Modified Oscillating Mirror Rapid Scanning Spectrometer as a Detector for Simultaneous Multi-element Determination

Transcription

1 FEBRUARY 1978 The Anlyst Modified Oscillting Mirror Rpid Scnning Spectrometer s Detector for Simultneous Multi-element Determintion Oliver Rose, Jr.," Willim R. Heinemn nd Joseph A. Cruso DeprtmeBt of Chemistry, University of Cincinnti, Cincinnti, Ohio 45221, USA nd Fred L. Fricke Vol. 103 No Cincinnti District Food nd Drug Administrtion, 1141 Centrl Prkwy, Cincinnti, Ohio 45202, USA An oscillting mirror rpid scnning spectrometer hs been modified to improve its light throughput nd resolution. Results of the ppliction of this spectrometer to simultneous microwve-induced tomic-emission spectrometry nd simultneous crbon furnce tomic-bsorption spectrometry re given. Keywords Oscillting mirror rpid scnning spectrometer ; microwueinduced tomic-emission spectrometry; crbon furnce tomic-bsorption spectrometry ; multi-element determintion An oscillting mirror rpid scnning spectrometer1 cn be clssified s sequentil liner scn spectrometer.2 The oscillting glvnometer provides scnning wvelength window nd ech spectrl element is detected in rpid sequence by single photomultiplier tube. n preliminry evlution3 of this type of spectrometer s pplied to simultneous multielement emission spectrometry, it ws found tht its photomultiplier detection system nd wide opticl rnge would mke it suitble s multi-element detector. However, modifictions were needed in order to improve the opticl throughput nd resolution. This pper describes the modified rpid scnning spectrometer (RSS) nd its ppliction to simultneous tomic-emission spectrometry using the microwve-induced plsm (MP) s the source nd to simultneous crbon furnce tomic-bsorption spectrometry. Experimentl Apprtus Detection system Bsiclly, the detection system, consisting of the Hrrick Rpid Scn Spectrometer (Model RSS-B), the RSS signl processing module, the Nicolet NMR-80 minicomputer with the FT-74 softwre pckge, ws similr to tht described previ~usly.~ n order to void lengthy re-lignment of the optics from one ppliction of the RSS to the next, the instrument being lso used extensively in other projects s dul-bem spectrophotometer with xenon rc lmp source, two modifictions of the opticl lyout of the Model RSS-B described by Denton et Z.* were mde. Firstly, telescope entrnce lens (5 cm focl length, 2.5 cm dimeter), ligned with the entrnce slit (S,) nd the first sphericl mirror (SM,), ws. dded with the first plne mirror removed, nd secondly, 1P28 photomultiplier tube ws inserted immeditely fter the exit slit. Fig. 1 is digrm of the modified opticl lyout nd indictes tht only the opticl T ws found to be necessry for tomic spectrometry. A photomultiplier power supply (MP Model MP-1031) nd lbortory-constructed signl mplifier were used. By using the 1 P28 photomultiplier tube nd the seprte mplifier, the signl to noise rtio ws improved. (The mplifier circuit supplied by the mnufcturer contined pre-mplifier section nd log mplifier section. Only the output of the pre-mplifier section ws displyed in the preliminry ~tudy.~) This * Present ddress: Shepherd Chemic Compny, 4900 Beech Street, Cincinnti, Ohio 45212, USA. 113

2 114 ROSE et l. OSCLLATNG MRROR RAPD SCANNNG SPECTROMETER Anlyst, Vd. 103 improvement of the signl to noise rtio permitted the slit widths to be reduced, resulting in better resolution. Slit widths of 0.06 nd 0.18 mm for entrnce nd exit slits, respectively, were used in the present emission studies. S1 Red 52 SM. Fig. 1. Digrm of modified opticl lyout of Hrrick RSS-B rpid scn spectrometer s used for trce metl emission work: S is the emission source; S,. the entrnce slit; SM,, the first sphericl mirror; GM, the glvnometer mirror; SM,, the second sphericl mirror; GR, the grting ; M,, the flt mirror ; S,, the exit slit; nd PMT, the photomultiplier tube. n order to evlute the performnce of the RSS with opticl components other thn those supplied by the mnufcturer, other sources including copper hollow-cthode lmp (Tekmr), zinc, cdmium, led nd copper multi-element hollow-cthode lmp (Tekmr) nd low-pressure mercury vpour lmp were utilised for well defined spectrl lines. The hollow-cthode lmps were powdered by d.c. supply (Model GPS-2, Brnes Engineering Co.) nd both lmps were operted t 8mA. Before spectr were recorded following chnge of n opticl component, the opticl lignment ws optimised. To improve the resolution, the grting supplied by the mnufcturer, which ws ruled t 300 lines mm-1 nd blzed t 300 nm, ws replced with one ruled t 1200 lines mm-1 nd blzed t 250 nm. A comprison of the two grtings is mde in Tble. n order to enhnce the sensitivity by incresing the light throughput, nd to improve the resolution by illuminting more lines of the grting, the glvnometer mirror (4.8 mm dimeter) ws replced by lrger mirror (8.9 mm dimeter) fter study of vrious glvnometer mirror sizes with respect to repetition rte. To fill the lrger mirror with rdition, lrger first sphericl mirror (SM,) ws instlled. An ultrviolet spphire hemicylinder, 20 x 25 mm with 10-mm rdius, fitted in well with the use of the 8.9 mm dimeter glvnometer mirror. n the preliminry study,3 the minicomputer determined the repetition rte of the RSS. However, it ws found tht identicl spectr could be stored in different 1K memory loctions nd coincide more precisely (permitting exct subtrction of bckground) if the glvnometer mirror oscillted continuously. A squre-wve genertor (Model 126C, Exct Electronics nc.) ws then used to trigger ech scn of the glvnometer mirror nd ech sweep of the minicomputer. Thus, while the RSS collected spectrl informtion continuously, the minicomputer stored dt only upon commnd. Crbon cu$ vporistion ssembly - MP system The crbon cup vporistion ssembly nd the procedure for introducing nlyte into the

3 Februry, 1978 AS A DETECTOR FOR SMULTANEOUS MULT-ELEMENT DETERMNATON 1 15 MP hs been described previ~usly.~ The set power level of the microwve genertor ws 50 W direct nd 0.5 W reflected, while the optimum flow-rte of rgon ws 600 ml min-1. Desolvtion $@rtus - MP system This system ws similr to tht described by Mrgoshes nd Veillon6 nd modified by Skogerboe nd Colemn,' except the microwve cvity ws n Evenson &-wve cvity.8 Smll droplets of wter splshed into the side-rm below the condenser tht led to the qurtz tube nd extinguished the plsm. A piece of PTFE tpe ws suspended in front of the side-m opening in order to minimise this interruption of the signl. The microwve power ws 90 W direct nd 0.5 W reflected. The flow-rte of rgon ws 800 ml min-1 nd the solution spirtion rte ws 2.2 ml min-l. Crbon furnce tomic-bsor$tion system Rdition from the zinc, cdmium, led nd copper multi-element hollow-cthode lmp operted t 8 ma ws pssed through the crbon rod tomiser (Model 63, Vrin Techtron) nd then detected by the RSS. The tomistion temperture ws chosen to be the optimum for the determintion of copper with corresponding flow-rte of rgon of 4 1 min-l. Regents Aqueous stndrds pipetted into the crbon cup or crbon rod were diluted from stock solutions, which were either Fisher tomic-bsorption stndrds or solutions prepred by dissolving pure metls or lbortory-regent grde slts in n cid nd diluting the solution with de-ionised wter. Solutions spirted into the desolvtion pprtus were mde 0.1 N in hydrochloric cid. All other regents were of ACS regent grde. Dt Acquisition As in the preliminry study,3 the difference between the signl observed for the smple nd tht observed for the solvent ws used s direct mesure of the emission (bsorption for the crbon furnce tomic-bsorption system) of the elements. For determintion using the crbon cup vporistion ssembly - MP nd the crbon furnce tomic-bsorption systems, 130-nm wvelength window ws observed by the RSS with repetition rte of 73 Hz. For both systems, the minicomputer ccumulted dt only when the trnsient signls were detected by the RSS. For the emission system, 200 spectr were ccumulted, while for the crbon furnce tomic-bsorption system, 140 spectr were ccumulted. For determintion involving the use of the desolvtion pprtus - MP system, n 80-nm wvelength window ws observed by the RSS with repetition rte of 20 Hz nd 500 spectr were ccumulted fter spirtion of the smple solution. Results nd Discussion Evlution of Modified RSS Grting The use of more finely ruled grting in the RSS results in compromise between incresed resolution nd reduced mximum wvelength window size s the ngulr displcement (excursion) of the glvnometer mirror needs to be greter to disply the sme wvelength window size thn tht with less spectrl dispersion. Tble lists the mximum wvelength TABLE COMPARSON OF PERFORMANCE OF THE RSS WTH MOUNTNG OF DFFERENT GRATNGS 300 lines mm-l grting lines mm-1 grting Mximum wvelength window sizelnm * 170t Line hlf-width of Cu nm linelnm: * Limited by spectrl response of 1P28 photomultiplier. t Limited by size of second sphericl mirror (see Fig. 1). $ Wvelength window size 17 nm.

4 116 ROSE f?t d. : OSCLLATNG MRROR RAPD SCANNNG SPECTROMETER Anlyst, VOl. 103 window size nd line hlf-width observed when the 300 nd the 1200 lines mm-l grtings were mounted with ll other components unchnged. The improvement in resolution ws substntil s the line hlf-width (full width t hlf mximum height) of the copper nm line decresed from 1.9 to 0.4 nm. Fig. 2 shows the sme 17-nm window of the rdition from the copper hollow-cthode lmp, which includes the nd nm lines recorded fter opticl components were chnged. A comprison of spectr () nd (b) indictes the enchncement of resolution with more finely ruled grting mounted in the RSS. However, the mximum possible wvelength window decresed from 500 to 170 nm. With the 300 lines mm-1 grting mounted in the RSS, the mximum wvelength window size ws determined by the spectrl response of the photomultiplier tube. For exmple, with the photomultiplier tube supplied by the mnufcturer (RCA C31025Q), the mximum wvelength window is 730 nm, while with the 1P28, the mximum width is 500 nm. The sphericl mirror, SM,, limits the mximum wvelength window size with the 1200 lines mm-l grting mounted in the RSS. Any greter ngulr displcement (excursion) of the glvnometer mirror does not result in n increse in wvelength window becuse the pth of the light bem produced by the oscillting mirror is greter thn the width of SM,. Glvnometer nziwor size The use of the 4.8 mm dimeter glvnometer mirror supplied by the mnufcturer provided repetition rte fst enough to permit sufficient ccumultion of trnsient signls.3 n order to select the size of lrger mirror tht would increse the light throughput nd resolution, but still provide resonble repetition rte, the performnce of severl glvnometer mirrors ws studied by mounting these on the sme opticl scnner (Generl Scnning, Model 0612). For exmple, the use of 12.7 mm squre glvnometer mirror (re mm2), which incresed the light throughput mrkedly, decresed the repetition rte to 16 Hz (t scnning rte of 41 Hz) with the wvelength window of 50 nm. (Any subsequent reference to wvelength window size implies tht the 1200 lines mm-1 grting ws mounted in the RSS.) Any fster repetition rte cused significnt distortion of the spectrum displyed. This decrese in repetition rte is due, in prt, to the unfvourble rtio of mirror inerti to scnner rmture inerti.9 The time required for the mirror to come to rest fter ech scn is much greter thn tht required for smller mirror nd there is much more inerti to overcome t the beginning of ech scn. The mximum repetition rte would decrese still further if the wvelength window size were incresed becuse the rte of ngulr displcement would be lrger. This incresed rte provides greter momentum to reverse s the mirror re-sets fter the scn is completed nd more time is required to bring to rest the resulting incresed surfce distortions of the mirror. However, in pplictions tht need only slower repetition rte (less thn 20 Hz), but high light throughput, the use of lrger glvnometer mirror cn be n importnt modifiction TABLE 1 COMPARSON OF PERFORMANCE OF THE RSS WTH MOUNTNG OF DFFERENT GALVANOMETER MRRORS Mximum repetition rte/hz*.. 73 Scn rte t mximum repetition rtelhz Reltive intensity of Cu nm line....,. 1 Line hlf-width of Cu Linerity of scn 4.8 mm dimeter mirror (re 18.1 mm2) nm linet mm dimeter mirror (re 62.2 mm) 20 (correltion coefficient of wvelength vevsm position plot$) * Wvelength window size 130 nm. t Wvelength window size 17 nm. $ Thirteen emission lines of rdition from low-pressure mercury vpour lmp

5 Februry, 1978 AS A DETECTOR FOR SMULTANEOUS MULT-ELEMENT DETERMNATON 117 to the RSS-B spectrometer. t ws not possible to utilise lrger opticl scnner without totlly rebuilding the RSS-B. By use of different monochromtor, much lrger opticl scnners nd corresponding mirrors (e-g., n 875-mm2 mirror such s re found on Generl Scnning 300 Series modelsg) could be employed. An 8.9 mm dimeter glvnometer mirror ws chosen s compromise between incresed light throughput nd repetition rte for the RSS-B spectrometer. Tble 11 gives comprison of the performnce of the RSS when the 4.8 mm dimeter nd 8.9 mm dimeter mirrors re mounted in the RSS with ll other opticl components unchnged. The light throughput ws incresed by fctor of 3.1 even though the ctul observtion time (scn period multiplied by the repetition rte) ws decresed nd the resolution of the RSS incresed. A comprison of spectr (b) nd (c) in Fig. 2 indictes the improvement of the spectrl output of the RSS with the use of the lrger glvnometer mirror. Spectrum (c) hs been ttenuted by fctor of two for disply purposes. The mximum repetition rte ws reduced to 20Hz with wvelength window of 130 nm, compred with 73 Hz for the sme wvelength window width with the smller mirror mounted in the RSS. With the Model 0612 opticl scnner, the lrger mirror is best utilised in pplictions in which the observed signls re continuous so tht more fvourble signl to noise rtio is obtined. t should be indicted t this point tht even with the lrger mirror mounted, the grting ws severely under-illuminted, which ppers not to be drwbck in other pplictionsl94 but needs to be improved for future tomic-emission work, probbly by building new instrument. \/ () Cd nm Znnd Cd lines unresolved nm nm nm Wvelengthhm Fig. 2. denticl wvelength windows (width 17 nrn) of the emission of copper hollow-cthode lmp. Slit widths: entrnce 0.06 mm, exit 0.18 mm. Disply ttenution : (), counts cm-l; (b), counts cm-l; (c), counts cm-l. Opticl components mounted in the RSS: (), 4.8 mm dimeter glvnometer mirror, 300 lines mm-l grting; (b), 4.8 mm dimeter glvnometer mirror, 1200 lines mm-l grting; (t), 8.9 mm dimeter glvnometer mirror, 1200 lines mm-1 grting. Fig. 3. denticl wvelength windows (width 35nm) of the emission of zinc, cdmium, led nd copper hollow-cthode lmp. Slit widths: entrnce 0.06 mm, exit 0.18 mm. Opticl components mounted in the RSS: (), 4.8 mm dimeter glvnometer mirror, 300 lines mm-1 grting; (b), 8.9 mm dimeter glvnometer mirror, ines mm-1 grting.

6 118 ROSE et Z. : OSCLLATNG MRROR RAPD SCANNNG SPECTROMETER Anlyst, VoZ. 103 n the preliminry study of the RSS,3 it ws noted tht zinc nd cdmium could not be determined simultneously s the zinc nm nd cdmium nm lines were unresolved. By mounting the 8.9 mm dimeter glvnometer mirror nd the lines mm-1 grting in the RSS, these two lines cn be resolved. Fig. 3 gives comprison of the sme 35-nm window of rdition from zinc, cdmium, led nd copper multi-element hollowcthode lmp recorded with different opticl components mounted in the RSS. An increse in the wvelength window size decreses the resolution becuse fewer dt points per nnometer re ccumulted. The mximum wvelength window size tht cn be used for multielement determintions while still resolving the zinc nd cdmium nm lines ccording to the Ryleigh criterion of resolution ws 80 nm. Atomic-emission nd -bsorption Spectrometry After modifying the RSS spectrometer s described bove, it ws gin evluted s detector in simultneous multi-element determintions. Both the trnsient signls produced by the crbon cup vporistion ssembly - MP system nd the crbon furnce tomicbsorption system, nd the continuous signls produced by the desolvtion pprtus - MP system, were observed. n order to obtin n enhnced signl to noise rtio, the 4.8 mm dimeter glvnometer mirror ws mounted in the RSS for the detection of the trnsient signls, lthough the 8.9mm dimeter mirror ws mounted for the detection of continuous signls. Tble 11 lists the limits of detection for the three systems. These limits re not s low s it is possible to chieve with the MP, s Skogerboe nd Colemn indicte t is likely tht these high limits result from the low duty cycle (see below) s well s from the rther limited opticl throughput vilble with the RSS-B. With the crbon cup vporistion ssembly - MP system, bismuth, cdmium, mngnese, led, mgnesium nd copper were determined simultneously; Fig. 4 shows typicl bckground-corrected emission spectrum of this determintion. Of note is the resolution of the mngnese triplet tht ppered s one pek in n identicl wvelength window recorded in the preliminry study.3 As result of the intensity nd rndomness of the rgon bckground, its subtrction ws not precise. Zinc ws determined seprtely owing to the spectrl overlp of the zinc nd cdmium nm lines. Typicl nlyticl grphs for this system re shown in Figs. 5 nd 6. Two sets of units re given in the bscisse s the system is mss sensitive. n this study, the smple volume used ws 5 pl. The liner rnge is pproximtely one order of mgnitude. This rnge is more compressed thn is desirble; however, tht is to be expected when the limited opticl throughput of the system, which in prt leds to higher detection limits thn might otherwise be possible, is considered. ndeed, scnners cpble of supporting mirrors greter thn ten times the lrgest re tht we could mount in the RSS-B re redily vilble with sufficient frequency response (see bove). n order to utilise such scnners, however, different instrument is required s size restrictions in the RSS-B prohibit further modifiction. The onset of self-bsorption gives rise to slope of bout 0.5 t higher concentrtions. This is not unexpected, prticulrly with the crbon cup experiments where the prticle density in the trnsient pulse is likely to be high enough to give rise to this phenomenon. Element Zn Bi Cd Mn Pb Mg cu Wvelength/nm TABLE 11 LMTS OF DETECTON* (p.p.m.) Crbon cup vporistion ssembly - MRf Desolvtion pprtus - MP Crbon furnce tomic bsorptiont * Defined s tht concentrtion which produces net signl twice the stndrd devition of the bckground signl. t Smple volume 5 p1.

7 Februry, 1978 AS A DETECTOR FOR SMULTANEOUS MULT-ELEMENT DETERMNATON 119 Unquestionbly, the MP is cpble of providing lower detection limits nd extended liner rnges.10 Nevertheless, the limits of detection for the crbon cup vporistion system nd the desolvtion system re in the sub-prts per million rnge nd re mutully comprble. These detection limits re surprising in view of the low duty cycle for the nlyticl line (line smpling time). For exmple, consider typicl crbon cup experiment with 130-nm window, repetition rte of 73 Hz nd scnning rte of 101 Hz. Also, ssume tht the 220 W vel en gth/nm 300 Fig. 4. Typicl bckground-corrected emission spectrum obtined for simultneous determintion of bismuth, cdmium, mngnese, led, mgnesium nd copper using the crbon cup vporistion ssembly - MP system. mximum residence time for cdmium in the plsm is 1 s nd tht 0.5 nm represents the cdmium slice of the 130-nm totl. Bsed on the repetition nd scnning rtes, 73 spectr re tken per second with the totl time per 130-nm spectrum given s s (l/lol). Thus, in the residence time of 1 s, totl of 0.73 s is spent in smpling the 130-nm window giving 2.8 ms on time for the cdmium line. Thus, while the detection levels for direct-reding plsm emission spectrogrphs re less, the observtion time per nlyticl line is t lest 1000 times greter. Cd nrn i Mn257.6nm 1 > c,.= 100 C + f + m 5 10 u 1 ( 1.o 10 Concentrtion, p.p.rn Arnountlng Fig. 5. Log - log plots of nlyticl grphs for cdmium, copper nd bismuth obtined using the crbon cup vporistion ssembly - MP system o 10.0 Concentrtion, p.p.m Arnount/ng Fig. 6. Log-log plots of nlyticl grphs for mngnese nd led obtined using the crbon cup vporistion ssembly - MP system.

8 120 ROSE et l. : OSCLLATNG MRROR RAPD SCANNNG SPECTROMETER AnZyst, VoZ. 103 As the lrger mirror ws mounted for determintion using the desolvtion pprtus- MP system, zinc nd cdmium were determined simultneously in n 80-nm window tht lso includes bismuth, mngnese nd mgnesium. The emission of led t low concentrtion levels ws obscured by bckground emission. Copper ws determined in seprte window. Typicl nlyticl grphs for this system re shown in Figs. 7 nd 8. The liner rnge is lso pproximtely one order of mgnitude for this system. g 100 C Q) 4-l c Q) + = 10.El00 c J 4-l C Q,.+ - m 0, o 10.0 Concentrtion, p.p.m o 10.0 Concentrtion, p.p.m. Fig. 7. Log - log plots of nlyticl Fig. 8. Log - log plots of nlyticl grphs for mgnesium, mngnese nd copper obtined using the desolvtion ssembly - MP system. grphs for cdmium nd zinc obtined using the desolvtion ssembly - MP system. Cdmium, led nd copper were determined simultneously by using the crbon furnce tomic-bsorption system. No dt were obtined for zinc owing to the presence of zinc in the de-ionised wter nd the problems ssocited with it for furnce tomic-bsorption determintion.11 However, the determintion of zinc is sensitive, being in the prts per lo9 rnge. As is to be expected, considering the sensitivity of the crbon furnce tomicbsorption method, the limits of detection for this system re superior to those for the two emission systems. However, the liner rnge is nrrow, especilly for cdmium, mking n optimum dilution difficult when working with complex smples. Mtrix Effects n order to study possible mtrix effects for the crbon cup experiment, 500 ng of sodium (5 p1 of 100 p.p.m. sodium solution from stock solution of sodium nitrte) were vporised into the plsm. As the sodium pssed through the plsm the intense ornge emission ws noted, but the plsm ppered to be stble. The direct power incresed only slightly nd the reflected power did not chnge. After severl vporistions, the qurtz tube bove the plsm ws coted with white residue. Tble V gives the limits of detection of cdmium, mngnese, mgnesium, copper nd zinc in 100 p.p.m. solutions of sodium slt. There ws little difference between these vlues in comprisonwith the limits of detection in queous solutions (Tble 111). As cn be seen, the concentrtion of the mtrix element is usully more thn times tht of the nlyte nd in one instnce bout times. Five microlitres of p.p.m. solution of sodium were injected into the cup nd ng of sodium were vporised into the plsm, cusing the plsm to contrct pprecibly. Also, there ws considerble memory effect s the sodium signl persisted for severl minutes. However, csul observtions indicte tht it my be possible to tune the MP with continuous flow of sodium slt solution comprble with the mount tht psses through the plsm upon vporistion of 5 OOO ng of sodium in crbon cup. Tuning the MP to the mtrix is very importnt in sustining plsm. Skogerboe nd Colemn spirted solutions contining 1000 p.p.m. of sodium into desolvtion - MP system without extinguishing the plsm.' However, when p.p.m. sodium solution ws spirted in the desolvtion pprtus described bove, it ws not possible to sustin plsm, lthough it must be mentioned tht the tuner hd been

9 Februry, 1978 AS A DETECTOR FOR SMULTANEOUS MULT-ELEMENT DETERMNATON 121 TABLE V LMTS OF DETECTON* (p.p.m.) N THE PRESENCE OF A MATRX ELEMENT (100 p.p.m. OF SODUM) Crbon cup vporistion Desolvtion Element Wvelength/nm ssembly - MP pprtus - MP :it Mn Mg cut * Defined s tht concentrtion which results in net signl twice the stndrd devition of the bckground. t Zinc determined seprtely in the crbon cup experiment; copper determined seprtely in the desolvtion experiment. rebuilt in our lbortories nd ws not operting completely stisfctorily. A plsm ws sustined with 100 p.p.m. solution of sodium spirting into the desolvtion pprtus fter the spirtion rte hd been decresed to 0.9 ml min-1 nd the rgon flow-rte incresed to 1.91 min-l. With the spirtion of the 100 p.p.m. sodium solution into the desolvtion pprtus the entrnce lens becme coted with slt, becuse the strem of gs leving the qurtz tube is directed t the lens. Limits of detection for zinc, cdmium, mngnese, mgnesium nd copper in 100 p.p.m. solution re lso given in Tble V. The higher detection limits in solution in 0.1 N hydrochloric cid in the presence of the 100 p.p.m. solution of sodium s compred with those in just the 0.1 N hydrochloric cid (Tble 111) re ttributed to the reduction in trnsport efficiency in the presence of sodium,' to chnges in the plsm energy chrcteristics s result of the lrge excess of sodium' nd, to lrge extent, the decrese in the spirtion rte. Conclusions Ech of the three systems hs distinct dvntges. The crbon cup vporistion ssembly - MP system hs the potentil for simultneous determintion when smple volume is criticl. The desolvtion pprtus - MP system offers the dvntge of observtion of continuous signl. For exmple, in this study, the lrger mirror could only be mounted in the RSS for the detection of continuous signl. The crbon furnce tomic-bsorption system gives superior sensitivity of detection, but is limited by the vilbility of multi-element sources of rdition. This study indictes tht rpid scnning spectrometry my provide useful detector for simultneous tomic spectrometry. However, future work must increse the light throughput still further in order to lower the limits of detection, thereby incresing the liner rnge nd fulfilling the cpbilities of the MP.lO Also, more fully illuminted optics will enhnce the resolution still further. The generl concept of n oscillting mirror rpid scnning spectrometer with photomultiplier detection my well provide n nswer to the growing need for n inexpensive detector for simultneous multi-element determintions. O.R., W.R.H. nd J.A.C. grtefully cknowledge prtil finncil support provided by the Ntionl nstitute of Occuptionl Sfety nd Helth nd by the Ntionl Science Foundtion with Grnts ROH nd CHE , respectively. References Strojek, J, W., Gruver, G. A., nd Kuwn, T., Anlyt. Chem., 1969, 41, 481. Winefordner, J. D., Fitzgerld, J. J., nd Omenetto, N., Appl. Spectrosc., 1975,29, 369. Rose, O., Jr., Mincey, D. W., Ycynych, A. M., Heinemn, W. R., nd Cruso, J. A., Anlyst, 1976, 101, 753. Denton, M. S., DeAngelis, T. P., Ycynych, A. M., Heinemn, W. R., nd Gilbert, T. W., Anlyt. Chem., 1976, 48, 20. Fricke, F. L., Rose, O., Jr., nd Cruso, J. A., AnZyt. Chem., 1975, 47, Mrgoshes, M., nd Veillon, C., Spectrochim. Ad, 1968, 23B, 503. Skogerboe, R. K., nd Colemn, G. N., AppZ. Spectrosc., 1976, 30, 504. Fehsenfeld, F. C., Evenson, K. M., nd Broid, H. P., Rev. Scient. nstrum., 1965, 36, 294. Technicl Dt vilble on Opticl Scnners, Generl Scnning nc., Wtertown, Mss., 02172, USA. Skogerboe, R. K., nd Colemn, G. N., Anlyt. Chem., 1976, 48, 611A. Evenson, M. A., nd Anderson, C. T., Clin. Chem., 1975, 21, 537. Received June loth, 1977 Accepted August 25th, 1977