be determined, it is often possible to detect all of them simultaneously by multiplexing two multielement lamps. Other disadvantages of both AA and AF would be common to any multielement instrument. A compromise in flame conditions and in the viewing position within the flame is always necessary, and the dynamic range presents a limitation during the simultaneous analysis of elements with widely differing concentrations. Design of a Multielement Atomic Spectrometer. To use a vidicon as a detector for AA, the arrangement need differ only slightly from the standard single-channel system. The vidicon tube is mounted on a suitably modified monochromator in place of the exit slit/photomultiplier assembly, and the usual detector electronics are replaced by the control console and an oscilloscope. A strip chart recorder can also be used if desired, although this is not essential. A typical instrumental arrangement for AA, which we have developed in our laboratory, is shown in Figure 1. The source is a multielement hollow cathode lamp (JarrellAsh, Waltham, Mass.), and a singleslot burner/nebulizer (Techtron, Palo Alto, Calif.) is used. A second lamp can be mounted in such a way that the radiation from both lamps is combined using a half-silvered quartz mirror. The lamp, lenses, burner/ nebulizer, and monochromator are mounted on an optical rail, and the system could easily be constructed in any analytical laboratory. This is a feature unrivaled by any other system reported to date. The detector shown in Figure 1 is the SSRI Model 1205 Optical Multichannel Analyzer (SSR Instruments Co.). This is a UV-responsive silicontarget vidicon, and its operation has been described in adequate detail elsewhere (1, 12, 13). The variation of spectral response with wavelength for this system is shown in Figure 2a. The quantum efficiency at 200 nm should be about 10%, similar to that of a IP 28 photomultiplier. One of the most sensitive photomultipliers at 200 nm is the R106 type, with a quantum efficiency of about 25%. This vidicon tube is suitable for AA, but for emission studies (AE or AF), an SIT should be more useful because it provides an electron gain of about 2000 over the silicon vidicon
Figure 1 . M u l t i e l e m e n t a t o m i c a b s o r p t i o n s p e c t r o m e t e r i n c o r p o r a t i n g silicon v i d i c o n d e t e c t o r A, source; B, burner; C, monochromator; D, vidicon tube; £, electronic console; F, oscilloscope display
(Table I). The SIT commonly uses an S-20 photocathode with a fiber-optic faceplate, and the response of this system does not extend into the UV (Figure 2b). It is not possible to provide UV response directly because a quartz fiber-optic faceplate is not available. An indirect method is thus required, and an SIT should soon be available in which a coating of a suitable fluorescent substance such a s p terphenyl is used to convert radiation at UV wavelengths into visible light. The quantum efficiency of the silicon vidicon (Figure 2a) is about 21% at 253.7 nm; the UV-responsive SIT has a predicted quantum efficiency of only 1.0-1.5% at this wavelength. Thus the electron gain of 2000 should result in a sensitivity increase of approximately 120. The dispersive system shown in Figure 1 is a 0.25-m Ebert monochromator (Jarrell-Ash). The spectral response of the vidicon is dependent upon the accurate focusing of radiation onto the surface of the target (or the faceplate, when an SIT is used). With only slight modification of this monochromator, the vidicon target can be positioned in its focal plane. This is achieved by removing the exit slit assembly and machining off the flange which remains on the monochromator housing to leave an aperture approximately 30 mm in diameter. The vidicon tube is mounted over this aperture, and the focal plane can be adjusted to give a sharp image on the target by moving the adjustable entrance slit in and out. A diffraction grating blazed at 190 nm and with 295 grooves/mm is suitable for AA and AF. This gives a linearly dispersed spectrum across the target covering a range of 168 nm. Thus, any elements with resonance lines from, say, 200-368 nm can be si-
Figure 2a. V a r i a t i o n of spectral response with w a v e l e n g t h for U V - s e n s i tive silicon vidicon showing p e r c e n t a g e q u a n t u m efficiency. b. Responsivity of S-20 p h o t o c a t h o d e s y s t e m u s e d in S I T
multaneously detected. The spectral range could be further increased by using a grating of even lower dispersion, but this would result in a loss of resolution. The grating described provides adequate resolution for most AA and AF determinations, since, for example, it enables the adjacent lines Pb 283.3 nm/Mg 285.2 nm and Cu 324.8 nm/Ag 328.1 nm to be easily resolved. The same instrumental system could be used for AF, except that a circular burner head would be used with a series of high spectral output sources irradiating the flame. The
number of elements to be simultaneously determined should be limited only by geometrical considerations in aligning the various light sources and by the spectral range provided by the monochromator. Measurement Technique. The electronic console of the SSRI Model 1205 vidicon contains two shift register memories. A video signal may be observed in the real time mode, in which case it is updated at the end of every electron search beam frame scan; or the accumulated data from a preset number of frame scans can be entered into one of the memory units.
ANALYTICAL CHEMISTRY, VOL. 45, NO. 14, DECEMBER 1973 · 1217 A
During AA measurements, a timeaveraged signal is obtained if the accumulate mode is used, and this results in a better S/N ratio. Also, the memory capabilities of the instru ment can be utilized to give a direct reading in terms of absorption. The pure solvent is first aspirated into the flame, and the accumulate preset control on the console is ad justed to count the data from, say, 100 frame scans. This information is stored in the first memory unit (memory A). The time to accumulate 100 such scans is approximately 4 sec. The analyte solution is then aspirat ed, and data for the same number of frame scans are stored in the second memory unit (memory B). A push button operation then takes the dif ference (A — B), and the resulting spectrum is the net absorption signal (10 —It) obtained simultaneously for each wavelength. This appears as a series of peaks on the oscilloscope screen. A cursor control permits the selection of any one of 500 vertical channels across the target area, and a digital readout is obtained in terms of intensity at the channel selected. Typically, each absorption peak has a half-width of about five channels. The peak height may be read by se lecting the channel at the center of the peak, or the peak may be inte grated by moving the cursor across the peak and summing the intensities for all the channels within the peak area. To obtain a signal in terms of per centage absorption, the net signal (A — B) is divided by the signal ob tained during aspiration of the sol vent A : A 100
h
~ It
= % absorption
* ο
The analog output, as displayed on the oscilloscope, can be plotted out on a strip chart recorder, if desired. The measurement technique for AF would be similar to that for AA, except that the analyte aspiration signal should be stored in memory A and the pure solvent aspiration signal in memory B. The difference (A - B) would then be a spectrum in which peak height (or area) would be a measure of fluorescent intensity at a particular wavelength and would be proportional to the analyte concen tration. Typical Results for Multielement AA
The photographs in Figure 3 are the spectra seen on an oscilloscope screen during the simultaneous deter mination of equal concentrations of the eight elements Zn, Cd, Ni, Co, Fe, Mn, Cu, and Ag in an air-acety-
Figure 3. Oscilloscope displays dur ing simultaneous multielement deter mination of eight elements by atomic absorption spectrometry Top, memory A; center, memory B; bottom, A - Β
lene llame using the apparatus de scribed above. The spectral range covered is 232.0 nm (Ni) to 328.1 nm (Ag), and the three spectra are, re spectively, the accumulated contents of memory A, memory Β, and the arithmetic difference A — B. Radia tion from two multielement hollow cathode lamps was combined, using a half-silvered quartz mirror as a beam splitter. These results are only qualitative, but it is clear that the resolution of the system permits at least eight ele ments to be determined simulta neously, and probably many more if suitable light sources can be multi plexed. Quantitative results from this system are now becoming available, and we hope to publish them together with sensitivity and S/N data in the near future. Results comparing sin gle-element AA on the SSRI vidicon
The vapor-deposited, thin-film thermopile sensor at the heart of the Mettler scanning calorimeter is revolutionary for a commercial instrument. As a sensor it is fast and responsive—assures excellent heat contact with crucibles. It is ideally suited for accurate purity determinations, as well as heat of fusion, specific heat, and many more measurements. The modular TA 2000 DSC system offers everything you want in thermal analysis. It is more sensitive and reproducible than most comparable instruments. It offers 0.17 millicalories per second full scale sensitivity. And calorimetric reproducibility of ± 0 . 5 % . Noise level is less than ± 1 microcalorie per second. Temperature range covers —20 to 500°C, with precision of ± 0 . 1 °C, and isothermal fluctuations of ± 0 . 0 0 2 ° C . Extremely linear selectable heating and cooling rates in steps of 0.1 °C. Temperature programs can be controlled remotely. It is completely computer compatible. For your copy of the 8-page Mettler TA 2000 brochure, circle the number.
Mettler CIRCLE 1 4 3 O N READER SERVICE CARD
A N A L Y T I C A L CHEMISTRY, VOL... 45, NO. 14, DECEMBER 1973 · 1219 A
with those obtainable by conventional analysis on a Techtron AA-120 spec trophotometer have already been published (11), and comparative sensitivities and S/N data are repro duced in Table II. The adequate reso lution of the vidicon system is mir rored in the sensitivities, which in most cases are comparable. The vidi con detection limits, however, aver age about a factor of seven poorer. The only reason why multielement results might differ from those by sin gle-element analysis on the vidicon is that the spectral intensity within a multielement lamp varies considera bly from element to element. For ex ample, in Figure 3 the Ag line at 328.1 nm is far more intense than the Cd line at 228.8 nm, although Cd has a higher sensitivity. It is thus possible that a slightly poorer S/N ratio will be obtained at these weaker lines. These peaks, however, appear weaker than they actually are, since the gain on the analog output was decreased to permit the larger peaks to appear fully on the screen. Signal/Noise Theory In the following discussion, the ef fects of detector and source are con sidered in comparing theoretically the S/N characteristics of (a) the multi channel AA spectrometer described above with (b) a single-channel dcoperated spectrometer using an R106 photomultiplier tube. (a) The signal current produced by a silicon vidicon may be expressed by: Is =
QeGAr(N)YT
amps
G = target gain, G — 1 for a silicon vidicon Ar = area of target scanned, cm 2 Ν = incident light flux, pho tons cm~ 2 sec" 1 y = slope of the current vs. luminous flux curve, γ = 1 for the silicon vid icon Τ = detector exposure time, sec Ta = active readout time to scan A r, sec For this system it is necessary only to consider source flicker noise and the video preamplifier noise, since dark current and shot noise do not significantly affect the detector per formance. (In the case of the SIT and SEC vidicons, which possess consid erable target gain, allowance has to be made for shot noise.) The preamplifier noise is expressed as an equivalent noise input current (ineq) which is defined by: ne ineq = ~7jr a m p s r m s where η = electron uncertainty per target element scanned in time Ta, sec. The noise due to source flicker is expressed (14) as a function of the signal current: if = Ι,,χΑ/
112
amps rms
(3)
where χ = flicker factor, typically ΙΟ" 2 Δ/ = observation bandwidth The overall noise then becomes:
(1)
i = V ine* + if2
* a
where Q = quantum efficiency of the photosurface e = electron charge, cou lombs
(2)
(4)
(b) The anode current of a photomultiplier tube is given by: /„ = GANQe
amps
(5)
Table II. Comparative Results by Single-Element AA: Vidicon vs. Techtron AA-120 Concn at which RSD deter mined, ppm
Vidicon Sensi tivity," ppm
RSD, %
Techtron AA-120 Detec tions limit, ppm
Sensi tivity," PPm
Z n 213.9 0.5 0.03 3.5 0.03 0.01 Pb 217.0 10.0 0.5 1.1 0.4 0.5 1.0 Cd 228.8 0.05 2.6 0.04 0.04 10.0 0.4 N i 232.0 1.9 0.3 0.2 10.0 Co 240.7 0.4 2.7 0.3 0.4 Fe 248.3 5.0 0.3 5.2 0.3 0.3 M n 279.5 2.0 0.09 0.1 1.8 0.1 10.0 Pb 283.3 0.9 0.9 5.0 0.5 Mg 285.2 0.5 0.02 2.0 0.01 0.01 2.0 Cu 324.8 0.1 0.1 1.8 0.07 5.0 Ag 328.1 0.1 2.1 0.1 0.1 Ag 338.3 5.0 0.2 2.6 0.2 0.2 Cr 357.9 10.0 0.4 4.3 0.1 0.2 Ca 422.7 5.0 0.2 1.7 0.08 0.1 " 1 % absorption. 6 S/N = 2/1. Reprinted from Réf. Il, p 320, by courtesy of Marcel Dekker, Inc.
RSD, % 0.42 0.49 0.46 0.25 0.78 1.1 0.63 0.48 0.39 0.51 0.53 0.48 0.37 0.65
«ill
Element and wavelength, nm
0.005 0.2 0.005 0.05 0.03 0.07 0.01 0.1 0.001 0.01 0.009 0.02 0.02 0.006
Want to automatically determine melting and boiling points without any possibility of misinterpreting results? Determine exact melting ranges of polymer films and fibers, liquid crystal preparations, and submicrogram quantities of organic compounds using thermal microscopy? Obtain digital dropping points of lubricants, food fats, paraffins and waxes? Softening points of fats, coal tar pitches, polymers and resins? The new generation of instruments, the modular Mettler FP system, will do them all faster, easier, more accurately. There's a separate Mettler FP furnace for each application. All provide high-precision, electronically controlled, linear (to ±0.1 °C) heating and cooling with extremely fast response over the - 2 0 to +300°C range. Temperatures at which significant thermal events occur are digitally displayed. Accurate, reliable platinum resistance sensors, closely matched by Mettler, enable you to choose the furnace which precisely suits your needs. For detailed information, circle the number.
Mettler CIRCLE 1 4 4 O N READER SERVICE CARD
A N A L Y T I C A L CHEMISTRY, VOL. 45, NO. 14, DECEMBER
1973 · 1221 A
where G = dynode gain, typically 106 A = area of the photocathode illuminated—i.e., slit size, cm 2 Ν = incident light flux at the slit, photons c m 2 s e c - 1 If it is assumed that source noise predominates when a photomultiplier is used, the overall noise may be ex pressed by: i = if = Ia/Af112
amps rms
(6)
To compare the two systems, cer tain parameters must be normalized: the optical geometry between the source and monochromator must be the same, an equal photon flux per unit area must fall on both detectors at the same wavelength, and an equal signal integration time must be con sidered. If a photon flux of 108 photons c m ' 2 s e c 1 at 300 nm is considered, the following values may be substi tuted in Equation 1: Q = 0.4, G = 1, Ar = area of a single channel = 1.25 X 1 0 - 3 cm 2 , Τ = 32 X 10" 3 sec, Ta = 32 x 10- 6 sec. Thus: 0.4 x 1.6 x 10" XIX 1.25 X 10"^ Χ (ΙΟ")1 Χ 3 32 X 10~ /.= amps 32 X 10" x 10 - 1 2 a m p s When we substitute in Equation 2, where η = 2500 electrons/channelframe (15), 2500 X 1.6 X 10" 32 X 10 6 = 1.25 X 1 0 " " am Also, in Equation 3: if
amps rms
= 8 Χ 10~12 Χ
ιο-'ίV 32
Χ
Γ
a m p s rms Χ 10"3/ 13 = 4.4 Χ 10" amps r m s If the values of i„eq and if are sub stituted in Equation 4, if is clearly insignificant (t = 1.25 Χ 10" 1 1 amps rms), and in this case the source noise does not contribute to the over all noise. Thus, for a single channel on the vidicon target,
h
( S / N ) v i d i c o n = — = 0.64 ι Under our experimental conditions, however, the data from 100 scans are accumulated (total observation time, 3.2 sec). This gives a further S/N ratio enhancement of 100 1/2 —i.e., (S/N)vldicon = 6.4. For the single-channel system a typical slit size is 1 cm x 100 μηι— i.e., A = 10~ 2 cm 2 , and the quantum efficiency of an R106 photomultiplier at 300 nm is 0.15. Thus Equation 5 becomes:
Ia
= 10 6 X 10~ 2 X 10 8 X 0.15 x 1.6 x 10" 1 9 a m p s = 2.4 x 10~ 8 a m p s
If the total observation time is 3.2 sec, i = if = 2.4 x 10 ~ 8 X / ι \"! 10 2 amps rms \3.2/ 1.3 X 10~ 1 0 a m p s r m s The S/N ratio of the single-channel system is thus: (S/N)pmi =
—
180
Under these conditions the S/N ratio of the single-channel instrument should be approximately 30 times greater than that of the multichannel apparatus. In the above case, the main noise contribution in the multichannel sys tem was preamplifier noise. However, as the source intensity increases and flicker remains constant, the source noise begins to contribute. For exam ple, if the photon flux is 109 photons c m 2 s e c - 1 , if = 4.4 X 10~ 12 amps rms. Substitution in Equation 4 now substantially alters the overall noise, and the S/N ratio for 100 frame scans becomes 61, which makes the S/N ratio of the single-channel instrument only three times greater than that of the multichannel system. In an S/N ratio comparison, the many experimental variables are dif ficult to quantitate, but three main points become apparent: As source intensity increases, vidicon preampli fier noise becomes less significant compared with source noise, and a better S/N ratio should be obtained; when preamplifier noise predomi nates, an S/N improvement should be obtained by accumulating the data from several scans or by integrating the charge on the target for a longer period before readout; with SIT and SEC vidicons, the sensitivity is great er than the silicon vidicon, but preamplifier noise should be the same. In the low light level region, where preamplifier noise is dominant, the S/N ratio for the SIT and SEC vidicons will always be greater than for the silicon vidicon. It can be concluded from these cal culations that, using presently avail able light sources for multielement analysis, the silicon vidicon will al ways display a somewhat poorer S/N ratio than an instrument using a pho tomultiplier. This is supported by our experimental results. Possible Future Applications In the method for AA analysis de scribed above, the vidicon detection
1222 A · ANALYTICAL CHEMISTRY, VOL. 45, NO. 14, DECEMBER
1973
system is rapid, but the subsequent recording and interpretation of data are somewhat lengthy procedures. For a routine analysis instrument, it would be useful to have some form of on-line data storage or handling capa bility, especially when rapidly chang ing signals are being detected in a real-time scanning mode. A comput erized system would be highly desir able. When measuring steady-state sig nals, the video signal produced by the scanning electron beam is usually digitized and stored in some binary format in a memory unit. These data can then be recalled after the com pletion of a measurement cycle to study the spectral intensities. Real time data processing is not necessary here, and mass storage of data of this kind can be performed ideally by magnetic tape systems. Real-time data processing would be necessary, however, for some applications, such as locating the channel at the center of a signal peak which covers several channels or measuring transient events. If 500 channels are available for data and each scan cycle has a duration of 32 msec (as in the SSRI instrument described above), the time required to scan each channel is about 64 Msec, and fast A/D conver sion must be performed on the video signal to store all the data during each frame scan. In this case the speed of a digital computer is essen tial. A fully automated system incorpo rating a minicomputer should be able to perform two real-time functions: to store in real time the data from suc cessive frame scans, so that transient signals may be simultaneously de tected at several wavelengths and the signal shape with time determined after the measurement cycle; and to sum the data from several channels to obtain the total intensity over the emission profile, rather than measur ing the peak height—i.e., to perform real-time integration. Multielement determinations in both AA and AF could then be per formed for atomization methods which produce only transient signals. With this unique capability for fol lowing the transient atomization of several elements simultaneously, the vidicon could be used as a detector for multielement analysis by the Delves cup technique (16) or by any of the nonflame atomization cells now commercially available, such as the Techtron carbon rod atomizer or the Perkin-Elmer HGA 2000 graphite Massmann-type furnace (PerkinElmer, Norwalk, Conn.). These sys tems require only very small analyte samples compared with a flame, and with a vidicon, 10 or more elements could be determined in a sample of
only a few microliters. In t h e N e w York S t a t e D e p a r t m e n t of H e a l t h laboratories a Delves c u p m e t h o d for d e t e r m i n i n g lead in fingerprick blood s a m p l e s is now in routine use (17). T h i s m e t h o d could be e x t e n d e d to t h e s i m u l t a n e o u s analysis of o t h e r t r a c e m e t a l s in t h e s a m e 50-μ1. blood s a m ples. O t h e r possible a p p l i c a t i o n s of tele vision c a m e r a d e t e c t o r s in AA a n d A F using b o t h flame a n d nonflame a t o m reservoirs are n u m e r o u s , a n d t h e s y s t e m would be especially useful in a n y case where a n u m b e r of ele m e n t s are now d e t e r m i n e d in t h e s a m e s a m p l e by a t o m i c s p e c t r o m e t r y . E x a m p l e s are p o t a b l e waters a n d waste w a t e r s a m p l e s , where m e a s u r e m e n t of u p t o 20 t r a c e m e t a l s m a y be required (18), a n d u s e d l u b r i c a t i n g oils, where w e a r - m e t a l c o n c e n t r a t i o n s are m o n i t o r e d as criteria ofJ;he condi tion of t h e engine (19-22).
Conclusions M u l t i c h a n n e l photoelectronic de tectors of t h e t y p e described in t h i s article have m a n y p o t e n t i a l uses in atomic spectrometry. A computerized m u l t i c h a n n e l AA or A F s p e c t r o m e t e r incorporating such a d e t e c t o r would e n a b l e extremely r a p i d d e t e r m i n a t i o n of a n u m b e r of e l e m e n t s . T h e sim plicity of operation, precision, a n d se lectivity of AA are c o m b i n e d with t h e versatility of m e t h o d s s u c h as emis sion s p e c t r o g r a p h y , X - r a y fluores cence, a n d s p a r k source m a s s spec trometry. T h e instrumental arrange m e n t is far less c o m p l i c a t e d t h a n most m u l t i e l e m e n t s y s t e m s , r e q u i r i n g m i n i m a l o p e r a t o r skill; a n d it is cer t a i n l y less expensive t o construct a n d m a i n t a i n t h a n an X - r a y fluorimeter or a m a s s s p e c t r o m e t e r . T h e principal a d v a n t a g e s of a spec t r o m e t e r w i t h a vidicon d e t e c t o r over previously designed m u l t i c h a n n e l AA or A F s p e c t r o m e t e r s are threefold a n d m a y be s u m m a r i z e d as follows: it can be a s s e m b l e d from c o m m e r c i a l l y available c o m p o n e n t s w i t h o u t a major engineering effort; it h a s a large n u m b e r of d e t e c t o r c h a n n e l s , allowing t h e s i m u l t a n e o u s d e t e r m i n a tion of at least 8 or 10 e l e m e n t s , as well as b a c k g r o u n d correction proce dures; a n d it is a t r u e s i m u l t a n e o u s analysis i n s t r u m e n t , while most o t h e r s y s t e m s d e t e c t signals sequentially, a n d it is t h u s s u i t a b l e for observing t r a n s i e n t signals. O n e d i s a d v a n t a g e of t h e s y s t e m is t h a t a c o m p r o m i s e in optical align m e n t , flame conditions, a n d t y p e of flame (e.g., a i r - a c e t y l e n e or nitrous o x i d e - a c e t y l e n e ) is always necessary. F u r t h e r m o r e , w h e n some e l e m e n t s are in excess over others, t h e differ ence in useful working r a n g e b e t w e e n e l e m e n t s is b o u n d to i m p o s e a l i m i t a tion. T h e l a t t e r d i s a d v a n t a g e is com
mon to most multielement tech niques. Another disadvantage which m i g h t limit t h e m e t h o d is t h e r a t h e r poor S / N r a t i o which h a s been en c o u n t e r e d d u r i n g AA work, b u t t h i s is likely t o be i m p r o v e d w i t h integra tion. In spite of t h e s e l i m i t a t i o n s , we b e lieve t h a t vidicons a n d similar d e vices will b e c o m e widely used in t h e construction of m u l t i e l e m e n t a t o m i c s p e c t r o m e t e r s a n d will greatly in crease t h e value a n d versatility of AA and AF spectrometry.
Acknowledgment T h e a u t h o r s are grateful t o M . R. Zatzick ( S S R I n s t r u m e n t s Co.) for his v a l u a b l e c o m m e n t s a n d criticisms d u r i n g t h e p r e p a r a t i o n of t h e m a n u script.
References (1) K. W. Busch and G. H. Morrison, Anal. Chem., 45, 712A (1973). (2) G. Horlick and E. G. Codding, Pitts burgh Conference on Analytical Chemis try and Applied Spectroscopy, Cleve land, Ohio, Paper No. 25, March 1973. (3) R. E. Benn, W. S. Foote, and C. T. Chase, J. Opt. Soc. Amer., 39, 529 (1949). (4) S. A. Johnson, W. M. Fairbank, and A. J. Schawlow, Appt. Opt., 10, 2259 (1971). (5) R. E. Santini, M. J. Milano, H. L. Pardue, and D. W. Margerum, Anal. Chem., 44,826(1972). (6) A. B. Laponsky and V. J. Santilli, "Recent Developments in Low Light Level Camera Tubes," Reprint 48, Westinghouse Electric Corp., Horseheads, N.Y., 1971. (7) M. Margoshes, Pittsburgh Conference on Analytical Chemistry and Applied Spectroscopy, Cleveland, Ohio, Paper No. 99, March 1970. (8) M. Margoshes, Opt. Spectra, 4, 26 (1970). (9) T. Hirschfeld, U.S. Patent 3,728,029, April 17, 1973. (10) R. L. Vogenthaler and M. Margoshes, ibid., 3,728,576. (11) K. W. Jackson, Κ. Μ. Aldous, and D. G. Mitchell, Spectrosc. Lett., 6, 315 (1973). (12) F. W. Karasek, Res./Develop., 23, 47 (1972). (13) "Tivicon Image Tubes from Texas In struments," Texas Instruments Inc., Dallas Tex (14) J. D. Winef'ordner and T. J. Vickers, Anal. Chem., 36, 1947 (1964). (15) Technical Data Available on SSRI Optical Multichannel Analyzer, SSR In struments Co., Santa Monica, Calif. (16) Η. Τ. Delves, Anahst (London), 95, 431(1970). (17) D. G. Mitchell, K. M. Aldous, and F. J. Ryan, Anal. Chem., in press. (18) C. R. Parker, "Water Analysis by Atomic Absorption Spectroscopy," Varian Techtron Pty, Ltd., Springvaie, Australia, 1972. (19) J. A. Burrows, J. C. Heerdt, and J. B. Willis, Anal. Chem., 37, 579 (1965). (20) E. A. Means and D. Ratcliff, At. Absorpt. Newslett., i, 174 (1965). (21) S. Sprague and W. Slavin, ibid., ρ 367. (22) M. Gardels, D. Demers, and D. G. Mitchell, "Advances in Automated Analysis, Technicon International Con gress, 1970," ρ 513, Thurman Assoc, Miami, Fla., 1971.
The multi-purpose Mettler Automatic Titrator actually offers more automatic titration capability than any other comparable system. You can carry out automatic end point titration, recording titration, pH or pTon stat procedures. Its accuracy and versatility meet the exacting demands of the research laboratory. Its ease of operation makes it ideal for the quality control lab. For continuous automatic control, just start it up and press a button. Programming controls are simple. Rapidly interchangeable burettes in 1, 10, 20 and 50 ml snap-in and snap-out of the drive to speed titrant changes. The Mettler Automatic Titrator meets all potentiometric needs with aqueous and nonaqueous solutions, and provides direct entry into data processing systems. For more information, circle the number.
Mettler CIRCLE 1 4 5 O N READER SERVICE CARD
ANALYTICAL CHEMISTRY, VOL. 45, NO. 14, DECEMBER 1973 · 1223 A
